# AI Visibility Desk: full text of guides and glossary > White-label AI visibility audits for agencies across ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity and Claude. This file holds the full text of every guide and glossary entry on https://aivisibilitydesk.com. The short index is at https://aivisibilitydesk.com/llms.txt. --- # AI Overview Tracker: How to Get Cited in AI Overviews URL: https://aivisibilitydesk.com/blog/get-cited-in-ai-overviews Updated: 2026-11-20 > How Google AI Overviews pick cited sources, how to check and track whether your site is cited, and what the evidence says raises your chances of a citation. Google AI Overviews cite pages chosen by Google's normal ranking systems, run across the original question and several related sub-questions, so the pages most likely to be cited are indexed, snippet-eligible pages that rank for the question or for the questions behind it. You check whether you are cited by searching your target questions and reading the source list, or by using an AI Overview tracker that records answers and citations across many questions and competitors. What raises your odds is mostly strong SEO for those sub-questions, direct and specific answers on the page, and being mentioned on other sites Google already trusts. ## How do AI Overviews choose which sources to cite? Google has published more about this than most AI companies, and the core message is that AI Overviews sit on top of regular Search. - **Normal ranking systems.** Google's [guide to generative AI features](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide) says AI Overviews and AI Mode rely on its core ranking and quality systems, and that SEO best practices still apply. - **Query fan-out.** The model generates a set of related queries at the same time to gather more results, then writes the answer from what it finds. A page can be cited because it ranks for a sub-question, even if it does not rank for the original query. - **Eligibility.** A page must be indexed and eligible to appear in Search with a snippet ([AI features and your website](https://developers.google.com/search/docs/appearance/ai-features)). Snippet controls such as `nosnippet` and `max-snippet` also apply to AI features. Google's newer guide adds that a site must be included in Search generative AI features in Search Console, so check that setting if you have ever changed it. - **Not every query gets one.** Google says AI Overviews appear only when its systems judge them additive to classic results, so they often do not trigger. Google also says there is no special markup, file or format needed: no llms.txt, no AI-specific schema, and no need to rewrite or chunk content for AI. ## What does the data say about cited pages? Independent studies fill in what Google does not say. They are correlational, so they show what cited pages have in common rather than what caused the citation. ### Ranking still matters, but less directly than it did Ahrefs analysed 1.9 million citations from 1 million AI Overviews in 2025 and found that [76% of cited pages ranked in the top 10](https://ahrefs.com/blog/search-rankings-ai-citations/) for the same keyword. A [2026 rerun](https://ahrefs.com/blog/ai-overview-citations-top-10/) on 4 million AI Overview URLs found the share had dropped to about 38%, with roughly 37% of cited pages not ranking in the top 100 for that keyword at all. Ahrefs attributes the change to fan-out: Google is citing pages that rank for related sub-queries. The two studies used different datasets and methods, so the numbers are not a like-for-like trend, but the direction is clear. The practical reading: ranking for the main keyword is no longer enough on its own. You need pages that rank across the cluster of questions a fan-out would generate. ### Video and YouTube show up In the same 2026 study, 18.2% of the cited pages that did not rank in the top 100 were YouTube URLs. Google's own guide notes that images and video can add visibility in its AI features. For how-to and product questions, a good video may be the easiest route into a citation. ### Brand mentions correlate with being named Citations and brand mentions are different things: a page can be cited without your brand being named, and your brand can be named from someone else's page. For mentions, Ahrefs found that web mentions of a brand had a [0.664 correlation with AI Overview brand visibility](https://ahrefs.com/blog/ai-overview-brand-correlation/) across 75,000 brands, stronger than backlinks or domain rating. Being discussed on other sites seems to matter more than links pointing at you. ## How do you check whether you are cited in AI Overviews? There are three ways, and they suit different jobs. | Method | Good for | Limits | |---|---|---| | Manual Google search | Spot-checking a few important questions | Results vary by location, device and login; slow for many questions; no competitor totals | | Search Console Generative AI performance report | Seeing how your own content performs in Google's AI features | Your site only; does not show competitors or the full answer text | | AI Overview tracker | Running many questions, recording the answer, cited sources and brands named, for you and competitors | Depends on the tool's location settings and the questions you choose | ### Checking by hand Search each target question in a private window, set to the country you care about, and expand the AI Overview to see every source. Note three things: whether an AI Overview appeared at all, whether your site is in the source list, and whether your brand is named in the text. Repeat on a few different days if the question is important; answers change. ### Using an AI Overview tracker Manual checks stop scaling at around a dozen questions. An AI Overview tracker queries Google for each question and records the answer, the cited sources and the brands mentioned, so you can see patterns: which domains Google cites most in your category, which competitors appear, and which questions do not trigger an AI Overview at all. When choosing an AI citation tracking tool, check that it shows the full answer and source list rather than only a score, that it lets you add competitors, and that it is clear about which country it queries from. Be wary of any tool that claims access to internal Google metrics; Google warns against those in its own guide. The most useful output is the list of most-cited domains for your questions. If Google keeps citing the same three comparison sites, review platforms or forums, those are the places to earn a mention. ## What increases your chances of being cited? These steps follow from Google's documentation and the studies above. None of them guarantees a citation. ### 1. Make sure you are eligible Confirm key pages are indexed, not blocked by robots.txt or a CDN rule, and not carrying `nosnippet` or a very short `max-snippet`. Check the Search Console setting for generative AI features. This is the only step with a hard yes or no. ### 2. Cover the questions behind the question For each target query, list the sub-questions a fan-out would ask: cost, alternatives, comparisons, "for small businesses", "how to set up". Make sure you have a page that genuinely answers each one, and that those pages are linked from the main page. This is where the 2026 Ahrefs data points. ### 3. Answer directly, near the top Start pages with a plain, specific answer to the question in the heading, then the detail. Include facts that can be quoted: numbers, prices, limits, dates, named steps. The [GEO research paper](https://arxiv.org/abs/2311.09735) found that adding statistics, quotations and citations raised source visibility by up to 40% in its benchmark. That was a lab setting, not live Google, but it matches what Google asks for. ### 4. Offer something that is not everywhere else Google's guide asks for a unique point of view and non-commodity content. If ten pages say the same thing, any of them can be cited. First-hand testing, original data, real screenshots and expert opinion give Google a reason to pick yours. ### 5. Add video and images where they help For tutorials, product demos and comparisons, a YouTube video with a clear title and description gives you a second route into the answer. ### 6. Earn mentions on sites that already get cited Use your tracker's citation list. Getting your brand into a review site, comparison article or forum thread that Google already cites for your questions is often faster than making your own page rank. ### 7. Avoid shortcuts Mass-produced pages written to catch AI answers fall under Google's scaled content abuse policy, and Google calls seeking inauthentic mentions ineffective. They risk more than they gain. ## Is an AI Overview citation worth the effort? It depends on what you expect from it. [Pew Research Center](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/) found that only 1% of visits to Google results pages with an AI summary included a click on a link inside it, and users clicked regular results less often too (8% of visits against 15% without a summary). So a citation is mainly about being present in the answer the buyer reads. For commercial questions, being named and linked next to competitors is the new equivalent of being on page one, even if fewer people click. Track citations together with brand mentions, because the mention is often what the reader remembers. For the bigger picture of how assistants pick brands, see [AI search visibility](/blog/ai-search-visibility), and for working beyond Google, [LLM SEO](/blog/llm-seo). ## Tracking AI Overview citations with AI Visibility Desk AI Visibility Desk includes Google AI Overviews alongside ChatGPT and Gemini. For each buyer question it reads the AI Overview, records which brands are named and which sites are cited, and reports how often the client's own site is among the sources, which other sites were cited, and which competitors appear. It also checks whether the client's robots.txt blocks the crawlers that feed AI search. Audits can re-run weekly or monthly, so you can watch citations change over time. Agencies publish the results under their own brand at a private link. See [how an AI visibility checker works](/ai-visibility-checker) or the [plans](/pricing). ## Frequently asked questions ### How does Google choose which sites to cite in AI Overviews? Google builds AI Overviews on its normal ranking and quality systems. The model runs several related searches at once (query fan-out) and cites pages from those results. A page must be indexed and eligible to show with a snippet. Pages that rank for the question, or for its related sub-questions, are the likeliest to be cited. ### How can I check if my site is cited in AI Overviews? Search your target questions on Google and look at the sources listed with the AI Overview, ideally from a clean browser and the right country. Google's Search Console also has a Generative AI performance report. To track many questions and competitors at once, use an AI Overview tracker that records the answer and its cited sources. ### Do I need special markup or llms.txt to appear in AI Overviews? No. Google says there are no extra requirements to appear in AI Overviews beyond normal Search eligibility, no special schema is needed, and Google Search ignores llms.txt files. Structured data is still worth having for regular rich results, but it is not a requirement for AI features. ### Does being cited in an AI Overview bring traffic? Some, but less than a top organic ranking used to. In Pew Research Center's 2025 browsing study, only 1% of visits to Google results pages with an AI summary included a click on a link inside the summary. A citation is mainly visibility and credibility: the brand is named and linked inside the answer. --- # AI Search Visibility: How Assistants Choose Brands URL: https://aivisibilitydesk.com/blog/ai-search-visibility Updated: 2026-11-20 > How ChatGPT, Gemini and Google AI Overviews decide which brands to name and which sources to cite, and what actually moves AI search visibility for a brand. AI search visibility is how often AI assistants name your brand, and cite your pages, when people ask questions about your category. ChatGPT, Gemini and Google AI Overviews build their answers from two things: what the model learned in training and the web pages it retrieves at the moment of the question. Brands that are described clearly and often across the sites those systems retrieve get named; brands that only describe themselves on their own site usually do not. This article explains how each engine assembles an answer, what that means for which brands and sources make it in, and which kinds of work tend to move the result. For the broader picture, see [what AI visibility is](/blog/what-is-ai-visibility). ## How do AI assistants build an answer? All three engines covered here combine a language model with web retrieval. The rough sequence is: 1. **Interpret the question.** The system works out what the user wants and, often, breaks it into several sub-questions. 2. **Retrieve.** It runs searches and pulls in pages that look relevant. 3. **Write.** The model writes an answer using the retrieved text plus what it already knows, and attaches some of the pages as sources. The model's training gives it a prior sense of which brands belong to a category. Retrieval updates and narrows that with current pages. A brand needs to be present in both to be named reliably: known to the model, and described in the pages that get retrieved for that kind of question. ## How Google AI Overviews choose sources Google's documentation is explicit that there is [no special optimisation needed for AI Overviews or AI Mode](https://developers.google.com/search/docs/appearance/ai-features), and no extra technical requirements. A page has to be indexed and eligible to show a snippet in normal Google Search. Google also says no new machine-readable files or special schema are required. What Google does describe is **query fan-out**: AI Overviews and AI Mode may run several related searches across subtopics and data sources to build one response. That has a practical consequence. The pages cited under an overview are not only the ones ranking for the exact query the user typed; they can come from searches on adjacent questions. A detailed page answering "how to invoice a client in another currency" can end up cited for a broader question about freelance accounting tools. Not every search shows an overview. In Pew Research's March 2025 browsing data, [18% of Google searches produced an AI summary, rising to 53% for searches of ten words or more](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/). Longer, question-shaped searches are where overviews show up most, which is also where buyer research happens. ## How ChatGPT chooses brands and sources When ChatGPT answers with web search, the pages come from OpenAI's search systems. OpenAI's crawler documentation draws three lines: - **OAI-SearchBot** surfaces websites in ChatGPT's search features. OpenAI states that [sites that opt out of OAI-SearchBot will not be shown in ChatGPT search answers](https://developers.openai.com/api/docs/bots), apart from navigational links. - **GPTBot** collects content that may be used to train models. Blocking it signals that your content should not be used for training. - **ChatGPT-User** fetches a page when a user's action triggers it, not as an automatic crawl. These are separate decisions. A site can block training (GPTBot) and still allow search (OAI-SearchBot). A blanket disallow rule or a CDN bot setting can block both without anyone deciding to, which removes the site from ChatGPT search answers. It is the first thing to check. Beyond access, ChatGPT recommendations reflect what the retrieved pages say. If the top comparison articles, review pages and forum threads for a category list four brands, those four are the likely answer. ## How Gemini chooses brands and sources Gemini is Google's assistant, and Google documents that Gemini models can [ground their answers in Google Search results](https://ai.google.dev/gemini-api/docs/google-search). In practice that means the same fundamentals that make a page retrievable by Google help it reach Gemini answers. ## Why do the answers keep changing? Because generation involves randomness and retrieval shifts with the index, the same question asked twice can return a different list. A 2025 study by SparkToro and Gumshoe, with volunteers running 12 prompts 2,961 times across ChatGPT, Claude and Google's AI, found [less than a 1 in 100 chance that ChatGPT or Google's AI would give the same list of brands in two responses](https://sparktoro.com/blog/new-research-ais-are-highly-inconsistent-when-recommending-brands-or-products-marketers-should-take-care-when-tracking-ai-visibility/). Order varied even more. The authors argue that how often a brand appears across many runs is meaningful, while its rank in a single answer is not. For measurement this means two things. Ask each question more than once, and report visibility as a rate across many answers, not a "position" from one screenshot. We cover the method in [how to run an AI visibility audit](/blog/how-to-run-an-ai-visibility-audit). ## What actually moves AI search visibility? There is less hard evidence here than the volume of advice suggests. What follows sticks to what is documented or directly observable in audits. ### 1. Let the right crawlers in Check robots.txt, firewall and CDN bot rules for OAI-SearchBot and Googlebot. If the search crawlers cannot fetch the site, nothing else matters. This takes minutes and is the most common fixable cause of a brand being absent from cited sources. ### 2. Be present where the assistants look When you audit the citations behind answers in a category, review sites, comparison articles, forum threads, YouTube and reference sites often appear next to the brands' own pages. If a competitor is named and you are not, look at which third-party pages were cited in that answer. Those are the pages that shape the answer, and the practical work is earning a place on them: getting included in comparison roundups, collecting reviews on the sites that get cited, answering real questions in communities where you have a genuine reason to be there. ### 3. Write pages that answer specific questions Because of query fan-out and the way assistants pull passages, specific pages do better than broad ones. A page that answers one buyer question directly in its first paragraph, then supports it, is easier to cite than a general product page. Cover the comparison, pricing, use-case and how-to questions buyers actually ask, not only "best X" lists. ### 4. Add facts, sources and quotations The research that coined the term generative engine optimization tested this. In [Aggarwal et al.'s GEO study](https://arxiv.org/abs/2311.09735), adding citations to sources, quotations and statistics to content were among the most effective changes, improving visibility in generated answers by up to about 40% in their benchmark. Keyword stuffing did little and in one test made things worse. The study used its own test engines, so treat the numbers as directional, but the direction matches what you see in cited pages: specific, checkable claims get used. ### 5. Keep the brand description consistent Assistants match a brand to a category through repeated descriptions across sites. If your homepage, directory listings, review profiles and press mentions describe the product differently, the association is weaker. Say what the product is, who it is for and how it differs, in the same words, everywhere you control. ### 6. Fix what is said, not only whether you are named A mention can be negative. If assistants repeatedly describe the brand as expensive or hard to set up, that comes from somewhere: reviews, forum threads, an outdated comparison. Find the sources and address the substance. ## What does not move it much - **Special files or markup aimed at AI.** Google states no special schema or AI text files are required for AI Overviews. Structured data is still useful for normal search features. - **Repeating keywords.** The GEO study found keyword stuffing ineffective. - **Chasing one answer.** Because answers vary run to run, rewriting a page because one screenshot left you out is reacting to noise. Look at rates across many runs. ## How AI search optimization fits with SEO AI search optimization builds on SEO. Google ties AI Overview eligibility directly to normal indexing, and ChatGPT search depends on its own crawl of the open web. The differences are in emphasis: more weight on third-party coverage, on specific answer-first passages, and on measuring brand mentions rather than rankings. For how the labels relate, see [GEO vs AEO vs SEO](/blog/geo-vs-aeo-vs-seo). ## How to measure AI search visibility Pick 20 to 30 buyer questions that do not name the brand, ask each on ChatGPT, Gemini and Google more than once, and record mention rate per engine, [share of voice](/blog/ai-share-of-voice) against competitors, and the sources cited. Repeat with the same questions later to see whether the work moved anything. [AI Visibility Desk](/ai-visibility-checker) does this for agencies: it asks each question on ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity and Claude (twice on ChatGPT and Gemini, because their answers vary), scores mentions, position, share of voice and citations, checks the site's crawler access, and publishes a report under the agency's brand. Audits can re-run weekly or monthly, so the report shows the trend. [See the plans](/pricing) if you want to try it on a client. ## Frequently asked questions ### What is AI search visibility? AI search visibility is how often a brand is named, recommended or cited when people ask AI assistants such as ChatGPT, Gemini or Google AI Overviews about its category. It is usually measured as the share of relevant answers that mention the brand, alongside which competitors and which sources appear. ### How does ChatGPT decide which brands to recommend? When ChatGPT searches the web, it builds the answer from pages it retrieves and from what the model learned in training. Brands that appear often and favourably in the pages it finds, such as reviews, comparisons and forum threads, are more likely to be named. Sites that block OAI-SearchBot are not shown in ChatGPT search answers. ### Do I need special markup to appear in Google AI Overviews? No. Google says there are no additional requirements and no special optimisation needed for AI Overviews or AI Mode. A page must be indexed and eligible to show a snippet in normal search. Good SEO fundamentals and content that answers the question clearly remain the route in. ### How is AI search optimization different from SEO? SEO aims to rank a page in a list of links. AI search optimization aims to have the brand named in the written answer and its pages cited as sources. Much of the groundwork is shared, but AI answers lean more on third-party mentions, clear factual passages and coverage across many sites. --- # AI Share of Voice: Formula, Example and Pitfalls URL: https://aivisibilitydesk.com/blog/ai-share-of-voice Updated: 2026-11-20 > What AI share of voice means, how to calculate it from ChatGPT, Gemini and Google AI Overviews answers, a worked example, and the traps that skew the number. AI share of voice is the portion of brand mentions in AI-generated answers that go to your brand, compared with a set of competitors you choose. You ask ChatGPT, Gemini and Google AI Overviews a fixed list of buyer questions, count the answers that name each brand, and divide your count by the total across all the brands. If answers name you 30 times and your competitors 70 times between them, your AI share of voice is 30%. Share of voice is one of the main numbers in an [AI visibility audit](/blog/how-to-run-an-ai-visibility-audit). This guide covers the formula, a worked example and the mistakes that most often make the number misleading. ## Where does the idea come from? Share of voice is an old advertising measure: a brand's portion of all advertising, or all media mentions, in its category. Applied to AI search, the "voice" is the written answer an assistant gives when someone asks about the category. Buyers often read that answer instead of clicking through to results; Pew Research found Google users clicked a regular result in [8% of visits when an AI summary appeared, against 15% when one did not](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/). If the answer is where the shortlist forms, your share of the names in it matters. ## The AI share of voice formula ``` AI share of voice (brand) = answers that name the brand ÷ sum of answers that name each brand in the comparison set × 100 ``` Three definitions make the formula usable: - **The comparison set.