How to get your business recommended by Gemini, Claude, Perplexity and Copilot (not just ChatGPT)
Each AI assistant answers from a different search index: Gemini and AI Overviews from Google, Copilot from Bing, Claude from Brave Search and its own crawler, Perplexity from its own index. Where each one gets its answers, the crawlers to allow, what to fix first for each, what paid options exist, and how to measure whether they mention you.
There is no single “AI index”. Gemini, Google's AI Overviews and AI Mode answer from the Google Search index; Copilot from Bing; Claude from Brave Search and its own crawler; Perplexity from its own index; ChatGPT from third-party providers such as Bing plus its own crawler. So the work is: be indexed in Google and Bing, let each assistant's search crawler read your site, answer your buyers' questions on your own pages, and earn mentions on the sites the assistants cite. Nobody sells a place in the answer itself.
Most advice about AI visibility is written about ChatGPT, and we have covered it in how to get ChatGPT to recommend your business. Your customers use the others too, and each one reads the web through a different door. This guide maps the doors, using what each company documents itself, and gives the first thing to fix for each.
Where each assistant gets its answers
| Assistant | Where live answers come from | Crawler to allow | Fix this first |
|---|---|---|---|
| Gemini app | The Google Search index, at the moment you ask | Googlebot; do not block Google-Extended | Google indexing; Business Profile; Merchant Center for products |
| Google AI Overviews and AI Mode | Google Search's ranking systems | Googlebot | Pages indexed and eligible for a snippet; the Search Console AI setting left off |
| Microsoft Copilot | The Bing search service | Bingbot | Bing Webmaster Tools, sitemap, IndexNow; Bing Places for local |
| Claude | Brave Search, plus Anthropic's own search crawler | Claude-SearchBot and Claude-User | Being in Brave Search; not blocking Anthropic's search bots |
| Perplexity | Its own index of the web | PerplexityBot and Perplexity-User | Crawlable, fast pages; third-party sources that describe you |
| ChatGPT | Third-party providers such as Bing, plus OpenAI's crawler | OAI-SearchBot and ChatGPT-User | Bing indexing; see our ChatGPT guide |
Google: Gemini, AI Overviews and AI Mode
Google describes Gemini's grounding as content from the Google Search index, retrieved when you ask, and its AI Overviews and AI Mode run on the same ranking systems as Search. Its documentation for site owners asks for one thing: pages that are indexed and eligible to show with a snippet. Its AI optimization guide (July 2026) treats optimizing for AI answers as part of SEO, recommends unique, first-hand content, and points businesses to Merchant Center for products and Google Business Profile for local details.
Three Google-specific details catch people out:
- Google-Extended controls the Gemini app, not just training. Google's crawler documentation says the Google-Extended token governs use of your content for Gemini training and for grounding in the Gemini app and Vertex AI, while not affecting inclusion in Google Search. Block it and Gemini stops using your pages in its answers.
- There is now an opt-out switch in Search Console. Since 31 August 2026 a “Search generative AI” setting removes a site from AI Overviews and AI Mode worldwide. Check that nobody on your team has turned it on.
- Ranking on page one is no longer the whole story. Ahrefs found that only 37.9% of URLs cited in AI Overviews ranked in the top 10 for the query in early 2026, down from about 76% in July 2025, after the switch to a newer Gemini model. Pages further down, including detailed answers to narrow questions, get cited.
Microsoft Copilot: it is Bing
Microsoft says Copilot may fetch information from the Bing search service with queries it generates, so a page Bing has not indexed cannot be cited. Bing gives you more tools than anyone else here:
- Verify your site in Bing Webmaster Tools (you can import it from Google Search Console) and submit your sitemap.
- Use IndexNow, which Bing supports, to notify it of new and changed pages the moment they go live.
- Check the AI Performance report, in preview since February 2026: it shows how often Copilot and Bing's AI features cite your pages and the queries behind them.
- For a local business, claim Bing Places; for a store, Microsoft Merchant Center feeds inform Copilot's shopping results.
