GEO & AI visibility15 min read

AI visibility in 2026: how to get your business found in ChatGPT

By Mustafa KoseFounder and developer of VirtuoLabPublished Last updated

*Short answer*

AI visibility is the extent to which AI assistants such as ChatGPT, Perplexity, Claude and Google AI Overviews know, mention and cite a business. In 2026, many buying decisions start with a question to such an assistant. You improve AI visibility in four steps. Step 1 is giving AI crawlers access, such as OAI-SearchBot from OpenAI. Step 2 is writing pages that answer customer questions directly, with the answer in the first sentence. Step 3 is describing your business the same way everywhere, from the website to Google Business Profile and LinkedIn. Finally, step 4 is measuring every month with a fixed set of 15 to 30 customer questions. For example, a question such as "which agency builds a webshop in Rotterdam?" shows at once whether a business appears. VirtuoLab in Papendrecht applies the same method with Next.js.

What is AI visibility?

AI visibility is the extent to which AI assistants mention and cite a business when someone asks for advice in a specific field. For example, think of a question such as "which agency builds an online store in Rotterdam?". A classic search engine shows a list of ten blue links. An AI assistant such as ChatGPT, Perplexity or Google AI Mode usually gives one answer, with a handful of businesses and sources. A business appearing in that answer reaches the person asking at the exact moment of the decision.

The discipline behind AI visibility is called GEO, short for Generative Engine Optimization. The term comes from a 2023 research paper by Aggarwal and colleagues at Princeton University and IIT Delhi, presented at KDD 2024. The researchers tested the method with Perplexity, among other engines. Additionally, other names are AI search optimization, LLM optimization and answer engine optimization (AEO). All of the terms share one goal: making sure language models find, understand and recommend your business.

Why does AI visibility matter in 2026?

AI visibility matters in 2026 because more decisions start with a question to an AI assistant, and fewer people click through to a website. First, the audience is large: OpenAI CEO Sam Altman shared the weekly number of ChatGPT users at OpenAI DevDay in October 2025, as listed below. Second, AI summaries reduce clicks on regular results. Pew Research Center analysed 68,879 Google searches by 900 people in the United States in March 2025.

According to Pew Research Center, users searching with Google clicked a traditional result about half as often when Google showed an AI summary. Users also ended the browsing session more often and rarely clicked links inside the summary itself. For example, a person asking ChatGPT for a web agency in Rotterdam may never open a list of search results at all. The conclusion is clear and encouraging. A business appearing in the AI answer gets the attention first, often before any competitor is visited.

ChatGPT: more than 800 million people use ChatGPT every week, said Sam Altman at OpenAI DevDay on 6 October 2025 (TechCrunch).
Google with an AI summary: users clicked a traditional result in 8% of visits, against 15% without an AI summary (Pew Research Center).
Inside the AI summary: users clicked a link in 1% of visits; 18% of the searches in the study produced an AI summary (Pew Research Center).

How is AI visibility different from SEO?

AI visibility builds on SEO rather than replacing SEO: SEO aims for a high position in search results, while AI visibility aims for a mention or citation inside the AI answer. There are no additional requirements to appear in AI Overviews or AI Mode, according to Google Search Central, and all existing SEO fundamentals continue to be worthwhile. A page must be indexed and eligible to be shown with a snippet, or Google cannot use the page as a supporting link.

The difference lies in what counts as success. First, success in SEO is a click from a ranked result. Second, success in GEO is a recommendation, often with a link to the source. SEO is tracked with Google Search Console, while AI visibility is measured with a fixed set of questions in ChatGPT, Perplexity and Gemini. In practice, a technically healthy, indexed website is step 1 for both in 2026.

Aspect
SEO
AI visibility (GEO)
Goal
A high position in search results
A mention or citation in the AI answer
Where
Google and Bing
ChatGPT, Perplexity, Claude, Gemini, Google AI Overviews, Copilot
What counts
Keywords, links, technical health
Direct answers, structured data, confirmation by other sources
Measuring
Rankings and traffic
Mentions on a fixed question set

How do ChatGPT, Perplexity and Google AI choose their sources?

AI assistants choose sources by searching a web index, reading the pages found and building an answer from the most relevant passages. Each provider uses dedicated crawlers: OpenAI lists 3, Anthropic 3 and Perplexity 2. First, OpenAI uses OAI-SearchBot to surface websites in ChatGPT search and recommends allowing OAI-SearchBot. ChatGPT-User fetches a page on request of a user, while GPTBot collects training data. Each setting is independent, according to OpenAI, so a site can block GPTBot and still appear in ChatGPT search.