** Your brand plus a fixed list of competitors, usually three to five. Brands outside the list are not in the denominator. - **The unit is the answer.** A brand counts once per answer that names it, whether the answer mentions it once or five times. This stops a long answer that repeats one name from inflating that brand's share. - **Only real answers count.** A Google search with no AI Overview, or a request that failed, contributes nothing to any brand. Some teams weight mentions by position or count every repetition. Both are defensible, but they make the number harder to explain and, given how much order varies between runs, position weighting adds noise. Counting answers is the simplest version that holds up. ## Share of voice vs mention rate These two numbers answer different questions, and reports should show both. | Measure | Formula | Question it answers | |---|---|---| | Mention rate | Answers naming the brand ÷ answers returned | How often does the brand make it into an answer at all? | | Share of voice | Answers naming the brand ÷ all brand mentions in the set | How much of the conversation does the brand get against these competitors? | Mention rate is absolute; share of voice is relative. A brand can improve its mention rate while its share of voice falls, if competitors improved faster. ## A worked example The numbers below are **illustrative, invented for this example**, not from a real audit. An agency audits a project-management tool, "Brand A", against four competitors. It uses 20 buyer questions, each asked twice on ChatGPT, twice on Gemini and once on Google: 100 checks in all. Eight checks return nothing (a few Gemini requests never come back, and several Google searches show no AI Overview), leaving 92 answers. Counting the answers that name each brand: | Brand | Answers naming it | Share of voice | |---|---|---| | Competitor B | 61 | 61 ÷ 180 = 33.9% | | Competitor C | 45 | 45 ÷ 180 = 25.0% | | Brand A (client) | 38 | 38 ÷ 180 = 21.1% | | Competitor D | 22 | 22 ÷ 180 = 12.2% | | Competitor E | 14 | 14 ÷ 180 = 7.8% | | **Total mentions** | **180** | **100%** | The total is higher than 92 because one answer usually names several brands. Brand A's mention rate is 38 ÷ 92 = 41%: it appears in about four answers out of ten. Its share of voice is 21%, third of five. The two numbers together tell the story better than either alone. Brand A is reasonably visible, but Competitor B is named in two answers out of three and owns a third of the conversation. Now watch what the comparison set does. Drop Competitor E from the list and the total falls to 166; Brand A's share rises to 38 ÷ 166 = 22.9% without a single answer changing. Add a dominant brand to the list and the client's share falls. That is why the list must stay fixed between audits. ## Breaking share of voice down One overall number hides most of the useful detail. Split it: - **By engine.** A brand can be strong on ChatGPT and almost absent from AI Overviews. Since the sources each engine retrieves differ, the fixes differ too. - **By question type.** Best-of, comparison, price, how-to and use-case questions often produce different winners. A competitor that owns price questions has usually earned a place on the pages that answer them. - **Head-to-head.** For each competitor, count the questions where only they were named, only the client, both, or neither. "Only they" questions are the concrete gaps to work on. ## Pitfalls that skew AI share of voice ### Brand names that are ordinary words A text search for "Wave", "Monday" or "Notion" will count sentences that have nothing to do with the brand. One common-word competitor can take a share of voice it never earned and push everyone else's down. Read the matches for these brands, correct the false ones by hand, and add the brand's other names (product names, abbreviations, the domain) so real mentions are not missed either. ### Small samples With 20 answers, one answer is five percentage points of mention rate. A share of voice built on a handful of questions can swing by ten points between two audits for no reason at all. Use at least 20 buyer questions, ask them more than once, and be cautious about reading meaning into differences of a few points. ### Run-to-run variance AI answers change between runs of the same question. A 2025 SparkToro and Gumshoe study, with volunteers running 12 prompts 2,961 times, found [less than a 1 in 100 chance that ChatGPT or Google's AI would return the same list of brands twice](https://sparktoro.com/blog/new-research-ais-are-highly-inconsistent-when-recommending-brands-or-products-marketers-should-take-care-when-tracking-ai-visibility/). The researchers concluded that how often a brand appears across many runs is the meaningful measure, which is what share of voice is built on, but only if there are enough runs. Ask each question at least twice on the chat assistants and report how many question-and-engine pairs gave different results, so readers know how settled the number is. ### Questions that name the brand "Is Brand A any good?" will almost always produce an answer naming Brand A. Including such questions inflates both mention rate and share of voice. Ask them separately to hear what is said about the brand, and keep them out of the scores. ### Brands you did not list The denominator only includes brands in your comparison set. If the assistants keep naming a sixth brand you left out, your share of voice looks better than the real conversation. Check which other brands appear in the answers and decide whether to add them before you report. ### Counting a mention as a recommendation An answer that says "Brand A is often criticised for its pricing" counts as a mention. Share of voice measures presence, not approval. Read the passages around each mention, or at least a sample, and report sentiment alongside the number. ### Comparing audits that are not comparable Different questions, a different competitor list, different engines or a different location all change the number. To compare two audits, keep everything fixed except time. ## How AI Visibility Desk calculates share of voice [AI Visibility Desk](/ai-visibility-report) uses the method above. Each audit asks the buyer questions twice on ChatGPT, twice on Gemini and once on Google, then counts, for the client and up to five listed competitors, the answers that name each brand. Each brand counts once per answer, checks with no answer are left out, and the two branded questions every audit asks are reported separately and never scored. The review screen shows each answer with a short passage so wrong matches can be corrected, lists other brands the assistants named so you can add or hide them, and reports how many question-and-engine pairs disagreed between runs. Share of voice sits in the report next to mention rate per engine, average position, head-to-head results, mention rate by question type and the most cited sources. With monitoring on, each re-run adds a point to the trend. If you run AI visibility work for clients, the [agency page](/for-agencies) explains the white-label report, and the [pricing page](/pricing) lists the plans, including a free audit to try. ## Frequently asked questions ### What is AI share of voice? AI share of voice is a brand's portion of all brand mentions in AI-generated answers to a set of category questions, compared with a defined group of competitors. If answers name your brand 30 times and your competitors 70 times in total, your AI share of voice is 30%. ### How do you calculate AI share of voice? Ask a fixed set of buyer questions on each AI engine, count the answers that name each brand, add those counts across your brand and its listed competitors, then divide each brand's count by the total. Count a brand once per answer, however many times that answer repeats its name. ### What is the difference between share of voice and mention rate? Mention rate is the share of answers that name your brand, measured on its own. Share of voice compares your mentions with competitors' mentions. A brand can appear in 40% of answers yet hold only 20% share of voice if competitors are named even more often. ### How many prompts do you need to measure AI share of voice? Use at least 20 buyer questions, asked more than once on engines whose answers vary. With very few answers, a single answer moves the result by several points. Keep the questions and competitor list fixed between measurements, or the numbers cannot be compared. --- # Answer Engine Optimization (AEO): A Practical Guide URL: https://aivisibilitydesk.com/blog/answer-engine-optimization Updated: 2026-11-20 > Answer engine optimization (AEO) makes your content the answer AI assistants and Google give to buyer questions. What it is, how it works and how to measure it. Answer engine optimization (AEO) is the practice of shaping your content so that search engines and AI assistants use it as the answer to a question, not just as a link beside it. It began with Google's featured snippets and voice assistants, and today it also covers Google AI Overviews, ChatGPT and Gemini. The core of AEO is simple: know the questions your buyers ask, answer each one directly and accurately on a page that engines can read, and back it up so the engine trusts it. ## Why does answer engine optimization matter? Search has been moving towards answers for years. Featured snippets put a paragraph from one website at the top of the results. Voice assistants read out a single answer. AI Overviews and chat assistants now write the answer themselves, often naming products and citing a few sources. Each step reduces the number of clicks that reach websites. In a [Pew Research Center study](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/) of 900 US adults' Google searches in March 2025, people clicked a regular result in 8% of visits when an AI summary appeared, against 15% when it did not. Google disputed the study, but the point for marketers stands: when the answer is on the page, being in the answer matters more than ranking beneath it. The search term has grown accordingly. In DataForSEO's US data for November 2026, "answer engine optimization" averaged 2,400 searches a month, with 3,600 in the latest month, and "aeo agency" went from 260 a year earlier to 1,600. ## What counts as an answer engine? Any system that responds to a question with an answer rather than a list. The ones that matter most for most brands today: | Answer engine | What the answer looks like | Where it gets information | |---|---|---| | Google AI Overviews | A written summary above the results, with linked sources | Google's search index | | Google featured snippets | A quoted passage from one page | Google's search index | | ChatGPT | A conversational answer, often a list of options, sometimes with sources | Its training data and live web search | | Gemini | Similar to ChatGPT, within Google's ecosystem | Its training data and Google Search | | Voice assistants | A short spoken answer | Usually a search engine | ## How do answer engines choose an answer? None of them publish a full recipe, but the public guidance is consistent on the basics. For featured snippets, Google says its systems decide whether a page [would make a good featured snippet](https://developers.google.com/search/docs/appearance/featured-snippets) for a query, and that you cannot mark a page as one. For AI Overviews, Google says a page must be indexed and eligible to show a snippet, and that there are [no additional requirements or special optimizations](https://developers.google.com/search/docs/appearance/ai-features) needed. Assistants that search the web, like ChatGPT, rely on their own crawlers being allowed to read your pages. So AEO is less about tricks and more about being the clearest, most trustworthy answer that the engine can reach. ## How to do answer engine optimization, step by step ### 1. Collect the questions Start with the questions, not the keywords. Good sources: - your sales and support teams ("what do people ask on calls?") - "People also ask" boxes in Google for your main topics - forums and communities in your industry - the questions buyers type into ChatGPT, which tend to be longer and more specific than search queries ("what's the best invoicing tool for a two-person design studio in the UK?") Group them by intent: definitions ("what is"), comparisons ("X vs Y"), recommendations ("best X for Y"), how-tos, and pricing questions. Recommendation and comparison questions are where brands get named or left out, so give them priority. ### 2. Check who answers them today Ask your priority questions on Google, ChatGPT and Gemini. For each, note whether your brand is mentioned, which competitors are, and which pages are cited. This baseline is what an [AEO audit](/aeo-audit) gives you, and it tells you where to focus. A question where a competitor is always named and you never are is a better target than one where nobody is named. ### 3. Write answer-first pages The most consistent AEO pattern is the inverted pyramid: - **Answer in the first two or three sentences.** State the answer plainly, as if it might be read out on its own. That is often exactly what happens. - **Use the question as a heading** where it reads naturally. It helps both readers and engines match the passage to the question. - **Then add depth:** reasons, examples, exceptions, numbers. - **Use lists and tables** for steps, comparisons and specifications. They are easy to extract and easy to read. - **One clear topic per page.** A page that tries to answer twenty loosely related questions answers none of them well. ### 4. Make the answer trustworthy An engine choosing between several pages that say similar things will prefer the one with substance. - Include specifics: prices, features, time frames, named examples. - Cite sources for facts and statistics, and link to them. - Show who wrote it and why they know. Author pages and real expertise help readers judge the content, and they do not hurt with engines either. - Keep pages up to date. A stale answer can be repeated to buyers long after it stopped being true. ### 5. Handle the technical basics - Make sure search and AI crawlers can access your pages. OpenAI, for instance, uses [OAI-SearchBot for ChatGPT search](https://developers.openai.com/api/docs/bots), separate from the crawler it uses for training. - Keep important content in HTML text, not only in images, PDFs or script-rendered widgets. - Use structured data where it describes your content accurately, but do not expect it to do the work on its own. Since August 2023, Google shows FAQ rich results [only for well-known, authoritative government and health sites](https://developers.google.com/search/blog/2023/08/howto-faq-changes). ### 6. Get your answer repeated elsewhere AI assistants weigh what many sources say, not just your own site. If review sites, industry publications and comparison articles describe you the same way your own pages do, that description is more likely to make it into the answer. This is where AEO blends into [generative engine optimization](/blog/generative-engine-optimization): the off-site work of earning mentions where engines look. ### 7. Measure and repeat Rerun the same questions on the same engines after your changes have been crawled. Track mention rate per engine, position in the answer and which pages are cited. Because AI answers vary from run to run, ask each question more than once and look at the pattern, not at any single answer. ## What does an answer-first passage look like? Take the question "How much does bookkeeping software cost for a freelancer?" A weak page opens with the history of bookkeeping and gets to prices in the fifth paragraph. An answer-first page opens like this: > Most bookkeeping software for freelancers costs between $10 and $50 a month (illustrative figures), and several tools offer a free tier with limits on invoices or clients. The main things that raise the price are payroll, multiple currencies and the number of users. Then it backs that up with a table of named tools and their current prices, with links to each pricing page and the date it was checked. The opening is short enough to be quoted whole, specific enough to be useful, and the support below it gives an engine a reason to trust it. On a real page, the bracketed range would be filled with current, sourced prices from your own research. ## Common AEO mistakes - **Burying the answer.** A 300-word introduction before the answer gives the engine nothing to quote. - **Writing for keywords instead of questions.** Repeating a phrase does not make a page a better answer. - **Only checking Google.** Your brand can appear in AI Overviews and be absent from ChatGPT, or the reverse. Check each engine on its own. - **Ignoring branded questions.** "Is [your brand] good for X?" and "[your brand] pricing" are questions too. Make sure your own site gives the clearest answer to them. - **Treating schema as a shortcut.** Markup describes content; it does not replace it. ## AEO vs GEO vs SEO AEO, GEO and SEO share most of their tactics. SEO gets your pages found and ranked. AEO shapes those pages into direct answers. GEO extends the effort to being cited and named across generated answers, which also depends on your reputation off-site. If you want a side-by-side view and advice on what to do first, read [GEO vs AEO vs SEO](/blog/geo-vs-aeo-vs-seo). ## Run an AEO audit [AI Visibility Desk](/aeo-audit) measures the answer side of AEO. It asks up to 30 buyer questions on ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity and Claude, then shows which answers name the brand, where in the answer, which competitors appear instead, which pages are cited, the question types where the brand is missing, and what the answers say about it. Agencies publish the result as a report under their own branding. See [pricing](/pricing) for plans. ## Frequently asked questions ### What is answer engine optimization? Answer engine optimization (AEO) is the practice of structuring content so that search engines and AI assistants use it as the direct answer to a question. It started with featured snippets and voice search and now also covers Google AI Overviews, ChatGPT and Gemini, where the answer is written for the user instead of shown as links. ### What is the difference between AEO and GEO? They overlap almost entirely in practice. AEO focuses on being the answer to a specific question, which favours clear, direct, well-structured pages. GEO focuses on being included and cited in a generated answer drawn from many sources, which also depends on brand mentions across the web. Most teams treat them as one effort. ### Does FAQ schema still help with AEO? Less than it used to. Since August 2023 Google shows FAQ rich results only for well-known, authoritative government and health sites, and it says no special schema is needed for AI Overviews. Clear question-and-answer content on the page still helps; the markup alone does not make you the answer. ### How do you measure answer engine optimization? Pick the questions you want to be the answer to, ask them on ChatGPT, Gemini and Google, and record whether your brand is named, where it appears, and whether your pages are cited. Repeat each question, since answers vary between runs, and rerun the same set later to see whether your changes worked. --- # Best AI Visibility Tools in November 2026: An Honest Comparison URL: https://aivisibilitydesk.com/blog/best-ai-visibility-tools Updated: 2026-11-19 > Peec AI and Otterly suit daily tracking, Profound suits enterprise, Rankscale and AI Visibility Desk suit white-label agencies. Prices checked 19 Nov 2026. If you are an enterprise brand with budget and a security review to pass, start with Profound. If you want daily prompt tracking at a mid-range price, look at Peec AI or Otterly.AI, and if you already pay for Semrush, Ahrefs or SE Ranking, try their AI add-on first. If you are an agency that needs branded client reports, compare Rankscale and AI Visibility Desk (our product), which both offer white-label output. ## How we compared them We read each vendor's own pricing, feature and agency pages on 19 November 2026 and wrote down what they say about engines covered, core features, agency and white-label options, and price. Prices change often in this market, so treat every figure below as "as seen on 19 November 2026" and check the vendor's page before you buy. Where a vendor's own pages disagree with each other, or a figure could not be confirmed, we say so instead of picking one. We did not run paid trials of every tool, and we did not score them on a single number. Each tool counts its usage differently (prompts, credits, checks, audits), so a