Evidence review · AI visibility

How ChatGPT, Gemini and Perplexity decide which businesses to recommend

Nobody outside OpenAI, Google and Perplexity knows the exact rules. But the companies document how their assistants search, and large studies show which signals go with being named. This is what that evidence says, and the order we work through it for a business.

They search first, then write

The most useful thing to understand about AI assistants is that, for questions about businesses, they do not answer from memory alone. They search, read what comes back, and write an answer from it. The companies say so in their own documentation:

  • ChatGPT search rewrites a question into narrower search queries, sends them to search partners, and can send more specific follow-up queries after reading the first results. For local questions it may pass along your approximate location (OpenAI).
  • Google AI Overviews and AI Mode may use “query fan-out”: many related searches across subtopics at once. A page can only appear as a supporting link if it is indexed and eligible to show a snippet in Google Search (Google).
  • Gemini can be grounded in Google Search, meaning content from the Search index is supplied to the model when it answers (Google).
  • Perplexity runs its own crawler, PerplexityBot, to find and link websites in its answers (Perplexity).

The consequence is simple. If a business is not in the search results an assistant draws on, it cannot be recommended, however good it is. The first job is always to be findable and readable, which is what our guide on whether your website is blocking ChatGPT covers.

What the evidence says gets a business named

Being findable gets you into the pool. Three pieces of research help explain who gets picked from it.

StudyWhat it measuredMain finding
Yext, Oct 20256.8 million citations in answers from Gemini, OpenAI and Perplexity, July to August 2025, in retail, financial services, healthcare and food service44% of citations came from businesses' own websites, 42% from listings, 8% from reviews and social, 6% from news, forums and other sources
Ahrefs, Dec 2025How 75,000 brands' visibility in ChatGPT, Google AI Mode and AI Overviews related to web signalsYouTube mentions (about 0.74) and branded web mentions (0.66 to 0.71) correlated most strongly; Domain Rating correlated weakly (0.27 to 0.33)
Aggarwal et al., KDD 2024Experiments changing page content and measuring visibility in generative engine answersAdding citations, quotations and statistics improved a source's visibility by up to about 40% in their benchmark

Listings carry as much weight as websites. In the Yext data, a business’s listings (its profiles on maps, directories and booking platforms) were cited almost as often as its own website. Yext also reports that the assistants differ: Gemini favoured businesses’ own websites, OpenAI leaned heavily on third-party listings, and Perplexity mixed in review platforms and local pages. General forums such as Reddit made up only 2% of citations (Yext). Yext sells listings management, so read the framing with that in mind, but the source counts are hard to argue with.

Being talked about beats being “authoritative”. Ahrefs found that how often a brand is mentioned, on YouTube and across the web, tracked AI visibility far more closely than Domain Rating, the link-based authority score SEO tools use (Ahrefs). These are correlations across big brands, not proof of cause, and a local business will not have 75,000 brands’ worth of mentions. The direction still matters: links alone will not get you named.

Specific, sourced content gets quoted. The GEO paper is the most cited academic work on the topic. Its finding that statistics, quotations and citations raise visibility matches what assistants are built to do: they look for a sentence that answers the question and a source that backs it up (Aggarwal et al.).

Four questions an assistant has to answer about you

Putting the documentation and the studies together, an assistant has to be able to answer four questions before it will recommend a business. We use them as the structure of every AI visibility project.

  1. Can it find you? Your pages are crawlable, indexed in Google and Bing, and readable without JavaScript.
  2. Can it tell who you are? Your name, address, phone number, services and areas are identical on your website, Google Business Profile, Bing Places, Apple Business Connect and the directories in your industry. Conflicting facts are the most common reason an assistant describes a business wrongly.
  3. Is there evidence? Reviews, case studies, awards, press and mentions on sites that are not yours. This is what the Ahrefs and Yext numbers point at.
  4. Does your content answer the question? Pages that state plainly what you do, for whom, where, at what kind of price, with numbers and sources, so there is a sentence worth quoting.

Blog posts only work on the fourth question. They cannot fix the second and third, which is where most of the evidence above points.