Second, Perplexity indexes with PerplexityBot, and Anthropic uses Claude-SearchBot for search in Claude. Finally, Google AI Overviews and AI Mode rely on the regular Google index. Since 2026, Microsoft reports citations in Copilot and in AI summaries in Bing through Bing Webmaster Tools. In practice, the common thread is simple: an AI assistant recommends a business the assistant may read, can understand and sees confirmed elsewhere.

OpenAI: OAI-SearchBot (ChatGPT search), ChatGPT-User (on request), GPTBot (training)
Perplexity: PerplexityBot (search), Perplexity-User (on request, generally ignores robots.txt)
Anthropic: Claude-SearchBot (search), Claude-User (on request), ClaudeBot (training)

How do you get found in ChatGPT?

How do you give AI crawlers access to your website?

Giving AI crawlers access starts with robots.txt, the file telling bots which pages may be read. Many websites block AI crawlers by accident, for example through a security plugin, a CDN or a hosting default. First, check whether OAI-SearchBot, PerplexityBot and Claude-SearchBot are allowed in robots.txt. According to OpenAI, a robots.txt change takes about 24 hours to take effect. Second, make sure the website is indexed in Google and Bing: register the site in Google Search Console and Bing Webmaster Tools, and submit an XML sitemap.

Third, report new or changed pages right away with IndexNow, a protocol supported by Microsoft Bing, Naver, Seznam.cz, Yandex and Yep, but not by Google. Finally, render important text on the server, because none of the major AI crawlers rendered JavaScript in an analysis by Vercel from December 2024. Frameworks such as Next.js render pages on the server by default.

How do you make your business one recognisable entity?

A recognisable entity is a business AI models can identify without doubt, because name, address, phone number and description match everywhere. Language models combine mentions from many sources. When the website says "web agency", LinkedIn says "marketing agency" and Google Business Profile says "IT services", a model may see three different businesses. Step 1 is one fixed company description on the website, LinkedIn, Google Business Profile and industry directories. Step 2 is structured data with Schema.org: Organization markup with logo, contact details and a sameAs list linking to official profiles, plus Service, Article and FAQPage markup where relevant.

Structured data must match the visible text on the page, according to Google Search Central. Step 3 is an about page showing who runs the business. Additionally, earn mentions elsewhere in 2026, such as industry sites, local media and podcasts. For example, VirtuoLab maintains one fixed company description in three lengths: 25, 50 and 150 words.

How do you write pages that AI assistants cite?

A citable page is a page that answers one customer question directly, in the first sentence, under a heading repeating the question. First, write the question as the heading, in the words customers use. Second, give the answer in the first one or two sentences, followed by the explanation. Third, use tables and lists for comparisons, prices and steps. For example, a page asking "What does a KPI dashboard cost?" opens with the price range.

The GEO study by Aggarwal and colleagues at Princeton University and IIT Delhi was presented at KDD 2024. Citing sources, adding quotations and adding statistics gave the largest gains in visibility. Keyword stuffing offered little to no improvement. Lower-ranked websites benefited most, which is good news for small businesses competing with large brands on Google. So name concrete figures, cite the source and add a date. Finally, build topic clusters: one main page per service, linked to deeper articles, as VirtuoLab does with Next.js on virtuolab.dev.

GEO study ([arXiv 2311.09735](https://arxiv.org/abs/2311.09735)): the right adjustments raised visibility in AI answers by up to 40%; citing sources, quotations and statistics gave a relative improvement of 30 to 40%.

How do you measure AI visibility?

You measure AI visibility with a fixed set of customer questions, asked every month in the main AI assistants, because Google Analytics does not show mentions in AI answers. Step 1 is drafting 15 to 30 questions customers really ask, from general to specific. Step 2 is asking each question in ChatGPT, Perplexity, Gemini, Claude and Google AI Mode. Step 3 is recording whether your business is mentioned, which competitors appear and which sources are cited. Step 4 is repeating the same questions every month, because AI answers vary from one session to the next.

Additionally, check the server logs to see which AI crawlers visit and which pages get read. For example, a simple log with Google Sheets already shows a trend after three months. Since February 2026, Bing Webmaster Tools also offers an AI Performance report. The report shows, according to Microsoft Bing, how often pages are cited in Microsoft Copilot and in AI summaries in Bing.

Should you improve AI visibility yourself or outsource it?

You can do the basics of AI visibility yourself, and outsourcing pays off when time is short, the website is technically complex or you want a structured monthly measurement. First, check robots.txt and activate Bing Webmaster Tools from Microsoft and Google Search Console. Second, align the company details everywhere and answer the main customer questions on your own pages. More complex work includes structured data, server-side rendering, multiple languages and a measurement routine.