straight price-per-prompt comparison would mislead. One disclosure up front: AI Visibility Desk is our product. We include it because leaving it out of a list on our own site would be odd, and we label it clearly. We have tried to be as plain about where other tools beat us as about where we think we fit. ## At a glance | Tool | Best for | Engines (as listed) | Entry price | White-label | |---|---|---|---|---| | AI Visibility Desk (ours) | Agencies sending client reports | 6 | Free, then $49/mo | Every plan | | Profound | Enterprise brands | Up to 10 | Custom only | Not mentioned | | Peec AI | Brands and agencies wanting daily tracking | 3 of 8 core, up to 26 on Enterprise | $95/mo | Not mentioned | | Otterly.AI | Budget daily tracking, broad free tools | 7 | €29/mo | Looker Studio branding | | AthenaHQ | Teams wanting many engines | 5 free, 11 on Starter | Free, then $295/mo | Not mentioned | | Scrunch | Brands wanting an action layer | 7 to 8 | $300/mo | Not mentioned | | Rankscale | Agencies wanting white-label dashboards | 17+ | €99/mo | Growth plan and up | | LLMrefs | Keyword-based tracking on one plan | 11 | $79/mo | Not mentioned | | Semrush AI Visibility Toolkit | Existing Semrush users | 2 tracked (more in checker) | $99/mo per user | Not stated for AI reports | | Ahrefs Brand Radar | Existing Ahrefs users, market-level data | 6, plus Claude | Needs a paid Ahrefs plan | Not mentioned | | SE Ranking | Existing SE Ranking users, agencies | 5, plus Claude via API | Included in Core $129/mo | Agency Pack, $69/mo | | Writesonic | Teams that also want content tools | 3, up to 8 on Enterprise | $79/mo (annual) | Agency Enterprise only | ## AI Visibility Desk (our product) **Who it is for:** agencies and consultants who need to send a client a finished, branded AI visibility report, and in-house marketers who want a full audit rather than a dashboard. **Engines:** ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity and Claude. **What it does:** each audit asks your buyer questions on all six engines, repeats them, and reports mention rate, position, share of voice, head-to-head results per competitor, citations and cited pages, and sentiment. Every audit also includes a built-in SEO, GEO and AEO site audit (including llms.txt and AI crawler checks). You can re-run audits weekly or monthly and see the trend. Reports are white-label on every plan. Paid plans include free pitch audits for prospects that do not use your plan allowance: 5 a month on Starter, 20 on Agency and 50 on Studio, each covering up to 5 questions on ChatGPT and AI Overviews. **Pricing:** Free, then $49, $149 and $349 a month, or yearly with two months free. There is no per-seat pricing: Agency includes a team of 5 people and Studio 20, with roles. **Where rivals beat us:** we do not offer daily tracking (our shortest schedule is weekly). We do not cover Copilot, Grok, DeepSeek or Meta AI. We have no API or MCP server, no client logins (clients get a private report link), no agent analytics from server logs (we check robots.txt and the homepage), and no custom report domain, so report links show our domain. If any of those is a must-have, look at the tools below. ## Profound **Who it is for:** large brands with an enterprise budget and procurement process. **Engines:** up to 10: ChatGPT, Perplexity, Google AI Mode, Gemini, DeepSeek, Claude, Google AI Overviews, Exa Search, Muse Spark and Qwen. **Standout features:** prompt tracking with query fan-out, sentiment, citations and fact-checking; Prompt Volumes for search volume; Agent Analytics that reads AI crawler logs through Cloudflare, Vercel, Akamai and others; shopping tracking; agent-based tools; MCP, API, SSO/SAML and SOC 2. **Pricing as seen:** Enterprise only, custom. Profound's pricing page says its Starter and Growth plans were retired for brands in September 2026. **Watch-outs:** no public price, no free tool listed, and white-label is not mentioned on its agencies page. [Profound pricing](https://www.tryprofound.com/pricing). Our comparison: [AI Visibility Desk vs Profound](/compare/profound). ## Peec AI **Who it is for:** brands and agencies that want daily tracking, MCP and API access at a published price. **Engines:** on Starter to Advanced you pick 3 of 8 core models (ChatGPT, Perplexity, AI Overviews, AI Mode, Gemini, Naver AI Briefing, ERNIE, Bing AI Mode). Enterprise adds API models including Claude, Grok and DeepSeek, up to 26. Extra models cost $30 to $140 per model per month depending on plan. **Standout features:** visibility, position, sentiment and sources; agent analytics from CDN logs; recommendations; MCP on all plans, API from Pro, Looker Studio from Advanced; daily tracking; unlimited users; SOC 2 Type II. Agency plans include free 7-day pitch projects and view-only client dashboards. **Pricing as seen:** Starter $95/mo (50 prompts), Pro $245/mo (150), Advanced $495/mo (350), Enterprise custom, 15% off annual. Agency plans: $245, $495 and $795 a month. **Watch-outs:** only 3 engines on self-serve plans unless you pay per extra model. White-label is not mentioned on the pricing page. [Peec AI pricing](https://peec.ai/pricing). Our comparison: [AI Visibility Desk vs Peec AI](/compare/peec-ai). ## Otterly.AI **Who it is for:** small teams that want daily tracking at a low entry price, and anyone who likes free tools. **Engines:** ChatGPT, Google AI Overviews, Perplexity, Microsoft Copilot, Google AI Mode, Gemini and Claude (Claude costs 3 credits per answer). The plan cards say 4 of 7 engines while the comparison table on the same page says every plan tracks every engine, so check which applies before you sign up. **Standout features:** sentiment, citation tracking, agent analytics, a GEO audit, prompt research, recommendations, API, MCP and a Data Studio connector from Standard, unlimited seats. Agency partners get pitch workspaces and Looker Studio reports with their branding. A long list of free tools. **Pricing as seen:** Lite €29/mo (16 prompts), Standard €189/mo (120), Premium €489/mo (403), Enterprise custom, 15% off annual. **Watch-outs:** 16 prompts on Lite is a small set for a real audit. [Otterly.AI pricing](https://otterly.ai/pricing). Our comparison: [AI Visibility Desk vs Otterly.AI](/compare/otterly-ai). ## AthenaHQ **Who it is for:** teams that want the widest engine list on a self-serve plan, and anyone who wants to try a free tier first. **Engines:** the free tier covers ChatGPT, Perplexity, AI Overviews, Gemini and Copilot. Starter covers 11: those plus AI Mode, Claude, Grok, DeepSeek, Meta AI and Mistral. **Standout features:** prompt and response analysis, competitor insights, on-page and off-page actions, a content optimisation agent, Shopify, Webflow, GA4 and GSC integrations. API is a paid add-on. Agencies get discounted licences and pitch-ready workspaces. **Pricing as seen:** Essentials free (300 credits), Starter $295/mo (3,600 credits, one credit per AI response), Enterprise custom. **Watch-outs:** credit-based pricing makes cost depend on how many engines and repeats you run. White-label is not mentioned. [AthenaHQ pricing](https://www.athenahq.ai/pricing). Our comparison: [AI Visibility Desk vs AthenaHQ](/compare/athenahq). ## Scrunch **Who it is for:** brands that want monitoring plus tools to change what AI agents see. **Engines:** ChatGPT, Claude, Gemini, Perplexity, Google AI Mode, AI Overviews and Meta on the pricing page; the homepage also names Copilot. **Standout features:** monitoring by prompt, topic, persona and region; crawl-error detection; AXP, a machine-readable version of your site served to AI agents; page audits; alerts and scheduled reports; Data API on Enterprise; SOC 2 Type II. **Pricing as seen:** Starter $300/mo ($250 annual) with 350 prompts and 3 seats; Growth $500/mo ($417 annual) with 700 prompts and 5 seats; extra seats $25/mo; 7-day trial. **Watch-outs:** seat limits, unlike most rivals. Agency pricing is not published and white-label is not mentioned. [Scrunch pricing](https://scrunch.com/pricing). Our comparison: [AI Visibility Desk vs Scrunch](/compare/scrunch-ai). ## Rankscale **Who it is for:** agencies that want white-label dashboards and links plus a long engine list. **Engines:** "17+": ChatGPT, Google AI Mode, AI Overviews, Bing AI Mode and Gemini in the app, plus about 15 API models. **Standout features:** rank tracking per engine, sentiment, citations, shopping and ads, query fan-outs, page audits, prompt research, GA4/GSC in beta, MCP, REST API from Pro+. White-label dashboards, reports and links from Growth. **Pricing as seen:** Pro €99/mo, Pro+ €229/mo, Growth €385/mo, Enterprise €780/mo, yearly about 15% less. Unlimited prompts and seats, limited by credits. **Watch-outs:** white-label starts at €385 a month. No free tool linked. [Rankscale pricing](https://rankscale.ai/pricing). Our comparison: [AI Visibility Desk vs Rankscale](/compare/rankscale). ## LLMrefs **Who it is for:** SEO-minded teams that think in keywords rather than prompts. **Engines:** ChatGPT, ChatGPT Search, AI Overviews, AI Mode, Gemini, Perplexity, Claude, Grok, Copilot, Meta AI and DeepSeek. **Standout features:** you track keywords and it generates fan-out prompts from each one; share of voice and average position per keyword; citations, content gaps, prompt volume estimates; weekly reports, CSV, API, share links, 20+ countries; unlimited projects and seats. A good set of free tools. **Pricing as seen:** one paid plan at $79/mo, marked "limited time only", with 500 prompts and a 7-day trial, plus a free account with unpublished limits. **Watch-outs:** reports are weekly, the price may change, and white-label is not mentioned. [LLMrefs pricing](https://llmrefs.com/pricing). Our comparison: [AI Visibility Desk vs LLMrefs](/compare/llmrefs). ## Semrush AI Visibility Toolkit **Who it is for:** teams already working in Semrush. **Engines:** Prompt Tracking and Brand Performance name ChatGPT and Google AI Mode. The free checker covers ChatGPT, Gemini, AI Mode and AI Overviews. Enterprise AIO, now sold through Adobe, lists ChatGPT, Perplexity, Claude, Gemini, Grok "and more". **Standout features:** AI Visibility Score, a large prompt research database, share of voice, sentiment and narratives, prompt tracking, and AI search checks inside Site Audit. **Pricing as seen:** $99/mo per user with 25 tracked prompts and 1 project; 50 more prompts cost $60/mo. Semrush One bundles start at $199.95/mo according to Semrush's knowledge base, while a Semrush blog post says $199, so the figures conflict. Enterprise AIO is custom. **Watch-outs:** per-user pricing and a small prompt allowance. The toolkit page does not say AI reports are white-label. [Semrush AI toolkit documentation](https://www.semrush.com/kb/1493-ai-toolkit). Our comparison: [AI Visibility Desk vs Semrush AI Toolkit](/compare/semrush-ai-toolkit). ## Ahrefs Brand Radar **Who it is for:** Ahrefs users who want market-level AI visibility data without writing their own prompts. **Engines:** AI Overviews, AI Mode, ChatGPT, Perplexity, Gemini and Copilot; Claude is available at 8 checks per update. **Standout features:** prompts drawn from a database Ahrefs puts at "581 million+" real queries; mentions, citations, share of voice and estimated impressions; custom prompt tracking with daily, weekly or monthly refresh; fan-out discovery. **Pricing as seen:** needs a paid Ahrefs plan (Lite is $129/mo). The Brand Radar AI add-on is "from $199/mo". Custom-prompt packages run from $50 (2,500 checks) to $250 (25,000). The Brand Radar page also shows "$699/mo for all models" without saying which tier that is. **Watch-outs:** pricing has several layers. White-label is not mentioned. [Ahrefs Brand Radar](https://ahrefs.com/brand-radar). Our comparison: [AI Visibility Desk vs Ahrefs Brand Radar](/compare/ahrefs-brand-radar). ## SE Ranking **Who it is for:** SE Ranking customers and agencies that want AI tracking inside an SEO suite with white-label. **Engines:** ChatGPT, Copilot, AI Overviews, AI Mode and Gemini on every plan; Claude via API (without web search) on the Agency plan. Perplexity is not listed. **Standout features:** 25 daily prompts included in every SE Ranking plan, competitor research, AI Overview overlap with organic top 20, Data Studio connector, API, MCP, AgencyAnalytics and Whatagraph connections. **Pricing as seen:** Core $129/mo, Growth $279/mo. The AI visibility add-on starts at $71/mo; prices for its larger tiers are not published on that page. The Agency Pack is $69/mo, annual only, with 30 client seats and a white-label platform. **Watch-outs:** no Perplexity, and the add-on pricing is not fully published. [SE Ranking AI visibility tracker](https://seranking.com/ai-visibility-tracker.html). Our comparison: [AI Visibility Desk vs SE Ranking](/compare/se-ranking). ## Writesonic **Who it is for:** teams that want AI visibility tracking next to content writing and outreach tools. **Engines:** Starter to Growth track ChatGPT, Gemini and AI Overviews. Enterprise tracks 8, adding Perplexity, Claude, Grok, AI Mode and Muse Spark. **Standout features:** citation gap analysis, fact check, an action center, AI articles, site audits, agent analytics, ads and shopping tracking, an outreach agent, MCP. The agency Starter plan includes 20 pitch projects a month. **Pricing as seen:** Starter $79, Basic $199, Growth $399 a month on annual billing; the page's meta description says "from $99". Agency Starter is $200/mo. White-label reporting is only on the custom agency Enterprise plan. **Watch-outs:** only 3 engines below Enterprise. [Writesonic pricing](https://writesonic.com/pricing). ## How to choose ### Agency that needs white-label reports Decide whether clients need a live dashboard or a finished report. Rankscale gives white-label dashboards and links from €385 a month. SE Ranking's $69 Agency Pack makes sense if you already use SE Ranking. AI Visibility Desk includes white-label reports on every plan from $49 a month, adds a site audit and free pitch audits, but does not offer client logins or a custom domain. Peec AI and Writesonic have strong agency plans with pitch projects, though white-label is limited or missing below enterprise tiers. ### In-house enterprise team Profound and Scrunch are built for this buyer, with SOC 2, alerts, log-based agent analytics (Profound) and an action layer. Peec AI's Enterprise plan is the option if you need the widest engine list. ### SEO team already on Semrush or Ahrefs Start with the add-on you can switch on today. Semrush's toolkit is cheap to try but tracks few prompts per user. Ahrefs Brand Radar is strong for market-level data you did not have to write prompts for. Move to a dedicated tool if you hit their engine or prompt limits. ### Tight budget Otterly.AI Lite (€29), LLMrefs ($79) and AthenaHQ's free tier are the cheapest ways into tracking. AI Visibility Desk has a free plan, and its $49 plan covers six engines with a site audit. Check how many prompts or credits each plan really includes before comparing prices. ## Try before you buy You can test the basics without signing up for anything. Run our free [AI visibility checker](/ai-visibility-checker) to see whether assistants name your brand, the [SEO checker](/seo-checker) for on-page basics, the [AI crawler checker](/ai-crawler-checker) to see whether GPTBot, ClaudeBot and others can reach your site, and the [llms.txt generator](/llms-txt-generator) to create a file for AI tools. Most vendors above also offer a trial, a free tier or a free checker, so use them on the same set of questions and compare the answers side by side. ## Frequently asked questions ### What is the best AI visibility tool? It depends on who is buying. Enterprise brands with a large budget tend to look at Profound. Teams that want daily prompt tracking at a mid-range price look at Peec AI or Otterly.AI. Agencies that need branded client reports look at Rankscale (white-label from its Growth plan) or AI Visibility Desk (white-label on every plan). Teams already paying for Semrush, Ahrefs or SE Ranking can start with that suite's add-on. ### How much do AI visibility tools cost? On 19 November 2026, entry prices ranged from free tiers (AthenaHQ, AI Visibility Desk) and about €29 a month (Otterly.AI Lite) to $295 to $300 a month (AthenaHQ Starter, Scrunch Starter). Profound only sells custom Enterprise plans. Most tools charge by the number of prompts or credits, so compare what one plan actually covers, not only the headline price. ### Which AI visibility tools offer white-label reports? On the vendors' pages as read on 19 November 2026: AI Visibility Desk includes white-label reports on every plan, Rankscale includes white-label dashboards, reports and links from its Growth plan, SE Ranking sells a $69 a month Agency Pack with a white-label platform, and Writesonic offers white-label reporting on its custom agency Enterprise plan. Otterly.AI offers Looker Studio reports with your branding. ### Do I need daily tracking of AI answers? Not always. Daily tracking helps when you are testing content changes and want to see movement quickly. For client reporting and quarterly reviews, weekly or monthly runs that repeat each question several times are often enough, because single AI answers vary from one run to the next anyway. ### Which AI engines should a visibility tool cover? At minimum ChatGPT, Google AI Overviews and Gemini, since those have the largest audiences. Google AI Mode, Perplexity and Claude are worth adding for most B2B and tech categories. Copilot, Grok, DeepSeek and Meta AI matter for some audiences and are covered by tools such as AthenaHQ, LLMrefs and Otterly.AI. --- # Generative Engine Optimization (GEO): A Practical Guide URL: https://aivisibilitydesk.com/blog/generative-engine-optimization Updated: 2026-11-19 > Generative engine optimization (GEO) helps your brand get named and cited in ChatGPT, Gemini and AI Overviews. A practical guide: what works and how to measure. Generative engine optimization (GEO) is the work of making your brand more likely to be named, summarised and cited when AI systems like ChatGPT, Gemini and Google's AI Overviews write an answer. It builds on SEO but aims at a different result: a mention inside the answer rather than a ranked link below it. In practice it comes down to four things: let the engines read your site, state clearly what you are, publish content that is easy to quote, and be present on the third-party sites the engines trust. ## Where does the term GEO come from? The term was popularised by a 2023 paper, [GEO: Generative Engine Optimization](https://arxiv.org/abs/2311.09735), by researchers at Princeton, IIT Delhi and independent collaborators, later accepted at KDD 2024. They built a benchmark of 10,000 queries and tested how rewriting a source page changed its visibility in AI-generated answers. Their findings are a useful starting point, with caveats. The best methods (adding quotations, adding statistics and citing sources) raised visibility by up to about 40% on their measures. Keyword stuffing, the reflex of old-school SEO, gave little or no improvement and sometimes made things worse. Citing sources helped lower-ranked pages the most: in one test, a page in fifth place gained about 115% in visibility, while the top-ranked page lost some. The caveat: this was a controlled experiment on a research setup, not a field study of ChatGPT or Gemini in production. Treat the direction as useful and the exact percentages as specific to that study. Search demand for the term has grown too. In DataForSEO's US data for November 2026, "generative engine optimization" averaged 4,400 searches a month, with 6,600 in the latest month. ## How do generative engines decide what to say? You do not need the internals to do GEO well, but a rough model helps. - **Training data.** The model has absorbed a large amount of text. If your brand was widely discussed before its training cutoff, it may "know" you without searching. - **Live retrieval.** For many questions, especially about products, the assistant searches the web, reads a handful of pages and writes an answer from them. Google's AI Overviews draw on Google's own index. ChatGPT uses its own search crawler, among other sources. - **Synthesis.** The model combines what it found into a short answer, usually naming a few options and sometimes citing where the information came from. GEO therefore has two horizons. Retrieval can respond within weeks if your pages are accessible and useful. Training data moves slowly and depends on how widely your brand is discussed across the web. ## What does a GEO programme involve? ### Step 1: Measure where you stand Before changing anything, take a baseline. Write 15 to 30 questions a buyer would ask without knowing your brand: "best X for Y", "X vs Y", "alternatives to X", "how to choose X". Ask each one on ChatGPT, Gemini and Google, ideally more than once, and record: - whether your brand is named (mention rate) - where it appears in the answer - which competitors are named - which websites are cited - how your brand is described This is a [GEO audit](/geo-audit). Keep the question set; you will reuse it to measure progress. Our guide on [how to run an AI visibility audit](/blog/how-to-run-an-ai-visibility-audit) walks through the details. ### Step 2: Make sure the engines can read you This is the least glamorous step and often the one with the quickest payoff. - **Crawler access.** Check robots.txt and any CDN or firewall rules. OpenAI documents that OAI-SearchBot is used to [surface websites in ChatGPT's search features](https://developers.openai.com/api/docs/bots), separate from GPTBot, which crawls for model training. You can allow one and block the other. Many sites block both by accident through a blanket bot rule. - **Google eligibility.** Google says a page needs to be [indexed and eligible to show a snippet](https://developers.google.com/search/docs/appearance/ai-features) to appear in AI Overviews, and that no special schema or AI-specific file is required. - **Content in text.** Key facts (what you do, who it is for, pricing model, locations) should be in plain HTML text, not only in images, PDFs or content that appears after scripts run. ### Step 3: Make your positioning unambiguous Generative engines summarise. If your site never says plainly what you are, the engine will borrow a description from somewhere else, or skip you. - Put a one-sentence description of what you sell and for whom near the top of the homepage. - Use the same wording on your About page, social profiles, directory listings and press materials. Consistency makes the description easier to pick up. - If you serve a specific niche, say so. "Accounting software for freelance designers" gives an engine a reason to name you for a narrow question where a big brand would otherwise win. ### Step 4: Publish content that is easy to quote Write pages that answer the questions in your audit and structure them so a model can lift a clear passage. - **Lead with the answer.