Skip to content

How Do AI Tools Like ChatGPT and Google AI Overviews Choose Which Websites to Cite?

Artificial intelligence (AI) answer tools mostly cite pages they can already find through search. Google says a page only needs to be indexed and eligible to show a snippet to appear as a link in AI Overviews or AI Mode, and that no special files or tricks are needed. ChatGPT search and Microsoft Copilot also show source links. The best thing a local business can do is be easy to crawl, accurate, and clearly the best answer.

When someone asks an AI tool a question, it often runs a search in the background, reads some of the results, and writes an answer. It then shows links to some of the pages it used.

  • Google AI Overviews and AI Mode use Google’s normal search index and ranking systems. Google’s guide to generative AI features says the usual search engine optimization (SEO) best practices still apply.
  • ChatGPT search shows citations and a “Sources” list. OpenAI says it sometimes sends rewritten searches to partner search providers, and that any public site can appear if its search crawler is allowed.
  • Microsoft Copilot uses Bing to ground its answers and shows citations (see 19.13.3 Bing, IndexNow, and AI Search).

Say a homeowner asks, “Who does emergency furnace repair near Coeur d’Alene?” The tool searches, finds local business pages and listings, and builds an answer. A heating company with a clear service page, an accurate Google Business Profile (GBP), and good reviews gives the tool something solid to cite.

More people now ask AI tools instead of clicking through ten links. If your pages can’t be crawled, aren’t indexed, or don’t clearly answer the question, you won’t be one of the sources.

There’s also a lot of bad advice out there. Google’s AI guide says you can ignore tactics like breaking content into tiny chunks or creating special AI text files. Paying someone for “AI optimization” tricks can waste money and may break spam rules.

Every business with a website. It matters most for businesses whose customers ask questions before they buy, like contractors, clinics, salons, and professional services.

  1. Make sure search engines can crawl you. Google’s AI features page says to allow crawling in robots.txt and at your host or content delivery network (CDN). See 19.5.1 Crawl Access and robots.txt.
  2. Make sure your important pages are indexed. Check Google Search Console and Bing Webmaster Tools (see 19.5.2 Index Control).
  3. Allow the AI search crawlers you want to appear in. For example, OpenAI says not to block OAI-SearchBot if you want to show in ChatGPT search (see 19.13.2 AI Crawler Access).
  4. Put important content in text. Google lists this as a best practice. Don’t hide prices, service areas, or hours inside images.
  5. Answer real questions clearly. Use the pattern in 19.9 Writing Answer-First Content.
  6. Add first-hand detail. Google’s AI guide stresses original content based on real experience. Your job photos, your process, and your local knowledge are what generic pages lack.
  7. Keep your business facts accurate everywhere. Google tells businesses to keep Business Profile info up to date. See 19.7 Entity Consistency.
  8. Make structured data match the visible page. Google says structured data isn’t required for AI features, but if you use it, it must match what’s on the page (see 19.6.1 Local Business Structured Data).
  9. Check what the tools say. Every month or so, ask a few AI tools the questions your customers ask. Note who gets cited (see 19.13.6 Tracking Traffic from AI Tools).
  • Dated screenshots of AI answers for your top customer questions, showing which sources were cited
  • Your robots.txt file and a record of which crawlers you allow
  • Search Console and Bing Webmaster Tools screenshots showing your key pages are indexed

Keep these in your website records folder and add a new set each quarter.

  • Blocking crawlers by accident. Some security tools and hosts block bots by default. Check your CDN or firewall settings, not just robots.txt.
  • Buying “AI SEO” packages that promise placement. No one controls what AI tools cite.
  • Rewriting every page into tiny chunks. Google says there’s no requirement to do this and no ideal length.
  • Making hundreds of near-copy pages for every question. Google warns that mass-producing pages for query variations is scaled content abuse.
  • Ignoring Bing. Copilot runs on Bing, and Bing has its own index (see 19.13.3).
  • Testing once and assuming. AI answers change from day to day and person to person. Look for patterns over time.

An SEO specialist: if your pages aren’t indexed, or you show up in normal search but never in AI answers. Questions to ask:

  • Are any of my important pages blocked from crawling or missing from the index?
  • What first-hand content is missing from my top service pages?
  • What will you do that Google’s own guidance doesn’t already cover?

[Bill: add a short story or lesson here. For example, the first time you saw a client’s business cited in an AI answer, or an “AI optimization” offer a client asked you about and what you told them.]

If you’d rather have this done for you, SiteSmarty builds websites and Google Business Profiles that are ready for search and AI answers.

  • [Date I did this:]
  • [What I actually did, step by step: which questions I tested, which AI tools I used, and what changed on the site]
  • [What it cost and how long it took:]
  • [What surprised me or what I’d do differently:]
  • [Screenshot or photo to add, with private info blurred: dated AI answers showing which sources were cited for a SiteSmarty question]
  • [Result so far:]

Do I need special optimization for AI Overviews? No. Google says there are no extra requirements. The page must be indexed and eligible to show a snippet.

How does ChatGPT search find websites? OpenAI says ChatGPT may send searches to partner search providers and uses its OAI-SearchBot crawler. Any public site can appear if that crawler isn’t blocked.

Does Bing Copilot cite websites? Yes. Copilot shows citations, and its answers are grounded in Bing’s search results.

Is GEO different from SEO? Generative engine optimization (GEO) is a newer name for the same basic work. Google’s AI guide says optimizing for its AI features is still SEO.

Can I pay to show up in AI answers? Not in the organic answer. Ads are separate and labeled. Be wary of anyone who promises AI placement.

  • Question: Can search and AI crawlers reach and index the business’s key pages, and do those pages clearly answer customer questions with first-hand detail?
  • Pass looks like: Key pages indexed in Google and Bing. AI search crawlers allowed. Important facts in text. Clear answers with real examples. Business facts match across listings.
  • Fail looks like: Pages blocked by robots.txt or a firewall. Key facts only in images. Thin or copied pages. Paid “AI placement” schemes.
  • Evidence to collect: robots.txt file, Search Console and Bing Webmaster Tools index screenshots, and dated AI answer screenshots for the top customer questions.

Where this fits

Do this first: 19.5.1 Crawl Access and Robots.txt, 19.5.2 Index Control, 19.9 Writing Answer-First Content

Do this next: 19.13.2 AI Crawler Access

Related pages: 14.1.1 Business Name, 19.7 Entity Consistency, 19.8 EEAT Signals, 19.13.3 Bing, IndexNow, and AI Search, 19.13.4 Brand and Entity Signals for AI, 19.13.6 Tracking Traffic from AI Tools

Sources

This page explains general rules and Bill's own experience. It is not legal, tax, or insurance advice. Talk to a qualified professional about your situation.

Get free updates

I'll send you a short email when new guide pages and videos go up.

No sales pitches. Unsubscribe any time.