Bing's webmaster guidelines now cover Copilot grounding by name and treat attempts to manipulate AI answers, such as hidden prompt text, as abuse. Bing matters twice: OpenAI says ChatGPT search uses third-party search providers such as Bing, so the same work helps there.
Claude: Brave Search and Anthropic's own crawler
When Claude searches the web, the results come from Brave Search: Anthropic lists Brave as the subprocessor for web search, which TechCrunch reported in March 2025 after finding Claude's citations matched Brave's results. Anthropic also runs three crawlers, described in its help center:
- ClaudeBot collects training data. Blocking it opts you out of training only.
- Claude-SearchBot indexes pages to improve search results. Anthropic says blocking it may reduce your visibility in Claude.
- Claude-User fetches a page when a user asks Claude to read it. Blocking it has the same effect.
Brave builds its own index rather than renting Google's or Bing's, so the practical steps are the usual ones: no login walls or heavy scripts hiding your main content, a sitemap, and pages linked from places that are already well indexed. Search your brand on search.brave.com to see what Claude is likely to find. Anthropic has said Claude will remain ad-free; there is nothing to buy.
Perplexity: its own index, and citations everywhere
Perplexity says its answers draw on its own index of hundreds of billions of webpages, built by PerplexityBot. Its crawler documentation separates PerplexityBot, which surfaces and links sites in results and is not used for training, from Perplexity-User, which fetches pages when someone asks and generally ignores robots.txt. Allow both if you want to be cited.
Perplexity shows its sources more prominently than any other assistant, so the pages that describe you elsewhere carry a lot of weight: reviews, news coverage, comparisons and discussion threads. Online stores can share product data through the free merchant program it launched in 2024. Perplexity phased out advertising and describes itself as being in the “accuracy business”.
A robots.txt that keeps every door open
Most sites need no AI-specific rules: an open robots.txt lets every search crawler in. If you want to keep your content out of model training but stay citable, block only the training crawlers and leave the search and user ones alone:
# Search and user-request crawlers: allowed (being cited depends on them)
User-agent: OAI-SearchBot
User-agent: ChatGPT-User
User-agent: Claude-SearchBot
User-agent: Claude-User
User-agent: PerplexityBot
User-agent: Perplexity-User
Allow: /
# Training only: block if you choose to
User-agent: GPTBot
User-agent: ClaudeBot
Disallow: /Leave Google-Extended out of that list if you want Gemini to use your pages, and check your CDN or firewall too: a bot-protection rule can block a crawler that robots.txt allows. OpenAI's crawler page and the others linked above list the user agents and IP ranges.
What every assistant rewards
Pages that answer the question asked
People ask assistants full questions: “which press release service has wholesale pricing for agencies”, “is this supplement legit”. A page that states the answer in its first lines, then covers the follow-up questions, gives every assistant something to quote. Google says no special markup or AI rewriting is needed; clear headings, tables and FAQs help readers and machines alike.
Mentions you do not control
Across more than 25 million links cited by ChatGPT, Claude and Gemini, Muck Rack found earned media made up 84% of citations and paid or advertorial content 0.3%. Semrush's analysis of 230,000 prompts found Reddit and Wikipedia the two most cited domains in ChatGPT, with shares that swung sharply within weeks, and AI Mode leaning on LinkedIn and Google's own sites. The steady part is the mix: news and trade coverage, reviews, community discussion and comparisons. A labeled press release on a known outlet anchors your facts, as how press releases get brands cited by AI explains, but the recommendation comes from what others say.
The same facts everywhere
Name, category, location, prices and what you do should read the same on your site, your Business Profile, Bing Places, your directory listings and your social profiles. Assistants combine sources; contradictions make them hedge or leave you out.
What you can and cannot buy
- ChatGPT began testing ads in the US on 9 February 2026 for logged-in adults on the Free and Go plans; OpenAI says ads do not influence answers.
- Google shows ads above, below and inside AI Overviews, drawn from ordinary Search, Shopping and Performance Max campaigns, and is expanding ads in AI Mode.