** The first two or three sentences should answer the question directly. Context and nuance can follow. - **Use specifics.** Numbers, named features, prices, time frames and concrete examples. The GEO paper's best methods were the ones that added this kind of substance. - **Cite your sources.** When you state a fact, link to where it comes from. - **Cover comparisons honestly.** "X vs Y" and "alternatives to X" are among the most common buyer questions. If you do not publish a fair comparison, a competitor or an affiliate site will. - **Keep it current.** Out-of-date pricing or features on your own site can be repeated back to buyers. ### Step 5: Earn mentions where engines look For many categories this matters more than anything on your own site. Your baseline audit shows which domains the engines cite when your category comes up. Typical patterns: - review and comparison sites in your industry - directories and marketplaces - "best X" listicles from publishers - forums and community threads - Wikipedia and other reference sites If the engines keep citing a comparison article that lists five competitors and not you, getting into that article (by contacting the author, providing a demo, or offering updated information) is a concrete GEO task. Digital PR, review generation and partnerships all belong here. ### Step 6: Measure again Rerun the same questions after your changes have had time to be crawled. Compare mention rate per engine, position, share of voice and cited sources against the baseline. Expect noise: AI answers vary between runs, which is why repeated runs and a decent number of questions matter more than any single answer. ## Which GEO tactics are overrated? A few ideas circulate widely with little evidence behind them. | Tactic | Our view | |---|---| | llms.txt files | Google says no special AI file is needed for AI Overviews. Harmless, but do not expect it to change results on its own. | | Special "AI schema" | Structured data helps search engines understand pages, but Google states there is no special schema required for its AI features. | | Keyword stuffing for AI | The GEO paper found it gave little to no improvement. | | Mass-producing AI-written pages | Thin, repetitive pages give an engine nothing new to cite. | | Prompt injection in hidden text | Manipulative and risky. Do not do it. | ## How is GEO different from AEO and SEO? The terms overlap heavily. [Answer engine optimization](/blog/answer-engine-optimization) (AEO) grew out of featured snippets and voice search and focuses on being the direct answer to a question. GEO focuses on being included in a generated, multi-source answer. SEO is the foundation both rely on. We compare them side by side in [GEO vs AEO vs SEO](/blog/geo-vs-aeo-vs-seo). ## What should you report to a client or your boss? Keep it to numbers you can defend: - mention rate per engine, from a fixed question set - typical position in the answer - share of voice against named competitors - how often your own site is cited, and the top cited domains - what the engines say about the brand, with example passages - a short list of actions tied to what the data showed Avoid promising rankings or placements. Promise measurement, a clear plan and an honest rerun. ## Run a GEO audit with AI Visibility Desk AI Visibility Desk runs the measuring part of GEO for you. You enter the brand, competitors and up to 30 buyer questions (or let it suggest them from the client's website). It asks each question twice on ChatGPT, twice on Gemini and once on Google, then reports mention rate per engine, position, citations, share of voice, unlisted competitors, sentiment and whether the site lets AI crawlers in. Agencies publish the result under their own brand at a private link. See the [GEO audit](/geo-audit) page for what is included, or [pricing](/pricing) to get started. ## Frequently asked questions ### What is generative engine optimization? Generative engine optimization (GEO) is the practice of making a brand and its content more likely to be mentioned, summarised or cited by AI systems that write answers, such as ChatGPT, Gemini and Google AI Overviews. It covers site access, clear positioning, quotable content and presence on the third-party sites those systems rely on. ### Does GEO replace SEO? No. Many AI answers are built from pages a search engine has already found and ranked, so technical SEO and good rankings remain the foundation. GEO adds a different goal on top: being named inside the answer itself, not just ranking as a link below it. Most teams run them as one programme. ### How long does generative engine optimization take to work? It varies. Fixing crawler access can matter as soon as the engines recrawl your pages. Earning mentions on review sites and in comparison articles, which often moves results the most, usually takes months. Measure a fixed set of questions before you start and rerun the same set later to see the change. ### Can you guarantee a brand appears in ChatGPT? No. Nobody outside the AI companies controls what a model says, and answers vary from one run to the next. Be cautious with anyone promising placement. What you can do is improve the odds and measure the result honestly across many questions and repeated runs. --- # GEO Agency Guide: How to Package and Sell AEO Services URL: https://aivisibilitydesk.com/blog/sell-geo-aeo-services-agency Updated: 2026-11-20 > How a GEO or AEO agency can package AI visibility work: the audit as the entry offer, what goes in a proposal, how to price it, and how to report results. The simplest way for an agency to sell GEO or AEO services is to start with a fixed-price AI visibility audit, then sell the fixes it uncovers as a project or retainer. The audit shows a client, in their own category's questions, how often ChatGPT, Gemini and Google's AI Overviews name them against competitors, and that evidence does most of the selling. Price the follow-on work like the SEO and PR work it largely is, and report mention rate across many questions rather than promising positions. ## Is there real demand for a GEO or AEO agency? Yes, and it is growing from a small base. In DataForSEO's US data for November 2026, "geo agency" and "aeo agency" each averaged 720 searches a month. In the latest month "geo agency" reached 880 (from 320 a year earlier) and "aeo agency" reached 1,600 (from 260). "ai visibility audit" went from 70 to 1,600 over the same year. The client side of that demand is easy to explain. A [Pew Research Center study](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/) of 900 US adults found Google users clicked a regular result in 8% of visits when an AI summary appeared, against 15% without one. Marketing leads see traffic flatten and want to know whether they are in the answers that replaced the clicks. Most do not know how to check. You do not need to rebrand as a "GEO agency" to sell this. Existing SEO and content agencies already have most of the skills. What they usually lack is a way to measure AI answers and a deliverable that makes the problem visible. ## What are clients actually buying? Strip away the labels and a client wants three things: 1. **To know where they stand.** Do assistants mention them for the questions their buyers ask? Who gets named instead? 2. **To know what to do about it.** A short list of changes with reasons, not a 60-page deck. 3. **To see whether it worked.** The same measurement, repeated after the work is done. GEO, AEO and LLM SEO are different names for overlapping work. Sell the outcome, and use whichever term your client already uses. ## Why lead with an audit? An audit is the natural entry offer for AI visibility work, for four reasons. - **It answers a question the client already has.** "Do AI assistants mention us?" is the question behind most inbound interest. - **The evidence persuades.** A table showing a competitor named in 70% of answers and the client in 15% does more than any pitch. - **It scopes the next project.** The gaps it finds (missing questions, sites cited instead of the client, crawler blocks, negative descriptions) become the work plan. - **It is low risk to buy.** A fixed fee for a defined deliverable is easier to approve than an open-ended retainer for a new discipline. Use it as a paid diagnostic, or run a shorter version free as part of a pitch. Either way, the full audit should be a real deliverable, not a teaser. ### What the audit should cover A useful [GEO audit](/geo-audit) or AEO audit measures: - **Mention rate per engine** across 15 to 30 buyer questions that do not contain the brand name, each asked more than once on engines whose answers vary. - **Position in the answer**: named first, in the middle, or as an afterthought. - **Competitors and share of voice**: who is named instead, and the client's share of all brand mentions. - **Citations**: which websites the assistants cite, and how often the client's own site is one of them. - **Sentiment and themes**: what the answers praise or criticise. - **Technical access**: whether robots.txt or a firewall blocks the crawlers that feed AI search. - **Branded questions**, kept separate: how the assistant describes the brand when asked directly. Our [step-by-step audit guide](/blog/how-to-run-an-ai-visibility-audit) covers the method in detail. ## How should you price GEO and AEO services? There is no reliable industry benchmark yet. Most published figures come from agencies pricing their own services, so treat every range below as a rough indication. | Source | What it says | Caveat | |---|---|---| | [First Page Sage](https://firstpagesage.com/reports/geo-aeo-pricing-statistics/) | Median GEO/AEO retainer of $6,850 a month; traditional SEO median $4,740 | Published by an agency that sells GEO | | [WebFX](https://www.webfx.com/blog/seo/geo-pricing/) | Agency retainers $1,500 to $50,000+ a month; projects $5,000 to $50,000; AI visibility tracking software $50 to $1,000+ a month | Very wide; agency-published | | [Digital Agency Network](https://digitalagencynetwork.com/geo-pricing/) | One-time GEO audits $2,000 to $8,000; add-on to an SEO retainer +$1,500 to $3,000 for small and mid-sized clients | No methodology given for the figures | A few patterns hold across these sources and are worth borrowing: - **Audits are priced as fixed-fee projects.** A few thousand dollars is a common quote for small and mid-sized clients; smaller agencies and local clients will sit below that. - **Ongoing work is priced as a retainer or an add-on.** Many SEO agencies add AI visibility to an existing SEO retainer rather than selling it separately, which is easier for existing clients to accept. - **Implementation costs more than advice.** Content production, digital PR and technical work are where the hours go. ### A simple three-step package One structure that maps onto the work: | Offer | What it includes | How to price | |---|---|---| | AI visibility audit | Question set, measurement on each engine, competitor and citation analysis, branded report, 30-minute walkthrough, prioritised fix list | Fixed fee | | Fix sprint | Crawler and indexing fixes, new or rewritten pages for missing questions, outreach to the sites that get cited | Fixed fee or monthly, 2 to 3 months | | Ongoing | Content and PR against the fix list, with a fresh audit each quarter to compare | Monthly retainer or add-on to SEO | Work out your cost per audit before you set a price: your tool subscription divided by audits per month, plus the analyst time to choose questions, check matches, write the summary and present it. The analyst time is usually the larger part, and it is what the client is paying for. ## What should go in a GEO or AEO proposal? Keep it short and specific to the client. A proposal that works usually has: 1. **The problem in their terms.** Three or four of their buyers' questions, and what the assistants answered. If you have run a free sample, include it. 2. **What you will measure.** The engines, the number of questions, how many times each is asked, and what counts as a mention. 3. **What you will deliver.** The report, the fix list, the walkthrough, and any implementation. 4. **What you will not promise.** Positions in ChatGPT, guaranteed citations, or results on engines you do not measure. 5. **How success is judged.** Mention rate and share of voice on the same question set, measured again after a stated period. 6. **Price and terms.** Fixed fee for the audit; monthly figure for anything ongoing. The "will not promise" section protects you. A [SparkToro study with Gumshoe.ai](https://searchengineland.com/ai-recommendation-lists-rarely-repeat-study-468076) found AI tools returned the same brand list less than 1% of the time across repeated runs. Promising "#1 in ChatGPT" sets up an argument you cannot win. Promising a measured improvement in how often the brand appears across a fixed set of questions is defensible. ## How should you report AI visibility results? Clients read the first page and skim the rest, so put the answer up top. - **Lead with one number per engine.** "ChatGPT named you in 18% of answers; your closest competitor was named in 64%." - **Show the competitor comparison next.** Share of voice in a single chart or table. - **Show the sources.** Which sites the assistants cited. These are your PR and content targets. - **Quote the answers.** A short passage from a real answer convinces more than a percentage. - **Be clear about what is a reading.** Sentiment is a model's interpretation of the text, not a hard measurement. Say so. - **Say what is missing.** If some checks failed or an engine did not answer, report how many. - **End with the fix list**, ranked by likely impact and effort. ## Where do white-label reports fit? Clients buy from the agency, not from the agency's tools. A white-label report lets you hand over a finished deliverable under your own name, logo and colours, without explaining which software produced it or building reports by hand. That matters most at the audit stage, where the report is the product. It also saves time. Assembling screenshots of AI answers into a slide deck for every client is slow and hard to repeat consistently; a generated report you check and annotate is faster and the same every time. What to check in any white-label setup: what the client actually sees (branding, links, any mention of the vendor), whether you can edit the summary and recommendations before publishing, whether you can correct wrong brand matches, and whether you can export the raw answers if the client asks how a number was reached. ## Mistakes to avoid - **Overclaiming coverage.** If you measure three engines, say three. Do not imply every AI assistant is covered. - **Measuring once.** A single run per question is noise. Ask more than once and report rates. - **Using branded questions as the score.** "What is [brand]?" will always name the brand. Keep those separate. - **Selling tactics without evidence.** llms.txt files and "AI schema" are cheap to add but have no evidence of impact, and Google says its Search ignores llms.txt ([Google's AI optimization guide](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide)). See [LLM SEO: evidence vs guesswork](/blog/llm-seo). - **Skipping the re-audit.** Without a second measurement on the same questions, you cannot show the client what changed. ## Running audits under your brand with AI Visibility Desk AI Visibility Desk was built for this entry offer. You add your agency's name, logo, colour and contact line once, enter a client and its competitors, and choose up to 30 buyer questions (or have it read the client's site and suggest them). It asks each question on ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity and Claude, then gives you mention rates per engine, position, share of voice, cited sites, sentiment, crawler checks and every answer to review. You can draft and edit the summary, then publish the report under your brand at a private link, print it to PDF, or download the answers as CSV. Paid plans add weekly or monthly monitoring for the retainer, pitch audits for prospects that don't use the allowance, and a white-label SEO audit. Plans run from one free audit to Studio at $349 a month for 100 audits. See [pricing](/pricing) or how it works [for agencies](/for-agencies). ## Frequently asked questions ### What does a GEO agency do? A GEO agency helps brands appear in answers from AI assistants such as ChatGPT, Gemini and Google AI Overviews. The work usually combines measurement (which questions name the brand and which competitors win), technical fixes such as crawler access, content that answers buyer questions directly, and PR to earn mentions on sites the assistants cite. ### How much do agencies charge for GEO and AEO services? Published ranges are wide and mostly come from agencies themselves. First Page Sage reports a median GEO and AEO retainer of $6,850 a month. WebFX quotes $1,500 to $50,000 or more a month. One-time audits are often quoted in the low thousands. Treat all of these as rough guides, not benchmarks. ### What is the best entry offer for selling AI visibility services? A fixed-price AI visibility audit. It answers the question clients already have (do assistants mention us?), shows competitors winning in plain view, and produces a list of fixes that becomes the scope for ongoing work. It is quick to deliver and easy for a client to say yes to. ### Can an agency guarantee results in ChatGPT or AI Overviews? No, and it should not try. AI answers change between runs, and research by SparkToro found that AI tools repeated the same brand list less than 1% of the time. Promise the work and the measurement method, report mention rate across many questions, and avoid promising positions or specific citations. --- # GEO vs AEO vs SEO: Differences and What to Prioritise URL: https://aivisibilitydesk.com/blog/geo-vs-aeo-vs-seo Updated: 2026-11-20 > GEO vs AEO vs SEO: how generative engine, answer engine and search engine optimization differ, where they overlap, and which one to prioritise first. SEO, AEO and GEO are three layers of the same job. SEO gets your pages found and ranked in search results. AEO (answer engine optimization) shapes your content so it becomes the direct answer to a question. GEO (generative engine optimization) aims to get your brand named and cited inside answers that AI systems such as ChatGPT, Gemini and Google AI Overviews write. They share most of their groundwork, so the real question is not which one to pick but what order to work in. ## What does each term mean? ### SEO: search engine optimization The established discipline. Its goal is for your pages to rank well in search results and earn clicks. The work spans technical health (crawling, indexing, speed), content that matches what people search for, and links from other sites. Success is measured in rankings, impressions, clicks and the business those clicks bring. ### AEO: answer engine optimization AEO grew up around featured snippets and voice assistants, where the engine picks one answer and shows or reads it. The goal is to be that answer. The work is mostly about format and clarity: knowing the questions, answering each one directly at the top of a page, using headings, lists and tables, and keeping facts accurate. Our [answer engine optimization guide](/blog/answer-engine-optimization) covers it step by step. ### GEO: generative engine optimization GEO is the newest term, popularised by a [2023 research paper from Princeton and IIT Delhi](https://arxiv.org/abs/2311.09735). It targets AI systems that write an answer from several sources. The goal is to be included and cited in that answer, which depends on your own pages and also on what the rest of the web says about you: review sites, comparison articles, directories, forums. See our [generative engine optimization guide](/blog/generative-engine-optimization) for the full method. ## How do GEO, AEO and SEO compare? | | SEO | AEO | GEO | |---|---|---|---| | Goal | Rank pages in results | Be the direct answer | Be named and cited in AI answers | | Main surfaces | Google and Bing results | Featured snippets, voice, AI Overviews | ChatGPT, Gemini, AI Overviews | | Unit of success | A ranking and a click | An answer box or quoted passage | A brand mention or citation | | Main levers | Technical health, content, links | Question research, answer-first formatting | Clear positioning, quotable content, third-party mentions | | Typical metrics | Rankings, clicks, traffic | Snippets won, questions answered | Mention rate, position, share of voice, citations, sentiment | | How results vary | Fairly stable day to day | Fairly stable | Answers vary between runs, so sample repeatedly | ## Where do they overlap? More than the separate names suggest. A single well-made page can serve all three: - **It can be crawled and indexed.** That is SEO, and it is also a requirement for AI Overviews: Google says a page must be [indexed and eligible to show a snippet](https://developers.google.com/search/docs/appearance/ai-features), with no extra requirements. - **It answers a real question directly.** That is AEO, and it also makes the page easier for a generative engine to quote. - **It contains specific, sourced information.** That helps all three. In the GEO paper's experiments, adding statistics, quotations and citations improved visibility in generated answers by up to about 40%, while keyword stuffing did not help. The shared foundation is why many agencies sell GEO and AEO as an extension of their SEO work rather than as a separate product. ## Where do they differ in practice? The differences show up in what you measure and where you spend effort beyond your own site. **Measurement.** SEO has mature tools and stable numbers: you can see a ranking and check it tomorrow. AI answers change between runs and between engines, and there is no "position 3" to report. You measure GEO and AEO by asking a fixed set of questions repeatedly and counting how often a brand is named, where, and next to whom. That is a different kind of report, and one most SEO tools were not built to produce. **Off-site work.** SEO values links from other sites. GEO values mentions on the sites AI engines actually cite in your category, which may or may not be the sites that pass link value. A review platform or a forum thread can matter a lot for GEO even when a link from it does little for rankings. **Brand description.** SEO rarely forces you to settle how the world describes your company. GEO does, because the engine will write a sentence about you. If your site, your listings and third-party articles describe you inconsistently, the engine picks one description, or none. **Engines.