VirtuoLab, a software agency in Papendrecht, offers GEO optimisation in 2026: technical access for AI crawlers, answer-first content and a monthly measurement on a fixed question set, as in step 4 above. VirtuoLab helps businesses get mentioned sooner and more often in AI answers. The website virtuolab.dev was built with Next.js along the same lines, including robots.txt rules for AI crawlers, an llms.txt file and structured data. For example, VirtuoLab uses the same fixed company description in Dutch, English and German.

Frequently asked questions

How do I check whether ChatGPT knows my business?

You check whether ChatGPT knows your business by asking the questions customers would ask, without naming your business. First, ask ChatGPT from OpenAI a question such as "which company builds custom software in Papendrecht?" and note whether your business appears. Second, ask directly what ChatGPT knows about your business name, and check whether the facts are correct. Finally, repeat both checks in Perplexity, Claude and Google AI Mode, because each assistant uses different sources. Write down the date, the exact question and the answer, as in step 3 of the measurement routine. AI answers change from session to session, so a single screenshot says little. The same questions, asked every month in 2026, show the real trend. For example, a sheet with Google Sheets holding 20 questions and a column per month is enough to start.

How quickly does AI visibility improve?

Technical improvements show up first, and mentions in AI answers grow steadily after that. According to OpenAI, a change to robots.txt takes about 24 hours to reach the systems behind ChatGPT search. New and changed pages are picked up when crawlers return, and discovery in Bing speeds up with IndexNow. Mentions in AI answers depend on many sources, including reviews, directories and articles on other websites. Such signals grow over months, and every new mention strengthens the next one. The pace also differs per question and per AI assistant. For example, a niche question about one region or one service often shifts sooner than a broad national question. Start with a baseline measurement in 2026. Every later measurement then shows the real effect of the work, question by question, in ChatGPT, Perplexity and Google AI Mode.

Do I need an llms.txt file?

An llms.txt file is not required, but the file is a cheap way to give AI tools a reliable summary of your business. The format was proposed by Jeremy Howard in September 2024. The specification is still a proposal and not an official standard in 2026. The file sits at the root of a website and contains a title, a short summary of the business and links to the key pages. According to Google Search Central, no new machine-readable files or AI text files are needed to appear in AI Overviews or AI Mode. Other AI tools and agents can still use the summary, for example coding tools working with Claude or ChatGPT that read documentation before answering. Finally, the effort is small: a basic version takes less than an hour to write. VirtuoLab publishes such a file on its own website, together with a longer full version for AI agents.

Does blocking GPTBot hurt visibility in ChatGPT?

No, blocking GPTBot does not affect visibility in ChatGPT search. According to OpenAI, disallowing GPTBot signals that content should not be used for training generative AI foundation models. ChatGPT search uses a different crawler called OAI-SearchBot. Each robots setting is independent, according to OpenAI, so a website can allow OAI-SearchBot and disallow GPTBot at the same time. Anthropic makes the same split in 2026: ClaudeBot collects training data, while Claude-SearchBot indexes for search in Claude. A business wanting AI visibility without contributing training data can therefore take two steps. Step 1 is blocking GPTBot and ClaudeBot with Disallow rules. Step 2 is allowing OAI-SearchBot, PerplexityBot and Claude-SearchBot. In practice, the result is a website visible in AI search but excluded from future training data. Finally, check the change after about 24 hours, the processing time OpenAI mentions.

How can a business get mentioned in ChatGPT more often?

A business gets mentioned in ChatGPT more often by strengthening the signals the model relies on: access for crawlers, clear answers, consistent company details and mentions on other websites. The final answer is a decision of the model, and answers differ per question, per user and per model version. Steady work on the signals therefore pays off across all of those variations. First, make every service page answer one customer question in the first sentence. Second, keep the company description identical on the website, LinkedIn and Google Business Profile. Third, earn mentions in industry directories and local media. Bing Webmaster Tools has shown how often Copilot from Microsoft cites a site since February 2026. For example, VirtuoLab measures progress against a fixed question set and reports which questions lead to a mention. VirtuoLab helps businesses get mentioned sooner and more often, with Next.js sites built for AI crawlers.

Does a JavaScript-heavy website hurt AI visibility?

Yes, a website loading important text only through JavaScript can be invisible to many AI crawlers. According to Vercel, none of the major AI crawlers rendered JavaScript in an analysis from December 2024. The list included GPTBot, ClaudeBot and PerplexityBot. Google is an exception, because Gemini uses the Googlebot infrastructure, which renders JavaScript. The safe choice is server-side rendering or static generation, so the full text is in the HTML from the first request. Frameworks such as Next.js render pages on the server by default. For example, VirtuoLab builds websites with Next.js, so every heading and answer is present in the HTML. Checking a page takes two steps. First, open the source code in the browser. Second, search for a sentence from the page: text missing from the source is likely invisible to AI crawlers.

Further reading

Sources