- Copilot carries Microsoft Advertising ads and, in the US, checkout for merchants through PayPal, Stripe and Shopify.
- Perplexity and Claude carry no ads.
An ad beside the answer is labeled as an ad. Being named inside the answer is not for sale anywhere, which is why third-party mentions and indexed pages are the whole game.
How to measure it
- Write down ten questions your buyers ask, in their words.
- Ask each assistant every question three times, logged out or in a fresh chat, and record whether you are named and which pages are cited. One answer proves nothing; the method is in our guide on checking ChatGPT mentions.
- Add the first-party reports: Bing Webmaster Tools' AI Performance (Copilot), Search Console's generative AI report (impressions in Google's AI features), and visits tagged utm_source=chatgpt.com in analytics.
- Repeat monthly. Count how often you appear, not where.
The full manual method, including the personalization traps, is in how to check if ChatGPT mentions your brand.
The part of this work we do is the third-party record: labeled press releases and sponsored articles on news sites that are indexed in Google and Bing, which assistants can read and cite. Every outlet in the directory shows its label and link type, and you can ask for a recent live sample to check that the page is indexed before you order.
How do I get my business recommended by Gemini?
Be indexed in Google Search with pages that answer your customers' questions, because Gemini grounds its answers in the Google Search index. Keep your Google Business Profile and, for products, your Merchant Center feed accurate, earn mentions on sites Google already trusts, and do not block Google-Extended in robots.txt: it controls grounding in the Gemini app as well as training.
Where does Claude get its information when it searches the web?
Anthropic's list of subprocessors names Brave Search for web search, and Anthropic also runs its own search crawler, Claude-SearchBot. To be findable, allow Claude-SearchBot and Claude-User in robots.txt and make sure your pages are in Brave Search; blocking ClaudeBot only opts you out of training.
How do I get my business on Perplexity?
Perplexity answers from its own web index, built by PerplexityBot, so allow that crawler and keep your key pages fast and crawlable. It cites sources heavily, so third-party pages that describe you accurately (reviews, news, comparisons, Reddit threads) matter as much as your own site. Online stores can join Perplexity's free merchant program to share product data.
Does Microsoft Copilot use Bing?
Yes. Microsoft says Copilot may fetch information from the Bing search service to answer, so Bing indexing decides what Copilot can cite. Verify your site in Bing Webmaster Tools, submit your sitemap, use IndexNow for new pages, and check its AI Performance report, which shows how often Copilot cites your pages.
Can I pay to be recommended by AI assistants?
Not in the answers themselves. OpenAI says ads in ChatGPT do not influence its answers, Google's ads in AI Overviews and AI Mode come from ordinary Search and Shopping campaigns and are labeled, Perplexity phased out its ads, and Anthropic says Claude will remain ad-free. You can pay for ads beside the answers on some of them, and for placements on sites the assistants read, but the recommendation itself has to be earned.
Do I need an llms.txt file for AI assistants?
No. Google's AI optimization guide says Google Search does not use llms.txt or similar AI text files, and its staff have called the idea speculative. None of the other assistants has said it uses one. Clear, crawlable pages and a normal sitemap do the job.
Should I block AI crawlers?
Block training crawlers if you want to; never block the search ones if you want to be recommended. The search and user-request crawlers (OAI-SearchBot, ChatGPT-User, Claude-SearchBot, Claude-User, PerplexityBot, Perplexity-User, Bingbot and Googlebot) are what let assistants read and cite your pages. Google-Extended is the exception to remember: blocking it keeps your pages out of Gemini app answers.
How do I know whether AI assistants recommend my business?
Ask each assistant the questions your customers ask, several times, and record how often you appear and which pages are cited; answers vary from run to run. Then add the first-party data: Bing Webmaster Tools' AI Performance report for Copilot, Search Console's generative AI report for impressions in Google's AI features, and visits tagged utm_source=chatgpt.com in your analytics.
Read next
Ready to be published where it counts?
Pick your publications, upload a draft and have verifiable coverage live within days.
No subscriptions · Live links · Refund if not published