** SEO mostly means Google, with Bing behind it. GEO means checking each AI system on its own. It is common to see a brand named often in Google's AI Overviews and rarely in ChatGPT, or the reverse. ## Why the click matters less than it did The reason GEO and AEO get attention is that AI answers reduce clicks. A [Pew Research Center analysis](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/) of US adults' Google searches found people clicked a regular result in 8% of visits when an AI summary appeared and 15% when it did not. Google disputed the methodology. Even taken cautiously, it means a ranking is worth less when an answer above it has already told the reader which brands to consider. Being one of those brands is the point of GEO. Search interest reflects this shift. In DataForSEO's US data for November 2026, "generative engine optimization" averaged 4,400 searches a month and "answer engine optimization" 2,400, while "llm seo" averaged 720 and was lower than a year earlier. The vocabulary is still settling, but GEO is currently the most searched label. ## Which should you prioritise? It depends on where you are starting from. A simple order works for most brands. ### 1. Fix SEO fundamentals first if they are broken If your site has indexing problems, blocks crawlers, hides key content behind scripts or has thin pages, fix that first. Neither answer engines nor generative engines can use pages they cannot read. Check robots.txt for AI search crawlers as part of this: OpenAI, for example, documents [OAI-SearchBot](https://developers.openai.com/api/docs/bots) as the crawler for ChatGPT search, separate from its training crawler. ### 2. Measure your AI visibility Before investing in GEO, find out whether you have a problem. Run your core buyer questions through ChatGPT, Gemini and Google AI Overviews and look at mention rate per engine, position, share of voice and which sites are cited. Some brands find they already do well; others find a competitor named in nearly every answer. Our [guide to AI visibility](/blog/what-is-ai-visibility) explains these metrics. ### 3. Do the AEO work on your own pages Rewrite or create pages for the questions where you are missing: answer first, specific, sourced, kept current. This also tends to help classic SEO, so the effort is rarely wasted. ### 4. Do the GEO work off your site Use the cited-sources data from your audit to decide where to earn coverage. If the engines keep citing the same review sites and comparison articles, that is your list. ### 5. Re-measure on the same questions Rerun the same question set later and compare. Keep the questions fixed so the change you see reflects the work, not a new set of questions. ## What does this mean for agencies? Most SEO agencies already have the skills: technical audits, content and digital PR. What is new is the measurement and the reporting. Clients are asking "do we show up in ChatGPT?", and a rankings report does not answer that. Agencies that can show a clear, repeatable AI visibility measurement, explain what it means and tie it to a plan are in a good position to add GEO and AEO to existing retainers. Our page [for agencies](/for-agencies) explains how a white-label audit fits into that. ## Measure before you choose AI Visibility Desk gives you the measurement that GEO and AEO decisions depend on. It asks your buyer questions on ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity and Claude, then reports mention rate per engine, position in the answer, share of voice, citations, competitors you had not listed, sentiment and whether the site lets AI crawlers in. Agencies can publish the result under their own brand. Schedule it to re-run weekly or monthly and the trend builds up on its own. Plans are on the [pricing page](/pricing). ## Frequently asked questions ### What is the difference between GEO, AEO and SEO? SEO aims to rank your pages in search results. AEO aims to make your content the direct answer to a question, in featured snippets, voice results and AI answers. GEO aims to get your brand named and cited inside answers written by AI systems like ChatGPT, Gemini and Google AI Overviews. They share most of the same groundwork. ### Should I focus on GEO or SEO? Start with SEO fundamentals if they are weak, because AI answers often draw on pages search engines can already find and rank. If your SEO is in reasonable shape, measure your AI visibility and add GEO work where the data shows you are missing from answers your competitors appear in. ### Is AEO the same as GEO? Mostly. The terms come from different eras: AEO from featured snippets and voice search, GEO from a 2023 research paper on generative engines. AEO leans towards on-page answer formatting; GEO puts more weight on being cited across many sources. In practice most agencies offer them as one service. ### Is SEO dead because of AI search? No. Google says pages need to be indexed and snippet-eligible to appear in AI Overviews, and assistants that browse the web read pages they can find. What has changed is the payoff: fewer people click when an AI answer appears, so being named in the answer matters alongside ranking. --- # How to Run an AI Visibility Audit, Step by Step URL: https://aivisibilitydesk.com/blog/how-to-run-an-ai-visibility-audit Updated: 2026-11-20 > A practical AI visibility audit method: pick buyer questions, ask ChatGPT, Gemini and Google AI Overviews more than once, then score mentions and citations. An AI visibility audit measures how often AI assistants name a brand when people ask the questions buyers ask, and who they name instead. You pick 20 to 30 buyer questions, ask each one on ChatGPT, Gemini and Google AI Overviews more than once, and record whether the brand is named, where in the answer, which competitors appear and which websites are cited. The result is a mention rate per engine, a share of voice against competitors and a list of the sources the assistants rely on. This guide walks through the method so you can do it by hand with a spreadsheet. At the end we show how [AI Visibility Desk](/ai-visibility-checker) runs the same steps for you. If you need the background first, start with [what AI visibility is](/blog/what-is-ai-visibility). ## Why run an AI search audit at all? Buyers increasingly read an answer instead of a list of links. When Pew Research tracked the browsing of 900 US adults in March 2025, people who saw a Google AI summary clicked a regular search result in [8% of visits, against 15% when no summary appeared](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/). If the answer names three competitors and not your client, ranking well below it does less than it used to. Classic SEO tools do not tell you what those answers say. An AI SEO audit fills that gap: it shows whether the brand is part of the answer, and gives you a baseline before any [generative engine optimization](/blog/generative-engine-optimization) work starts. ## Step 1: Define the brand, its other names and its competitors Write down: - **The brand name and every other name it goes by.** Product names, abbreviations, the old name after a rebrand. If an answer says "FB" or the product name instead of the company name, you still want to count it. - **The website domain.** You will need it to tell when the assistants cite the brand's own pages. - **Three to five competitors, with their websites.** Pick the ones a buyer would actually compare against, not the biggest names in the industry. Five is enough to make share of voice meaningful without turning the spreadsheet into a wall. Watch for names that are ordinary words. A brand called "Wave" or "Notion" will match sentences that have nothing to do with the brand. Decide now that you will read those matches rather than trust a text search. ## Step 2: Choose the buyer questions This is the step that decides whether the audit is useful. Write the questions a prospect would type before they know which brand they want. Do not include the brand's name: a question that names the brand will almost always produce an answer that names it, which tells you nothing about visibility. Spread the questions across the kinds of intent buyers have: | Question type | Example (accounting software) | |---|---| | Best-of | "What is the best accounting software for freelancers?" | | Comparison and alternatives | "What are good alternatives to QuickBooks for a one-person business?" | | Price-led | "What is the cheapest invoicing tool that handles VAT?" | | How-to or problem-led | "How do I track expenses and invoices in one place as a contractor?" | | Use case | "Accounting software for a design studio with three people" | Aim for 20 to 30 questions. Pull wording from sales calls, support tickets, Reddit threads and the "People also ask" box in Google. Write them the way people talk, including the longer, specific ones; Pew found [AI summaries appeared on 53% of Google searches of ten words or more](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/). Separately, ask two or three questions that name the brand ("What do people say about Brand X?"). Keep them out of the scores. They are there to hear how the assistants describe the brand when asked directly. ## Step 3: Pick the engines Cover the assistants your client's buyers use. For most audits that means: - **ChatGPT**, which for many buyers is the assistant they mean when they say "I asked AI". - **Gemini**, Google's assistant. - **Google AI Overviews**, the summary at the top of Google results. Not every search shows one, so record "no overview" separately from "overview that does not name the brand". Use a logged-out or private browser session and the same location for every check. Personalisation and location change answers, and you want results anyone could reproduce. ## Step 4: Ask each question more than once AI answers are not stable. In a 2025 study run by SparkToro and Gumshoe, volunteers ran 12 recommendation prompts 2,961 times, and the researchers found [less than a 1 in 100 chance that ChatGPT or Google's AI would return the same list of brands twice](https://sparktoro.com/blog/new-research-ais-are-highly-inconsistent-when-recommending-brands-or-products-marketers-should-take-care-when-tracking-ai-visibility/). The order changed even more often. Their conclusion was that how often a brand appears across many runs carries real signal, while its position in any single answer does not. So ask each question at least twice on the chat assistants. If the two runs disagree (named once, not named once), that disagreement is information: the brand is on the edge of the answer for that question. Note how many question-and-engine pairs disagreed; it tells the client how much of the score is settled and how much could swing. ## Step 5: Record each answer For every answer, log: 1. Whether the brand is named (yes, no, or "no answer"). 2. Which competitors are named. 3. The order: is the brand the first of your listed brands to be named, second, third? 4. The sources cited, by domain and, where shown, by page. 5. A short passage around the mention, so someone else can check the match. 6. Any brands named that are not on your competitor list. A spreadsheet with one row per question, engine and run works. Expect it to be long: 20 questions asked twice on ChatGPT, twice on Gemini and once on Google is 100 rows. ## Step 6: Score the results ### Mention rate per engine For each engine: answers that name the brand divided by answers returned. Leave out checks that produced no answer (a Google search with no AI Overview, or a request that failed) rather than counting them as misses. An overall score can be the average of the three engine rates, so one engine with more checks does not dominate. ### Position in the answer For answers that name the brand, count how many of your listed competitors were named before it. If none were, it ranks first. Report the average rank and the share of answers where the brand came first. Treat it as supporting detail, given how much ordering varies. ### Share of voice Count the answers that name each brand, add them up across the client and its competitors, and divide each brand's count by the total. We cover the formula, a worked example and the traps in [our guide to AI share of voice](/blog/ai-share-of-voice). ### Questions where the brand is missing everywhere List the questions where the brand was not named in any run on any engine. These are usually the most useful lines in the report, because each one points at a piece of content or coverage that does not exist yet. ### Head-to-head and question types For each competitor, count the questions where only they were named, only the client, both, or neither. Then group questions by type (best-of, comparison, price, how-to, use case) and compare mention rates. A brand that wins comparisons but disappears on price questions has a specific, fixable gap. ## Step 7: Analyse the citations The sources the assistants cite show where their information comes from. For each cited domain, note how often it appears and what kind of site it is: the client's own site, a competitor's site, a review site such as G2 or Capterra, a forum such as Reddit, video, a reference site such as Wikipedia, or editorial and comparison content. Two lists matter most: - **How often the client's own site is cited** when the client is named. If assistants name the brand but cite other people's pages, the brand's own content is not doing the work. - **Third-party sources cited in answers that name a competitor but not the client.** These are the pages to target for coverage, reviews or corrections. Our article on [getting cited in AI Overviews](/blog/get-cited-in-ai-overviews) goes into what tends to earn a citation. Also check that the client's site lets the AI search crawlers in. OpenAI states that [sites that block OAI-SearchBot will not be shown in ChatGPT search answers](https://developers.openai.com/api/docs/bots), and Google says a page must be [indexed and eligible for a snippet](https://developers.google.com/search/docs/appearance/ai-features) to appear as a supporting link in AI Overviews. ## Step 8: Read what is said about the brand A mention is not always good news. Read the passages around each mention and mark them positive, neutral or negative, with what the brand is praised or criticised for. Group similar points ("easy to use", "expensive for teams") so the client sees themes rather than a pile of quotes. Be clear in the report that this is a reading of the text, not a measurement. ## Step 9: Write it up A good audit report has: - The headline mention rate and the rate per engine. - Share of voice against the listed competitors. - The questions where the brand is missing, with who was named instead. - The top cited sources and the gaps among them. - What the assistants say about the brand. - Three to five concrete recommendations tied to those findings. - The method: number of questions, runs per engine, date, and how many checks returned nothing. See [what goes into an AI visibility report](/ai-visibility-report) for a fuller outline. ## How long does a manual audit take? Asking 20 questions five times each, reading every answer and logging brands and sources is easily most of a working day, before any analysis. It is a good way to learn how the assistants behave. It is a slow way to produce audits for several clients a month. ## How AI Visibility Desk automates the audit [AI Visibility Desk](/for-agencies) runs the method above for agencies and publishes the result under the agency's own brand. - **Setup.** Enter the client's website and the tool reads the site, searches the web and fills in the brand, category, competitors (up to five) and buyer questions for you to review and edit. Plans allow up to 10, 20 or 30 questions per audit. - **Runs.** Each question is asked twice on ChatGPT, twice on Gemini and once on Google. Two extra questions name the brand to hear what is said about it; they are reported separately and never scored. - **Turnaround.** In our measurements ChatGPT answers arrive in about 20 seconds, Google in about 80 seconds and Gemini in 8 to 12 minutes. Gemini is the least reliable: some checks never come back. Those are left out of the scores and the report states how many. - **Scores.** Mention rate per engine and overall, average position and how often the client is named first, share of voice, head-to-head against each competitor, mention rate by question type, and questions where the client is missing everywhere. - **Citations.** Top cited domains by type, cited pages, how often the client's own site is cited, and sources that cite competitors but not the client. A check of the client's robots.txt for AI crawlers and of the homepage. - **Review.** Every answer is shown with a short passage, so a wrong match (the "Wave" problem) can be corrected by hand. Brands the assistants named that you did not list are shown so you can add or hide them. A model can read the mentions for sentiment and group them into themes. - **Report.** Preview, then publish to a private link carrying the agency's name, logo and colours. Print to PDF from the browser, or download every answer as a CSV. It covers ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity and Claude, and on paid plans an audit can re-run weekly or monthly to show the trend. It does not cover Copilot, Grok or Meta AI. ## Try it on a client If you want to see the method run on a real brand, the free plan includes one audit with up to 10 questions. [See the plans](/pricing) or read how the [AI visibility checker](/ai-visibility-checker) works. ## Frequently asked questions ### What is an AI visibility audit? An AI visibility audit measures how often AI assistants such as ChatGPT, Gemini and Google AI Overviews name a brand when buyers ask category questions. It records where the brand appears in each answer, which competitors are named instead, and which websites the assistants cite as sources. ### How many prompts do you need for an AI visibility audit? Twenty to thirty buyer questions is a workable size for one brand. Fewer than ten makes each answer swing the score too much. Ask every question more than once on engines that vary, because the same question can return a different list of brands from one run to the next. ### Can I do an AI search audit by hand? Yes. Open each assistant in a logged-out or private session, ask your list of questions, and record in a spreadsheet whether the brand and each competitor is named and which sources are cited. It works, but 20 questions asked five times each is 100 answers to read and log. ### How often should an AI visibility audit be repeated? Answers change as the models and the web change, so treat one audit as a snapshot. A sensible rhythm is once a quarter, or after a significant content or PR push, using the same questions and competitors so the two results can be compared fairly. --- # LLM SEO: What Works, What's Evidence, What's Guesswork URL: https://aivisibilitydesk.com/blog/llm-seo Updated: 2026-11-20 > LLM SEO means getting your brand named in ChatGPT, Gemini and Google AI answers. What the evidence supports, what is speculation, and how to measure it. LLM SEO is the work of getting your brand named, described correctly and cited when people ask large language models like ChatGPT, Gemini or Google's AI Overviews about your category. In practice most of it is ordinary SEO and brand building, because the assistants that answer buying questions usually search the web first and draw on pages and brands that are already well established. The part that is new is measurement: you track mentions inside answers, not positions in a list of links. This guide separates what has evidence behind it from what is still guesswork, and shows how to measure LLM visibility without fooling yourself. ## What does "LLM SEO" actually mean? The phrase covers the same ground as generative engine optimization (GEO) and answer engine optimization (AEO). Different people use different labels for the same goal: show up in AI-written answers. The differences are mostly about emphasis. Search interest in the exact phrase has cooled while the broader topic has grown. In DataForSEO's US data for November 2026, "llm seo" averaged 720 searches a month, with 480 in the latest month against 1,000 a year earlier. Over the same period "ai visibility" went from 390 to 3,600 a month. The work is the same; the vocabulary is moving toward "AI visibility". For an agency, the useful definition is practical: LLM SEO is anything that raises the share of relevant AI answers that name your client, and LLM visibility is that share. ## How do language models decide which brands to mention? There are two routes from your brand into an answer, and they respond to different work. ### Route 1: what the model already "knows" A model learns from a large snapshot of text during training. If your brand was discussed often and consistently across the web before that snapshot, the model may name you from memory, without searching. You cannot edit this directly, and changes take effect only when a new model is trained. The lever is the same one PR has always used: be talked about, in the places your category is discussed. ### Route 2: what the model finds when it searches Most buying questions now trigger a live search. ChatGPT search fetches pages through OpenAI's crawler, and OpenAI's [crawler documentation](https://developers.openai.com/api/docs/bots) says sites that block OAI-SearchBot will not be shown in ChatGPT search answers, though they can still appear as navigational links. Google's AI Overviews and AI Mode are built on Google's own index and ranking systems, and Google describes a "query fan-out" step in which the model runs several related searches at once to gather more results ([Google's AI optimization guide](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide)). This route is where most practical LLM SEO happens, because it rewards things you can change this quarter: whether crawlers can reach your pages, whether those pages rank for the questions behind the question, and whether other sites that rank mention you. ## What does the evidence actually support? Plenty of LLM SEO advice is repeated without a source. These are the claims that do have one. | Claim | Evidence | Strength | |---|---|---| | Your pages must be indexable to appear in Google's AI features | Google says a page must be indexed and eligible to show in Search with a snippet ([AI features and your website](https://developers.google.com/search/docs/appearance/ai-features)) | Official documentation | | Blocking OAI-SearchBot keeps you out of ChatGPT search answers | Stated in [OpenAI's crawler docs](https://developers.openai.com/api/docs/bots) | Official documentation | | Pages that rank are more likely to be cited in AI Overviews | Ahrefs found 76% of AI Overview citations ranked in the top 10 in 2025, falling to about 38% in a [2026 rerun](https://ahrefs.com/blog/ai-overview-citations-top-10/) as Google pulled from more sub-queries | Large correlational study | | Brands mentioned widely on the web appear more in AI Overviews | Ahrefs measured a 0.664 correlation between web mentions and AI Overview brand visibility across 75,000 brands ([study](https://ahrefs.com/blog/ai-overview-brand-correlation/)), stronger than backlinks | Correlation, not proof of cause | | Adding citations, quotations and statistics can raise visibility | The [GEO paper](https://arxiv.org/abs/2311.09735) (Aggarwal et al., KDD 2024) reported gains of up to 40% on its benchmark | Controlled experiment, not live engines | Two cautions apply to the whole table. Correlation studies tell you what visible brands have in common, not what made them visible. And the GEO paper tested a research setup, so "up to 40%" is a ceiling observed in a benchmark, not a forecast for your site. Still, the pattern is consistent: be crawlable, rank for the questions buyers ask (and the related ones), get mentioned by other sites, and write pages that contain specific, checkable facts. ## What is still speculation? Some popular tactics have no evidence, or have evidence against them. - **llms.txt files.