Where we started: our own brand

We run the same checks on Jeedesigns that we run for clients. On October 12, 2026, Bing returned no results at all for our brand name. Our website was fast, crawlable and full of structured data, but almost nothing outside it mentioned us, and our listings were incomplete. By the evidence above, that is a business an assistant has little reason to name.

So we are fixing it in the order this guide describes, starting with indexing and identical listings, and we will publish the results as we go. An agency that cannot make its own name visible should not sell you the service.

Test it yourself in 15 minutes

You do not need a tool to see where you stand. Pick five prompts a customer would really type, run each one in ChatGPT (with search on), Gemini, Perplexity and Google’s AI Mode, and write down what happens.

Prompt typeExampleWhat it tells you
Category and place"Best dental clinic in Dubai Marina" / "Web design company in Kochi" / "Emergency plumber in Austin, TX"Whether you are in the pool at all
Problem"My AC is leaking water, who can fix it today in Plano?"Whether you are linked to the problems you solve
Comparison"[Your business] vs [competitor]"Which sources the assistant trusts about you
Facts"What does [your business] do, and where is it based?"Whether your facts are right
Price"How much does [your service] cost in [city]?"Whose numbers get quoted

For each answer, record whether you were named, where in the answer, whether the facts were right, and which sources were cited. Use a private or logged-out window so past chats do not colour the result, and run each prompt more than once, because answers change between runs. Repeat the same prompts every month. The cited sources are the most useful column: they tell you exactly which listings, review sites and publications you need to be on.

What to do, in order

  1. Be findable. Allow the AI search crawlers, make sure your content is in the HTML, and verify your site in Google Search Console and Bing Webmaster Tools. Our AI crawler checker tests the technical side for free.
  2. Make your facts identical everywhere. One name, one address format, one phone number and one description of what you do, on your website and every listing. Remove or correct the old ones.
  3. Earn evidence. Ask happy customers for reviews on the platforms your prompt test showed being cited. Publish real case studies. Get mentioned by local press, industry associations and partners.
  4. Publish answers worth quoting. Clear service pages, prices or price ranges where you can, and guides with numbers and sources.
  5. Measure monthly with the same prompt set, and change one thing at a time so you can see what worked.

What we would not do: buy reviews, list the business on hundreds of low-quality directories, or hide text on pages that only AI is meant to read. Each of these creates the conflicting or untrustworthy signals that keep a business out of answers, and fake reviews break the rules of the review platforms themselves.

If you want this done for you, our AI brand visibility service follows exactly this order.

Sources

  1. ChatGPT search, OpenAI Help Center
  2. AI features and your website, Google Search Central
  3. Google's common crawlers (Google-Extended), Google Search Central
  4. Perplexity crawlers, Perplexity
  5. AI Doesn't Rank, It Cites. And 86% of Its Sources Are Brand-Managed, Yext, Oct 2025
  6. AI brand visibility correlations (75,000 brands), Ahrefs, Dec 2025
  7. GEO: Generative Engine Optimization (Aggarwal et al., KDD 2024), arXiv

Every number in this guide links to one of these sources or comes from our own work. Read how we research and update guides.

FAQ

Questions readers ask

No AI company publishes a timeline. Access fixes are fast: OpenAI says ChatGPT search adjusts to a robots.txt change in about 24 hours. Listings, reviews and outside mentions take months to build, and they are what most of the evidence points to.
Not for Google. Google says there are no special files, markup or schema needed to appear in AI Overviews or AI Mode. llms.txt is cheap to add and does no harm, and structured data helps machines read your facts correctly, but neither replaces being findable and mentioned.
Usually because the web disagrees about you: an old address on a directory, a past service you no longer offer, or a business with a similar name. Assistants combine what they find. Fix the facts at the source, on every listing and profile, and keep them identical.
It helps, because assistants search the web before they answer and Google's own AI features are built on its index. But the studies show mentions, listings and reviews matter as much as rankings. A business can rank for its name and still be left out of a 'best in my city' answer.
Next step

Want this done for your business?

Tell us what you are trying to fix. We reply within one business day with honest advice and a fixed quote if there is work to do.

Get a fixed quoteChat on WhatsApp
  • Fixed quote within one business day
  • Free review, no obligation
  • You own everything we build
Chat