** Google says Search ignores llms.txt and other AI text files, so they neither help nor hurt ([AI optimization guide](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide)). No major assistant has said it uses them to choose sources. - **Special "AI schema".** Google says no special schema.org markup is needed for its AI features. Structured data is still worth doing for normal rich results. - **Rewriting or chunking everything "for AI".** Google lists this among things that are not necessary. Clear pages with direct answers help human readers and machines alike, but there is no separate AI format. - **Ranking first inside ChatGPT.** Answers vary too much for a stable position. A SparkToro study with Gumshoe.ai found AI tools returned the same brand list less than 1% of the time across repeated runs ([Search Engine Land summary](https://searchengineland.com/ai-recommendation-lists-rarely-repeat-study-468076)). The authors suggested tracking how often a brand appears instead of its rank. Note that Gumshoe sells AI tracking, which they disclosed. - **Manufactured mentions.** Google names seeking inauthentic mentions as ineffective, and mass-produced pages aimed at AI answers fall under its scaled content abuse policy. If someone sells you a guaranteed position in ChatGPT, ask how they would measure it given that the answer changes from one run to the next. ## How do you measure LLM visibility? Because answers vary, a single screenshot proves very little. Useful measurement looks like a small survey. 1. **Write buyer questions without your brand in them.** "Best invoicing software for freelancers", not "Is FreshBooks good?". Fifteen to thirty questions is a workable set. 2. **Ask each question more than once per engine.** Even two runs per engine catch some of the variation that a single check hides. 3. **Record more than yes or no.** Was the brand named? How early in the answer? Which competitors appeared? Which websites were cited, and was the client's own site among them? Was the description positive, neutral or negative? 4. **Report engines separately.** ChatGPT, Gemini and Google often disagree on the same question, and the fixes differ. 5. **Keep branded questions apart.** Asking "What is [brand]?" tells you how the brand is described, but it should not inflate the visibility score. The headline numbers are mention rate per engine and [AI share of voice](/blog/ai-share-of-voice), which compares your mentions with competitors' across the same answers. Our guide on [how to run an AI visibility audit](/blog/how-to-run-an-ai-visibility-audit) walks through the full process. ### What to look for in an LLM visibility tool A good LLM visibility tool should show you the answers, not only a score. You want to read the passage where the brand was matched, correct a wrong match (a competitor called "Wave" will match the everyday word), and see which sources the assistant leaned on. Be wary of tools that report a single "AI rank", and of any that claim access to internal Google metrics, which Google itself warns about. Also check which engines a tool covers. Coverage differs a lot between products, and a number from one assistant does not tell you much about another. ## A practical LLM SEO checklist Start with the items that have evidence and that you can verify. 1. **Check crawler access.** Confirm robots.txt and your CDN or firewall do not block Googlebot or OAI-SearchBot. Bot-protection settings can block crawlers without anyone noticing, so check rather than assume. 2. **Confirm indexing and snippets.** Pages with `nosnippet` or missing from Google's index cannot be used in AI Overviews. 3. **Map the questions behind the question.** For each buyer question, list the sub-questions a fan-out search would run (pricing, alternatives, comparisons, "for small teams"), and make sure you have a strong page for each. 4. **Put the answer near the top.** Open key pages with a direct, specific answer, then support it. Include facts a model can quote: prices, limits, numbers, dates. 5. **Earn mentions where the category is discussed.** Comparison articles, review sites, industry roundups and community threads that already get cited are worth more than another blog post on your own site. Your audit's citation list tells you which sites those are. 6. **Keep facts consistent.** Your name, what you do, who it is for and what it costs should read the same on your site, profiles and directories, so the model is not choosing between versions. 7. **Fix negative or outdated descriptions at the source.** If answers repeat an old complaint or a retired price, find the page they come from. 8. **Measure before and after.** Run the same question set before you start and again after the work has had time to be crawled, and compare. None of this is exotic. Most of it is SEO and PR with a new scoreboard, which is why agencies are well placed to sell it. If that is you, read [how to sell GEO and AEO services](/blog/sell-geo-aeo-services-agency). ## Measuring LLM visibility with AI Visibility Desk AI Visibility Desk runs the measurement half of this for agencies. You enter a client, its competitors and up to 30 buyer questions (or let it read the client's site and suggest them). It asks each question twice on ChatGPT, twice on Gemini and once on Google, then reports mention rate per engine, typical place in the answer, share of voice, which sites were cited, how often the client's own site was among them, sentiment, crawler access to the site, and brands the assistants named that you did not list. Every answer is there to read and correct. The report is published under your agency's name and logo at a private link. It covers ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity and Claude, and can re-run weekly or monthly to show the trend. Plans start with one free audit; see [pricing](/pricing) or what it looks like [for agencies](/for-agencies). ## Frequently asked questions ### What is LLM SEO? LLM SEO is the work of getting a brand mentioned, accurately described and cited in answers from large language models such as ChatGPT, Gemini and Google's AI Overviews. It overlaps heavily with classic SEO, because many assistants search the web before answering, but it is measured by mentions in answers rather than by rank positions. ### What is LLM visibility? LLM visibility is how often an AI assistant names your brand when people ask questions about your category. It is usually measured as a mention rate: ask a fixed set of buyer questions several times, then count the share of answers that name you, where they name you, and which competitors and sources appear alongside. ### Does llms.txt help with LLM SEO? There is no public evidence that it does. Google says its Search ignores llms.txt and similar AI text files, so they neither help nor hurt visibility in AI Overviews or AI Mode. Other AI companies have not said they use it for ranking or citation. It is cheap to add, but do not expect results from it. ### What should an LLM visibility tool measure? At minimum: whether each answer names your brand, how early in the answer, which competitors are named, which websites are cited, and how the answer describes you. It should run each question more than once, because AI answers vary between runs, and report each engine separately, because ChatGPT, Gemini and Google often disagree. --- # What Is AI Visibility? How to Measure and Improve It URL: https://aivisibilitydesk.com/blog/what-is-ai-visibility Updated: 2026-11-20 > AI visibility is how often AI assistants like ChatGPT, Gemini and Google AI Overviews name your brand. Learn how to measure it and what moves the number. AI visibility is how often, and how prominently, AI assistants such as ChatGPT, Gemini and Google's AI Overviews name your brand when someone asks a question about your category. You measure it by asking a fixed set of real buyer questions and counting how many answers mention you, where in the answer you appear, which competitors are named instead, and which websites the assistant cites. It is the AI-era equivalent of asking "do we show up when people search for what we sell?" ## Why does AI visibility matter now? Because a growing share of product research now ends with a written answer instead of a list of links. The audiences are large. OpenAI said in February 2026 that ChatGPT had [900 million weekly active users](https://techcrunch.com/2026/02/27/chatgpt-reaches-900m-weekly-active-users). Google said in July 2025 that AI Overviews reached [more than 2 billion monthly users](https://techcrunch.com/2025/07/23/googles-ai-overviews-have-2b-monthly-users-ai-mode-100m-in-the-us-and-india). Not all of those sessions are buying decisions, but a fair number of them are questions like "best accounting software for freelancers" or "which CRM is easiest for a small agency". The answer format also changes what people click. In a [Pew Research Center study of 900 US adults' browsing](https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/), Google users who saw an AI summary clicked a regular result in 8% of visits, against 15% when no summary appeared. Only 1% of visits included a click on a link inside the summary itself. Google disputed the methodology, but the direction is hard to ignore: when the answer is on the page, fewer people go looking for it. If your brand is not in that answer, you may not get a second chance. Interest in the topic has followed. In DataForSEO's US data for November 2026, "ai visibility" averaged 1,000 searches a month, with 3,600 in the latest month against 390 a year earlier. "ai visibility audit" went from 70 to 1,600 over the same period. ## What does AI visibility actually measure? "AI brand visibility" is a loose phrase, so it helps to break it into parts you can count. Most serious audits look at five things. ### 1. Mention rate The share of answers that name your brand. This is the headline number and what most people mean by an "AI visibility score". If Gemini names you in 9 of 30 answers, your Gemini mention rate is 30%. Report it per engine. ChatGPT, Gemini and Google AI Overviews are built differently, use different sources and often disagree about the same question. A single blended number hides the engine where you are missing. ### 2. Position in the answer Being named first is not the same as being named seventh in a list of ten. Assistants tend to lead with the options they consider strongest, and readers skim. Track where in the answer your brand typically appears: first mention, top three, or somewhere further down. ### 3. Citations Many answers come with sources. Two questions matter here: how often is your own website cited, and which other sites are cited when your category comes up? The second list is often more useful than the first. It tells you which review sites, directories, comparison articles and forums the assistants lean on, and therefore where your brand needs to be present. ### 4. Share of voice Mention rate looks at you alone. AI share of voice looks at everyone: of all the brand mentions across all the answers, what share is yours? A 25% mention rate is strong in a crowded category and weak in a category with three players. Share of voice also surfaces brands you did not think of as competitors but the assistants clearly do. ### 5. Sentiment When you are mentioned, what is said? An answer that names you as "expensive and hard to set up" is a mention, but not a helpful one. Sentiment is a reading of the text around each mention: positive, neutral or negative, and what you are praised or criticised for. It is an interpretation, so treat it as a guide, not a hard metric. | Metric | Question it answers | Why it matters | |---|---|---| | Mention rate | Do assistants name us? | Headline visibility, per engine | | Position | How early are we named? | Earlier mentions get read | | Citations | Which sites do assistants trust here? | Shows where to earn coverage | | Share of voice | How much of the conversation is ours? | Puts the score in context | | Sentiment | What do they say about us? | A mention can hurt as well as help | ## How do you measure AI visibility? The method is simple to describe and tedious to do by hand. Our step-by-step guide, [how to run an AI visibility audit](/blog/how-to-run-an-ai-visibility-audit), goes into more detail; the short version follows. **Write the questions.** Use 15 to 30 questions a real buyer would ask before they know which brand they want: "best X for Y", "X alternatives", "how do I choose X", "is X worth it for a small team". Leave your brand name out. Branded questions ("is Acme any good?") tell you something different and should be reported separately. **Ask more than once.** AI answers vary between runs. Asking each question twice on each chat assistant gives you a steadier figure than a single sample. **Cover the engines your buyers use.** ChatGPT, Gemini and Google AI Overviews cover a large share of AI-assisted search today. Each one needs to be checked on its own. **Record more than yes or no.** For each answer, note whether you were named, where, which competitors appeared, which URLs were cited and how you were described. **Check the basics on your site.** If your robots.txt blocks the crawlers that AI search products use, or your homepage gives little plain text to read, that is worth knowing before anything else. Done by hand, 20 questions across three engines with repeats is 100 answers to read and tally. That is why most agencies use a tool for it. When we ran a real 20-question audit through our own tool, the 100 checks cost $0.143 in API fees and most answers were back within 12 minutes. ## What is a good AI visibility score? There is no universal benchmark, and you should be wary of anyone who quotes one. A good score depends on: - **How crowded the category is.** In a market with five serious vendors, appearing in half the answers is strong. In a market with fifty, 15% might put you in the leading group. - **How specific the questions are.** Broad questions ("best project management software") favour the largest brands. Narrow ones ("project management for architecture firms") give specialists a fair chance. - **The engine.** It is common for a brand to do well on one assistant and barely appear on another. The most useful comparison is against your named competitors on the same questions, on the same day. That is what share of voice and head-to-head comparisons give you. ## How do you improve AI visibility? No one controls what an AI model says, and anyone promising guaranteed placement is overselling. What you can do is make your brand easier to find, easier to understand and more widely discussed in the places assistants read. The disciplines that cover this are usually called [generative engine optimization](/blog/generative-engine-optimization) (GEO) and answer engine optimization (AEO). In practice, the work looks like this. ### Fix access first Make sure AI search crawlers can reach your site. OpenAI, for example, documents a separate crawler, OAI-SearchBot, for [surfacing sites in ChatGPT search](https://developers.openai.com/api/docs/bots). Google says that to appear in AI Overviews a page only needs to be [indexed and eligible to show a snippet](https://developers.google.com/search/docs/appearance/ai-features), with no special markup required. Blocking crawlers by accident, or hiding key content behind scripts, quietly removes you from consideration. ### Say clearly what you are and who you are for Assistants summarise. If your homepage and product pages do not state plainly what you sell, for whom, and how you differ, the model has to guess, or it uses someone else's description of you. Write the sentence you would want an assistant to repeat, and put it where it can be read. ### Answer the questions buyers ask Build pages that answer the actual questions in your audit: comparisons, "best for" use cases, pricing explanations, how-to guides. Lead with a direct answer, then support it. Pages written this way are easier to quote. ### Be present where the assistants look Your citation data shows which third-party sites assistants rely on in your category. If they keep citing two review sites, an industry directory and a handful of comparison articles that do not mention you, that is your outreach list. Coverage on sites the assistants already trust often moves the needle more than another page on your own domain. ### Add evidence In the original [GEO research paper from Princeton and IIT Delhi](https://arxiv.org/abs/2311.09735), adding quotations, statistics and cited sources to a page improved its visibility in generative engine answers by up to about 40% in their tests, while keyword stuffing did not help. That was a controlled experiment, not a guarantee for your site, but it points the same way as common sense: specific, sourced content is easier for a model to use. ### Measure again Run the same questions again after the changes have had time to be picked up. Keep the question set fixed so the comparison is fair. ## AI visibility vs SEO AI visibility and SEO are related but not the same. SEO asks where your pages rank. AI visibility asks whether your brand is named in the answer. Strong rankings help, because assistants that search the web often draw from pages that already rank, but we regularly see brands that rank well and are still missing from AI answers, and the reverse. For a fuller comparison, see [GEO vs AEO vs SEO](/blog/geo-vs-aeo-vs-seo). ## Run an AI visibility audit If you want to see where a brand stands today, [AI Visibility Desk](/ai-visibility-checker) asks ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity and Claude a set of buyer questions, then reports mention rate per engine, position in the answer, citations, share of voice, competitors you had not listed and sentiment. Agencies can publish the result as a white-label report under their own branding at a private link. On paid plans it can re-run weekly or monthly, so the trend builds up on its own. Plans are on the [pricing page](/pricing). ## Frequently asked questions ### What is AI visibility? AI visibility is how often, and how prominently, AI assistants such as ChatGPT, Gemini and Google AI Overviews mention or cite a brand when people ask questions about its category. It is usually measured by asking a fixed set of buyer questions and counting how many answers name the brand, where, and alongside which competitors. ### What is an AI visibility score? An AI visibility score is usually the share of answers to a set of relevant questions that mention your brand. If ChatGPT names you in 12 of 40 answers, your mention rate there is 30%. Good scores are reported per engine, because ChatGPT, Gemini and Google often disagree about the same question. ### How do I check my brand's AI visibility? Write 15 to 30 questions a real buyer would ask, without your brand name in them. Ask each one on ChatGPT, Gemini and Google, more than once, and record whether you are named, where in the answer, which competitors appear and which sites are cited. A tool can run and score this for you. ### Is AI visibility the same as SEO? No, but they overlap. SEO measures where your pages rank in a list of links. AI visibility measures whether an assistant names your brand inside a written answer. Strong SEO helps, because assistants often draw on pages that already rank, but a brand can rank well and still be left out of AI answers. --- # Glossary ## AI crawler URL: https://aivisibilitydesk.com/glossary/ai-crawler An AI crawler is an automated bot run by an AI company that fetches web pages, either to collect training data for models, to build an index for AI search, or to load a page a user asked about. Each crawler identifies itself with a user agent name, and site owners can allow or block them in robots.txt. Blocking the wrong one can quietly remove a site from AI answers. ## Which AI crawlers should site owners know? **OpenAI** - **GPTBot** collects content that may be used to train OpenAI's models. - **OAI-SearchBot** finds and indexes pages to show and link in ChatGPT's search features. It is not used for training. - **ChatGPT-User** fetches a page when a ChatGPT user's request needs it, rather than crawling on its own. Because these visits are user-initiated, check OpenAI's current documentation for how robots.txt applies. **Anthropic** - **ClaudeBot** collects public web content that may be used to train Claude models. - **Claude-SearchBot** crawls to improve the quality of search results in Claude. - Anthropic also uses **Claude-User** for pages fetched in response to a user's request. **Perplexity** - **PerplexityBot** indexes pages so they can be shown and linked in Perplexity's answers. Perplexity says it is not used to train foundation models. **Google** - **Google-Extended** is a robots.txt control token, not a separate crawler. Googlebot does the crawling. Disallowing Google-Extended tells Google not to use your content to train Gemini models or ground answers in Gemini apps. It does not remove your site from Google Search, and it does not remove you from [AI Overviews](/glossary/ai-overviews) or AI Mode, which are part of Search. ## Should you block AI crawlers? That is a business decision. Many brands block training crawlers (GPTBot, ClaudeBot) but allow search and user crawlers (OAI-SearchBot, ChatGPT-User, Claude-SearchBot, PerplexityBot), so they can still be cited in answers. The [LLM SEO guide](/blog/llm-seo) covers how crawl access fits into the wider work. The common mistake is a blanket rule, or a firewall or bot-protection setting, that blocks everything without anyone noticing. ## Example A hypothetical furniture retailer, Oakwell, has a robots.txt rule copied from its staging site that disallows all user agents from its /products/ folder, plus a bot-protection setting that challenges unknown agents. OAI-SearchBot and PerplexityBot can't read any product page, and none are cited in answers. After both are fixed, the pages can be indexed again, and the agency tracks whether citations follow. ## How AI Visibility Desk checks access Every AI Visibility Desk audit includes an AI crawler access check, and the built-in site audit flags AI crawler blocking alongside other issues. The free [SEO checker](/seo-checker) runs a quick version. ## AI share of voice URL: https://aivisibilitydesk.com/glossary/share-of-voice AI share of voice is the percentage of all brand mentions in AI answers, for a defined set of questions, that belong to one brand. It answers the question "when AI tools talk about this category, how much of the conversation is about us compared with our competitors?" The idea comes from advertising, where share of voice meant a brand's share of total ad exposure in a market. ## How is AI share of voice calculated? The common method: 1. Pick a set of [buyer questions](/glossary/buyer-question) for the category. 2. Ask each question on each AI engine, ideally several times, since answers vary. 3. Count every mention of your brand and of each competitor across all answers. 4. Divide your brand's mentions by the total mentions of all tracked brands. Some tools weight mentions by position, giving more credit to a brand named first. Others count each brand once per answer. Either is fine as long as you use the same method every time you compare. ## How is it different from mention rate? [Mention rate](/glossary/mention-rate) is about your brand alone: the share of answers that name you. Share of voice is relative. A brand can have a 40% mention rate and still hold a small share of voice if competitors are named in nearly every answer. ## Example A hypothetical CRM vendor, PipeNorth, tracks 20 questions, asked 3 times on 5 engines (300 answers). Across them, tracked brands are mentioned 900 times in total. PipeNorth is mentioned 135 times, so its share of voice is 135 / 900 = 15%. The category leader has 315 mentions (35%). Two other competitors split most of the rest. The agency sets a goal of moving PipeNorth to 20% within two quarters. ## What is a good AI share of voice? It depends on how many serious competitors exist. In a category with three main players, a third is a fair baseline. In a crowded category with fifteen, 10% can be strong. The trend over time is usually more useful than the single number. ## How AI Visibility Desk measures it AI Visibility Desk reports share of voice across six engines, plus head-to-head results against each competitor and brands that appear in answers but weren't on your list. With [monitoring](/monitoring), the audit re-runs weekly or monthly so you can chart the trend. The [AI share of voice guide](/blog/ai-share-of-voice) covers the method in detail. ## AI visibility URL: https://aivisibilitydesk.com/glossary/ai-visibility AI visibility is how often, where and how favorably a brand appears in the answers AI assistants and AI search features give when people ask about its category. If a buyer asks ChatGPT "what's the best payroll software for a 20-person company?" and your client is named near the top with a fair description, that client has AI visibility for that question. If the answer lists five competitors and skips them, they don't. ## Why does AI visibility matter? More buyers now start research by asking an assistant instead of scrolling a list of links. The answer often names a short list of brands and stops there. A brand that ranks well in classic search can still be missing from those answers, because AI systems pull from a mix of training data, live search results and third-party pages such as reviews, forums and comparison articles. Search rankings alone no longer tell you whether buyers hear your name. ## How is AI visibility measured? There is no single official score. Most audits combine a few measurements taken across a set of buyer questions: - **Mention rate:** the share of answers that name the brand at all. See [mention rate](/glossary/mention-rate). - **Position:** where the brand appears in the answer, first or fifth. - **Share of voice:** the brand's mentions compared with competitors' mentions. See [share of voice](/glossary/share-of-voice). - **Citations:** whether the answer links to the brand's own pages. - **Sentiment:** whether the brand is described positively, neutrally or with caveats. Because answers change from one run to the next, a useful measurement asks each question several times and reports the averages. ## Example A hypothetical bakery supply brand, FlourCo, is checked on 20 buyer questions, each asked 3 times on 4 engines (240 answers). FlourCo is named in 72 of them, a 30% mention rate. Its main competitor is named in 132 (55%). FlourCo is cited by link in only 9 answers, all from Perplexity. That gap tells the agency where to focus: comparison content and third-party reviews that AI engines read. ## How AI Visibility Desk measures it AI Visibility Desk asks a brand's buyer questions repeatedly across ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity and Claude, then reports mention rate, position, share of voice and sentiment in a white-label report. You can try a quick version with the free [AI visibility checker](/ai-visibility-checker), or read the full [guide to AI visibility](/blog/what-is-ai-visibility). ## Answer engine URL: https://aivisibilitydesk.com/glossary/answer-engine An answer engine is a search or AI tool that responds to a question with a direct, written answer instead of a ranked list of links. Perplexity, ChatGPT with search, Claude, Gemini, Google AI Mode and Google AI Overviews all work this way. Earlier examples include featured snippets and voice assistants, which read out a single answer. ## How does an answer engine work? Most modern answer engines combine a large language model with live search. When a question comes in, the engine: 1. works out what is being asked, sometimes splitting it into several sub-queries, 2. retrieves relevant pages from a search index, 3. writes an answer based on those pages and on what the model already knows, 4. shows some of the pages as [citations](/glossary/citation). This pattern is called [retrieval-augmented generation](/glossary/rag). Some assistants only search when the question needs fresh information. Otherwise they answer from the model's training. ## Why do answer engines matter for marketers? In a classic search engine, ten brands can each get a blue link on page one. In an answer engine, the reply might name three brands, or one, or none. The buyer may never see a results page. That makes it more important to know which brands each engine names for the questions your customers ask, and why. Answer engines also differ from each other. They use different search indexes, different models and different rules for picking sources, so a brand can be visible in one and absent from another. ## Example A hypothetical bike shop, Spoke & Chain, asks four answer engines "where can I get a bike serviced quickly in Bristol?" Perplexity names Spoke & Chain and cites its Google Business Profile. ChatGPT names two competitors and cites a local blog. Claude gives general advice without naming shops. Google AI Mode lists three shops including Spoke & Chain. Four engines, four different answers. ## How AI Visibility Desk covers answer engines AI Visibility Desk checks six answer engines: ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity and Claude. It reports mention rate, position and cited sources for each, so you can see where a brand is strong and where it is missing. Start with the free [AI visibility checker](/ai-visibility-checker), or read about [answer engine optimization](/blog/answer-engine-optimization). ## Answer engine optimization (AEO) URL: https://aivisibilitydesk.com/glossary/answer-engine-optimization Answer engine optimization (AEO) is the practice of shaping content so that search features and AI assistants pick it as the direct answer to a question. The term predates today's AI chat tools: it grew out of work on featured snippets, "People also ask" boxes and voice assistants, and now also covers AI Overviews, AI Mode, ChatGPT and Perplexity. ## How is AEO different from GEO? The two overlap heavily and many people use them interchangeably. When a distinction is made, AEO focuses on the question and answer format: finding the exact questions people ask and writing clear, quotable answers to them. [Generative engine optimization](/glossary/generative-engine-optimization) is broader and includes off-site reputation, brand mentions and how AI models describe a brand overall. In practice, an agency doing one is usually doing parts of the other. ## What makes content a good answer? - **Answer first.** Put a plain one or two sentence answer right below the question heading, then add detail. - **Use the question as a heading.** Headings phrased the way people ask make it easy for systems to match a passage to a query. - **Be specific.** Numbers, steps, names and conditions are easier to quote than general claims. - **Use structured data where it is accurate.** FAQ, HowTo, Product and Organization markup help machines read the page, though markup alone does not guarantee selection. - **Keep it crawlable.** Answers hidden behind scripts, tabs that never load, or blocked crawlers can't be used. ## Example A hypothetical dental clinic, BrightSmile, has a page titled "Our Services" with a long paragraph about implants. The agency adds a section headed "How long do dental implants last?" with a two-sentence answer, followed by detail on care and costs. Over the next few weeks, the agency tracks whether answer engines start quoting that passage for the question. ## How AI Visibility Desk fits in AI Visibility Desk reports results by question type and lists missing questions, where competitors are named and the client isn't, which gives AEO work a clear target list. The built-in [AEO audit](/aeo-audit) also checks the site for topic and question gaps. Read more in the [AEO guide](/blog/answer-engine-optimization). ## Brand mentions URL: https://aivisibilitydesk.com/glossary/brand-mentions Brand mentions are references to a brand by name, either on other websites or inside AI-generated answers, whether or not they include a link. In SEO, the term usually means mentions on third-party sites such as news articles, reviews, forums and lists. In AI visibility work, it also means the brand being named in an AI engine's answer. ## How do web mentions affect AI answers? AI models learn about brands from text across the web. A brand that is discussed often, in context, on many independent sites is more likely to be part of the model's knowledge and to be described accurately. When an engine searches before answering, it often retrieves the same kinds of third-party pages: comparison articles, review sites, community threads. If a brand is named on those pages, it can appear in the answer even without a [citation](/glossary/citation) to its own site. Links still matter for classic SEO, but for AI answers an unlinked mention on a trusted page can carry real weight. ## Where do useful brand mentions come from? - Comparison and "best of" articles in the brand's category, - review platforms with recent, detailed reviews, - forums and community sites where buyers ask for recommendations, - industry news, interviews and podcasts, - directories and listings relevant to the category or location. The aim is accurate, contextual mentions on pages buyers and AI engines already read, rather than volume for its own sake. ## Example A hypothetical recruitment agency, TalentBridge, has a 10% mention rate across AI engines. Its two main competitors appear on four "top recruitment agencies in Manchester" lists that AI engines cite often. TalentBridge appears on none. The agency contacts the publishers with case studies and updated details. Two lists add TalentBridge. The next audit shows TalentBridge named more often in answers that cite those lists. ## How AI Visibility Desk tracks mentions AI Visibility Desk counts brand mentions in answers from six AI engines and reports [mention rate](/glossary/mention-rate), share of voice, head-to-head results against each competitor and [sentiment](/glossary/sentiment). It also lists the third-party sources engines cite, which gives a target list for outreach. See the [AI share of voice guide](/blog/ai-share-of-voice). ## Buyer question (prompt) URL: https://aivisibilitydesk.com/glossary/buyer-question A buyer question is a question a potential customer would ask an AI assistant while researching a purchase, such as "what's the best email marketing tool for a small online shop?" In AI visibility work these are often called prompts. They are the test inputs: you ask them on each AI engine and record which brands appear in the answers, which feeds measures like [mention rate](/glossary/mention-rate). ## How are buyer questions different from keywords? Keywords are short and typed into a search box ("email marketing software"). Buyer questions are longer and written the way people talk to an assistant, often with context about budget, size, location or needs. AI engines answer the whole question, so a brand can be visible for the keyword and still missing from the answer to a more specific question. ## What makes a good set of buyer questions? A useful set covers the different moments of a buying decision: - **Category questions:** "What are the best project management tools for agencies?" - **Problem questions:** "How can I stop missing client deadlines?" - **Comparison questions:** "Asana vs Trello for a 10-person team?" - **Specific need questions:** "Which project tools work well with time tracking?" Avoid putting the client's brand name in most questions. Branded questions show what engines say about the brand, but unbranded questions show whether the brand gets recommended at all. Aim for questions real buyers would ask, not ones written to make the client look good. Twenty to fifty questions is a practical starting range for one brand. ## Example A hypothetical coffee equipment shop, Grindhouse Supply, starts with 25 questions. Most are about espresso machines, but customers also ask about grinders, so the agency adds 8 grinder questions. The audit shows Grindhouse is named in 40% of espresso answers but only 5% of grinder answers. That gap becomes the first content brief. ## How AI Visibility Desk uses buyer questions AI Visibility Desk can suggest buyer questions, competitors and a category by reading the client's website, and shows AI search volume per prompt so you can see which questions people ask most. It reports results by question type and lists missing questions where competitors are named and the client isn't. See the [audit guide](/blog/how-to-run-an-ai-visibility-audit) for how to build a question set. ## Citation (in AI answers) URL: https://aivisibilitydesk.com/glossary/citation A citation in an AI answer is a linked source that an AI engine shows with its response to indicate where the information came from. Perplexity numbers its sources inline, Google AI Overviews and AI Mode show link cards, and ChatGPT and other assistants show source links when they search the web to answer. ## What is the difference between a citation and a mention? A mention is when the answer names your brand in the text. A citation is when the answer links to a page as a source. They often occur separately: - Your brand can be **mentioned without being cited**, because the engine learned about you elsewhere or drew on a third-party page. - Your page can be **cited without your brand being named**, for example when your blog post explains a concept and the answer quotes it without recommending you. Both are worth tracking. Mentions shape what buyers think. Citations can send traffic and show which pages engines trust. See [brand mentions](/glossary/brand-mentions). ## Which sources do AI engines cite? It varies by engine and topic. Common sources include brand websites, review platforms, forums, news sites, comparison articles and encyclopedic pages. Citations come from the retrieval step of [retrieval-augmented generation](/glossary/rag): the engine searches, picks pages and summarizes them. Pages that are crawlable, answer the question directly and contain specific facts are easier to use. ## Example A hypothetical HR software company, StaffLoop, is cited in 14 of 120 answers. Twelve of those citations go to a single page: a clear guide to calculating holiday entitlement. Meanwhile, competitors are cited from a comparison article on an industry blog that doesn't list StaffLoop. The agency's plan: build more guides like the one that works, and pitch StaffLoop's inclusion in that comparison article. ## How AI Visibility Desk tracks citations AI Visibility Desk lists the sources each engine cites for a client's buyer questions and separates out the client's own cited pages, so you can see which URLs earn citations. The [guide to getting cited in AI Overviews](/blog/get-cited-in-ai-overviews) covers how to improve the odds. ## E-E-A-T URL: https://aivisibilitydesk.com/glossary/e-e-a-t E-E-A-T stands for Experience, Expertise, Authoritativeness and Trustworthiness. It is a framework from Google's Search Quality Rater Guidelines, the document Google gives to people who evaluate the quality of search results. Google added the first "E", for Experience, in 2022. The guidelines describe Trust as the most important of the four. ## What does each part mean? - **Experience:** the creator has first-hand experience of the topic, such as having used the product or visited the place. - **Expertise:** the creator has the knowledge or skill the topic needs. This matters most for health, money, legal and safety topics. - **Authoritativeness:** the creator or site is a recognized go-to source for the topic. - **Trustworthiness:** the content is accurate, honest and safe, and the site is transparent about who runs it. ## Is E-E-A-T a ranking factor? Not directly. Quality raters don't change rankings, and there is no E-E-A-T score in Google's systems. Google uses rater feedback to check whether its ranking systems are working. In practice, the signals that suggest E-E-A-T (clear authorship, real experience, accurate information, a good reputation on other sites) line up with what Google's systems try to reward. ## How does E-E-A-T relate to AI answers? AI engines that search the web before answering also need to choose sources they consider reliable. Google's AI Overviews and AI Mode draw on the same Search systems. Other engines have their own methods, but they too tend to favor pages that are specific, well sourced and backed by reputation elsewhere. So the work that supports E-E-A-T, such as named authors with real credentials, original data, first-hand detail and independent [brand mentions](/glossary/brand-mentions), also helps a page become a [citation](/glossary/citation) in AI answers. ## Example A hypothetical physiotherapy clinic, MoveWell, publishes injury guides with no author name. The agency adds author bios for the clinic's licensed physiotherapists, includes photos and notes from real treatment sessions (with consent), links to the clinic's registration details and cites medical sources. The pages now show who wrote them and why they can be trusted. ## How AI Visibility Desk fits in AI Visibility Desk shows which sources each AI engine cites for a client's questions, which helps you see which kinds of pages engines trust in that category. Its built-in [SEO audit](/seo-audit) flags content issues with suggested fixes. ## Entity URL: https://aivisibilitydesk.com/glossary/entity An entity is a distinct, identifiable thing, such as a company, person, product, place or concept, that search engines and AI models recognize as one item and connect to facts about it. "Acme Plumbing" the business is an entity. So are its founder, its town and the services it offers. Google stores entities and their relationships in its Knowledge Graph, and language models build up a similar, less explicit picture during training. ## Why do entities matter for AI visibility? When an AI engine recommends brands, it has to know the brand exists, what category it belongs to, and what it is known for. If that picture is clear and consistent, the brand is easier to include in the right answers. If it is muddled, the model may leave the brand out, confuse it with a similarly named company, or describe it wrongly. Common entity problems: - the brand name is shared with other companies or common words, - the site, social profiles and directory listings describe the business differently, - the category is unclear (is it a software company or a consultancy?), - there are few independent sources that mention the brand. ## How do you strengthen a brand entity? - Describe the brand the same way everywhere: name, category, location, main products. - Add Organization structured data with accurate details and sameAs links to official profiles. - Keep key profiles complete and current (Google Business Profile, LinkedIn, relevant directories, Wikidata if the brand qualifies). - Earn [brand mentions](/glossary/brand-mentions) on independent sites that describe the brand in context. - Publish clear About and product pages that state plainly what the company does. ## Example A hypothetical consultancy called Northstar is often confused with other businesses using the same name. AI answers describe it as a travel agency. The agency updates its site to use "Northstar Data Consulting" consistently, adds Organization markup with links to its profiles, and corrects two directory listings. Later audits show more accurate descriptions in answers. ## How AI Visibility Desk helps AI Visibility Desk shows how each of six AI engines describes a brand, including [sentiment](/glossary/sentiment) and themes, and flags brands named in answers that weren't on your list. See the [LLM SEO guide](/blog/llm-seo) for more on entity work. ## Generative engine optimization (GEO) URL: https://aivisibilitydesk.com/glossary/generative-engine-optimization Generative engine optimization (GEO) is the practice of improving how often, and how well, a brand is mentioned and cited in answers written by generative AI tools such as ChatGPT, Gemini, Perplexity, Claude and Google's AI features. Where classic SEO aims for a ranked link, GEO aims for a place inside the generated answer itself. ## How is GEO different from SEO? SEO and GEO share a lot of groundwork. A site that is crawlable, fast, clearly written and well linked tends to do better in both. The differences are in what you measure and what influences the result: - **The output is a paragraph, not a list of links.** Success means being named, described accurately and ideally cited, rather than holding position three. - **Off-site sources weigh heavily.** AI answers often lean on review sites, forums, comparison articles and news coverage, so [brand mentions](/glossary/brand-mentions) on other sites matter as much as your own pages. - **Results vary by run.** The same question can produce a different answer an hour later, so GEO is measured as rates over many runs. For a side-by-side comparison, see [GEO vs AEO vs SEO](/blog/geo-vs-aeo-vs-seo). ## What does GEO work involve? Typical GEO work includes writing pages that answer buyer questions directly in the first sentences, adding facts and specifics that are easy to quote, making sure AI crawlers are allowed in robots.txt, keeping the brand's description consistent across the web, and earning coverage on the third-party pages that AI engines tend to cite. ## Example A hypothetical agency client, a project management tool called TaskHarbor, is named in 12% of answers to 25 buyer questions. The agency finds that competitors are cited from two comparison articles and a popular review site where TaskHarbor has few reviews. Over three months the agency publishes a clear comparison page, fixes a robots.txt rule that blocked GPTBot, and runs a review campaign. A follow-up audit shows TaskHarbor named in 27% of answers. ## How AI Visibility Desk fits in GEO needs a baseline and a way to track change. AI Visibility Desk runs a [GEO audit](/geo-audit) across six AI engines and can re-run it weekly or monthly with [monitoring](/monitoring), so agencies can show clients the trend. ## Google AI Mode URL: https://aivisibilitydesk.com/glossary/ai-mode Google AI Mode is a conversational search experience inside Google Search that answers questions with an AI-written response, links to sources and lets the user ask follow-up questions. It sits in its own tab or entry point, separate from the regular results page, and is built for longer, more complex questions than a typical keyword search. ## How is AI Mode different from AI Overviews? [AI Overviews](/glossary/ai-overviews) are a summary placed on top of a normal results page, and Google decides when to show one. AI Mode is a dedicated mode the user chooses, where the AI answer is the main content and the conversation can continue. Google has described AI Mode as using [query fan-out](/glossary/query-fan-out): it splits a question into related sub-queries, searches them, and combines what it finds. Answers tend to be longer and cite more sources than an AI Overview. Because the two features can cite different pages for the same question, it is worth tracking them separately. ## How do brands show up in AI Mode? The same basics apply as for other Google AI features. Pages need to be indexed and eligible for snippets. Content that answers specific sub-questions clearly gives AI Mode more to draw from. Comparison pages, detailed guides and pages with concrete facts (prices, specs, locations, opening hours) are often useful to it. Third-party pages that mention your brand can also be the source of a mention. ## Example A hypothetical travel insurer, Waypoint Cover, asks "what travel insurance covers skiing in Austria for a family of four?" in both features. The AI Overview cites two comparison sites and names neither Waypoint nor its competitors. AI Mode, with its longer answer, names three insurers including Waypoint and cites Waypoint's winter sports page. Tracking only AI Overviews would have missed that mention. ## How AI Visibility Desk tracks it AI Visibility Desk checks Google AI Mode as its own engine, next to Google AI Overviews, ChatGPT, Gemini, Perplexity and Claude, and reports mention rate and cited sources for each. Learn more about measuring Google's AI features in the [AI search visibility guide](/blog/ai-search-visibility), or see what a full [AI visibility report](/ai-visibility-report) includes. ## Google AI Overviews URL: https://aivisibilitydesk.com/glossary/ai-overviews Google AI Overviews are AI-generated summaries that appear at the top of some Google search results. They answer the query in a few sentences or a short list, and link to the web pages Google used to build the answer. They grew out of Google's earlier Search Generative Experience (SGE) test and rolled out widely in 2024. ## When do AI Overviews appear? Google decides per query. They tend to show for informational and comparison questions, such as "how to choose a CRM" or "best running shoes for flat feet", and less often for simple navigational searches like a brand name. Whether one appears can also change over time for the same query. ## How do sites get cited in AI Overviews? Google has said there are no special requirements beyond being indexed and eligible to show a snippet in Search. In practice, cited pages usually: - rank reasonably well for the query or for related sub-questions (Google uses [query fan-out](/glossary/query-fan-out) to search several related queries), - answer the question clearly near the top of a section, - are crawlable and not blocked with nosnippet or similar controls. Note that blocking the Google-Extended token in robots.txt does not remove a site from AI Overviews. See [AI crawler](/glossary/ai-crawler) for the details. ## Do AI Overviews reduce clicks? Often, yes, for the queries where they appear, because some searchers get what they need from the summary. This is part of the wider trend of [zero-click search](/glossary/zero-click-search). Being one of the cited sources, or being named in the text, is how a brand stays visible on those results. ## Example A hypothetical outdoor brand, Ridgeline Gear, tracks 30 buyer questions. AI Overviews appear on 18 of them. Ridgeline is named in the text of 4 and cited as a source in 6, mostly from its sizing guide. The agency uses that as a model and writes similar guides for the other product lines. ## How AI Visibility Desk tracks it AI Overviews is one of the six engines AI Visibility Desk checks, alongside Google AI Mode, ChatGPT, Gemini, Perplexity and Claude. Reports show mention rate, cited sources and the client's own cited pages. The guide on [getting cited in AI Overviews](/blog/get-cited-in-ai-overviews) covers tactics in more depth. ## LLM SEO URL: https://aivisibilitydesk.com/glossary/llm-seo LLM SEO is the work of improving how large language models (LLMs) such as ChatGPT, Gemini and Claude mention, describe and cite a brand when people ask them questions. It is another name for much of what is also called [generative engine optimization](/glossary/generative-engine-optimization). The label "LLM SEO" tends to be used by SEO teams who are extending their existing skills to AI assistants. ## How do LLMs decide which brands to mention? An LLM answer usually draws on two sources: - **What the model learned in training.** Brands that are written about often and consistently across the web are more likely to be part of the model's general knowledge. - **What it retrieves at answer time.** Many assistants search the web before answering, then summarize the pages they find. This is [retrieval-augmented generation](/glossary/rag), and it means current search visibility and crawl access matter. Which source dominates depends on the product and the question. A question about recent prices or new products leans on live retrieval. A broad "what is the best" question may lean more on learned knowledge plus a few retrieved pages. ## What does LLM SEO work look like? Much of it is familiar SEO, done with AI answers in mind. Make sure AI crawlers can reach the site. Describe the brand, its category and its products the same way everywhere, so the model forms a clear picture of who you are. Publish pages that answer buyer questions directly. Earn mentions on review sites, forums and comparison articles that models are likely to read. Then measure, because you can't see inside the model, only its answers. ## Example A hypothetical accounting firm, Ledgerline, ranks first on Google for "small business bookkeeping Leeds" but is never named when ChatGPT is asked the same question. An audit shows ChatGPT is citing two local directories where Ledgerline has no listing. Adding accurate listings there is a quick LLM SEO fix. ## How AI Visibility Desk fits in AI Visibility Desk checks a brand across ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity and Claude, and shows which sources each engine cites. The free [SEO checker](/seo-checker) is a quick way to spot crawl and content issues first. See the [LLM SEO guide](/blog/llm-seo) for more. ## llms.txt URL: https://aivisibilitydesk.com/glossary/llms-txt llms.txt is a proposed file, placed at the root of a website (for example example.com/llms.txt), that gives large language models a short Markdown summary of the site and links to its most useful pages. The idea, published in 2024, is that a clean, text-only guide is easier for an AI system to read than a full website with navigation, scripts and ads. ## What goes in an llms.txt file? The proposal suggests a simple structure: - an H1 with the site or company name, - a short summary in a blockquote, - optional notes about the site, - H2 sections with lists of links to key pages, each with a one-line description. Some sites also publish an llms-full.txt with the full text of their documentation in one file. ## Do AI engines actually use llms.txt? This is the honest part: it is a proposal, not a standard, and the major AI companies have not confirmed that their crawlers or answer engines use it to decide what to show. Google has said it does not use llms.txt for Search. Some developer tools and coding assistants do read llms.txt files for documentation. So treat it as a low-cost, low-risk addition, not a ranking lever. It takes little time to write, it forces a useful exercise (deciding which pages matter most), and it may become more widely used. It does not replace crawlable pages, clear content or an [AI crawler](/glossary/ai-crawler) policy in robots.txt. ## Example A hypothetical accounting software company, Tallybook, writes an llms.txt with a two-sentence summary and three sections: Product (pricing, features, integrations pages), Guides (its five most-read help articles) and Company (about and contact). It takes about an hour. The agency notes it in the client report as a completed housekeeping item, without promising any traffic change. ## How AI Visibility Desk checks it The built-in white-label site audit in AI Visibility Desk checks whether a site has an llms.txt file, alongside AI crawler blocking and other SEO, GEO and AEO issues. You can also run the free [SEO checker](/seo-checker) or a full [SEO audit](/seo-audit). ## Mention rate URL: https://aivisibilitydesk.com/glossary/mention-rate Mention rate is the percentage of AI-generated answers, across a set of buyer questions, that name a specific brand at least once. If a brand is named in 30 out of 100 answers, its mention rate is 30%. It is the simplest and most widely used measure of [AI visibility](/glossary/ai-visibility). ## How is mention rate calculated? Mention rate = answers that name the brand / total answers checked. Two details make the number more reliable: - **Repeat each question.** AI answers vary between runs, so one answer per question gives a noisy result. Asking each question several times and averaging smooths that out. - **Report by engine.** ChatGPT, Gemini, Perplexity, Claude and Google's AI features often disagree. A 25% overall rate can hide 60% on one engine and 5% on another. It also helps to note the brand's position when it is named. Being listed first is worth more than being the sixth item in a list. ## What is a good mention rate? There is no universal benchmark, and anyone quoting one should show their method. A useful rate depends on the category, the questions chosen and the number of competitors. A local business may reach a high rate on location-specific questions and a low one on broad category questions. Compare the brand with its direct competitors on the same questions, and compare it with itself over time. Those two comparisons say more than any fixed target. ## Example A hypothetical pet food brand, Kibble & Co, is checked on 15 buyer questions, each asked 3 times on 6 engines (270 answers). It is named in 81 answers, a 30% mention rate. By engine: Perplexity 49%, Google AI Mode 38%, ChatGPT 27%, Gemini 24%, Google AI Overviews 22%, Claude 20%. The agency looks at what Perplexity is citing to learn which sources help, and targets those for the weaker engines. ## How AI Visibility Desk measures it AI Visibility Desk asks each buyer question repeatedly on six engines and reports mention rate per engine, typical position in the answer, and results by question type. The free [AI visibility checker](/ai-visibility-checker) gives a quick first read, and the [how to run an AI visibility audit](/blog/how-to-run-an-ai-visibility-audit) guide walks through the full process. ## Query fan-out URL: https://aivisibilitydesk.com/glossary/query-fan-out Query fan-out is a technique where an AI search system breaks one question into several related sub-queries, searches for each, and combines what it finds into a single answer. Google has described Google AI Mode as working this way, and other answer engines that search the web follow a similar pattern. ## How does query fan-out work? Take the question "what's the best electric car for a family that drives long distances in winter?" Instead of searching that whole sentence, the system might run sub-queries such as: - electric cars with the longest real-world range, - EV range loss in cold weather, - electric SUVs with seven seats, - fast-charging speeds by model, - family EV reviews. It then reads the top results for each, picks the useful passages and writes one answer. The [citations](/glossary/citation) can come from any of those sub-queries, not only the original question. ## Why does query fan-out matter for visibility? It changes which pages get a chance to be cited. A page doesn't need to rank for the full question. A page that answers one sub-question well (for example, a clear table of cold-weather range) can be pulled into the answer. It also means a brand that only covers the head topic, and nothing around it, has fewer ways in. In practice, this rewards content that covers a topic in depth, with distinct sections that each answer one specific question clearly. Thinking about the sub-questions behind a [buyer question](/glossary/buyer-question) is a good way to plan that content. ## Example A hypothetical home insurance broker, Hearthside, is never named for "best home insurance for an old house". The agency lists likely sub-queries: insuring listed buildings, thatched roofs, older wiring, and non-standard construction. Hearthside has no pages on any of them. The agency writes four focused guides. In later checks, AI Mode cites the thatched roof guide for the original question. ## How AI Visibility Desk helps You can't see the exact sub-queries an engine runs, but you can see the result. AI Visibility Desk shows which sources each engine cites for a client's questions, and the built-in site audit finds topic and question gaps. A [GEO audit](/geo-audit) is a good starting point. See the guide on [getting cited in AI Overviews](/blog/get-cited-in-ai-overviews) for more. ## Retrieval-augmented generation (RAG) URL: https://aivisibilitydesk.com/glossary/rag Retrieval-augmented generation (RAG) is a method where an AI system first retrieves relevant documents, such as web pages, then uses them as context to write its answer. Instead of relying only on what the model learned during training, the answer is "grounded" in sources found at the moment the question is asked. The term comes from AI research published in 2020 and is now the basis of most AI search products. ## How does RAG work in AI search? In an AI search engine, the process looks roughly like this: 1. The user asks a question. 2. The system turns it into one or more search queries (sometimes many, see [query fan-out](/glossary/query-fan-out)). 3. It retrieves the most relevant pages or passages from a search index. 4. The language model reads those passages and writes an answer. 5. Some of the pages are shown as [citations](/glossary/citation). Perplexity, Google AI Overviews, Google AI Mode and the search features in ChatGPT, Gemini and Claude all use a version of this. ## Why does RAG matter for brands? RAG is the part of AI visibility you can influence fastest. Changing what a model learned in training takes a long time. Changing which pages a retrieval step finds can happen as soon as pages are crawled and indexed. That puts weight on things marketers already know how to work on: - pages must be crawlable by the right bots (see [AI crawler](/glossary/ai-crawler)), - pages must be findable for the queries the engine runs, - passages must answer the question clearly enough to be used, - third-party pages that get retrieved should mention the brand accurately. ## Example A hypothetical solar installer, BrightRoof, launches a new battery storage service. A month later, an assistant answering from training data alone doesn't know about it. But when the same assistant searches the web, it retrieves BrightRoof's new service page and a local news article, and names BrightRoof in the answer. The difference is retrieval. ## How AI Visibility Desk helps AI Visibility Desk shows which sources each of six engines cites, including the client's own cited pages, so you can see what retrieval is picking up. Its site audit checks crawl access for AI bots. Read the [LLM SEO guide](/blog/llm-seo) for more. ## Sentiment (in AI answers) URL: https://aivisibilitydesk.com/glossary/sentiment Sentiment in AI answers is whether an AI engine describes a brand positively, neutrally or negatively when it mentions it, along with the recurring themes behind that tone. Being named is only half the picture. "Brand X is a solid choice for small teams" and "Brand X is cheaper, though users report slow support" are both mentions, but they leave very different impressions. ## Why does sentiment in AI answers matter? A buyer who asks an assistant for recommendations often reads one short paragraph per brand. Any caveat in that paragraph carries a lot of weight, because there may be no other information on the page. AI engines build these descriptions from their training data and from pages they retrieve, such as reviews, forum threads and comparison articles. If those sources repeat a complaint, the answer may repeat it too, even if the problem was fixed long ago. ## How is sentiment measured? The usual approach is to collect every answer that mentions the brand, then classify how the brand is described in each one. Good analysis goes beyond a single score and groups the descriptions into themes, such as pricing, ease of use, customer support or reliability. Themes tell you what to do next. A score on its own doesn't. Read a sample of the actual answers too. Automated classification is helpful at volume, but it can miss sarcasm, mixed statements and context. ## Example A hypothetical meal kit company, FreshCrate, is mentioned in 90 answers. Sentiment is positive in 52, neutral in 26 and negative in 12. Nearly all the negative mentions share one theme: "hard to cancel the subscription". That theme traces back to two old forum threads and a review site page. FreshCrate has since simplified cancellation, so the agency updates the help page with a clear "How to cancel" section and asks the review site to note the change. ## How AI Visibility Desk measures it AI Visibility Desk analyzes sentiment in every answer that mentions the brand across six engines, groups it into themes, and includes it in the white-label [AI visibility report](/ai-visibility-report). Combined with [mention rate](/glossary/mention-rate) and [brand mentions](/glossary/brand-mentions), it shows both how often and how well a brand is described. ## Zero-click search URL: https://aivisibilitydesk.com/glossary/zero-click-search A zero-click search is one where the user gets what they need from the search results page, or from an AI-generated answer, and leaves without clicking through to any website. The term was in use well before AI search. Featured snippets, knowledge panels, weather boxes, calculators and local map packs had already been answering many queries directly on the page. ## How does AI search affect zero-click searches? AI features extend the same pattern to more complex questions. [Google AI Overviews](/glossary/ai-overviews) and AI Mode summarize several sources into one answer. Assistants like ChatGPT, Perplexity, Gemini and Claude answer in a chat, where links are shown as sources rather than as the main result. For many informational questions, the user can stop there. This doesn't mean clicks disappear. People still click to buy, book, compare prices, read reviews in full or check details. But for top-of-funnel questions, more of the decision is shaped before any visit happens. ## What should marketers measure instead of clicks? If fewer people click, traffic alone undercounts how often a brand reaches buyers. Useful additions: - **Mentions in answers:** is the brand named when buyers ask category questions? See [mention rate](/glossary/mention-rate). - **Citations:** are the brand's pages used as sources? - **Description quality:** what does the answer actually say about the brand? - **Branded search and direct traffic:** buyers who heard the name in an AI answer may search for it later. ## Example A hypothetical garden centre, Greenacre, sees fewer visits to its "when to plant tulip bulbs" guide, while the page's search impressions hold steady. Checking AI Overviews shows the guide is cited and Greenacre is named in the summary. The visit drop is partly zero-click behavior. The agency reports the citation and mention as visibility the client is still getting, and adds a clear call to action on product pages that do receive clicks. ## How AI Visibility Desk helps AI Visibility Desk measures the part of visibility that click data misses: how often a brand is named, where it appears in the answer, and which of its pages are cited, across six AI engines. The [AI search visibility guide](/blog/ai-search-visibility) explains how to report on it.