ChatGPT vs Perplexity vs Google AI Overviews: how each cites sources

By Roshan, founder of SearchMaxx - Sep 28, 2026 - updated Oct 18, 2026 - 6 min read

ChatGPT search, Perplexity and Google AI Overviews all search the web and cite the pages they use, but each relies on a different crawler and a different index. ChatGPT needs OAI-SearchBot, Perplexity needs PerplexityBot, and Google AI Overviews use the normal Googlebot and Google's search index. Allow all three, make sure your text is in the HTML without JavaScript, and measure each one separately, because only Google gives you a dashboard.

Everything below comes from each company's own documentation or a published study, linked inline. Where a company does not document something, we say so instead of guessing.

The short version

ChatGPT search

Crawler and index. OpenAI's crawler docs say "OAI-SearchBot is for search." It is the bot that surfaces sites in ChatGPT's search features. OpenAI's ChatGPT search help page and launch post say search also draws on third-party search providers and content from OpenAI's partners. OpenAI does not name the providers on those pages, so we will not either.

Citations. Answers that used search may include inline citations. A Sources panel lists related links, including some that were not cited in the text. OpenAI's help page says placement is not guaranteed and that ChatGPT ranks results using several factors meant to surface relevant, reliable information.

robots.txt. Three bots, three jobs:

• OAI-SearchBot: OpenAI says sites that opt out "will not be shown in ChatGPT search answers, though can still appear as navigational links."
• GPTBot: training only. "Disallowing GPTBot indicates a site's content should not be used in training generative AI foundation models."
• ChatGPT-User: visits a page when a user asks. It "is not used to determine whether content may appear in Search," and because a user started the visit, "robots.txt rules may not apply."

"Each setting is independent of the others," so you can block training and still allow search. OpenAI says changes take about 24 hours to apply. It also asks sites to allow its published IP ranges, so check your firewall, not only robots.txt.

Measuring it. OpenAI's publishers FAQ says ChatGPT adds utm_source=chatgpt.com to referral links. Filter for it in your analytics. That counts clicks, not mentions. To count mentions, ask ChatGPT your buyers' questions and record whether you are named.

Perplexity

Crawler and index. Perplexity's crawler docs say PerplexityBot is "designed to surface and link websites in search results on Perplexity" and "is not used to crawl content for AI foundation models." Its help center says Perplexity searches the web for each question.

Citations. Per that help page, "each response includes citations and links to original sources." Citations are the core of the product, not an extra panel.

robots.txt. Two bots:

• PerplexityBot: allow it to appear in results. Perplexity also recommends allowing its published IP ranges, listed at a JSON endpoint on its docs page, and gives example firewall rules for Cloudflare and AWS.
• Perplexity-User: fetches a page a user asked about. "Since a user requested the fetch, this fetcher generally ignores robots.txt rules."

Perplexity says changes can take up to 24 hours. There is no separate training bot to block, because Perplexity says neither bot is used for training.

Measuring it. Perplexity does not document a referral tag. Look for visits referred by perplexity.ai in your analytics, and ask Perplexity your buyers' questions to count mentions. Perplexity is also the engine the GEO paper tested live. Adding quotations from relevant sources raised a page's visibility there by 22%, and keyword stuffing made it about 10% worse.

Google AI Overviews

Crawler and index. Google's AI features guide says a page must be indexed and eligible to show a snippet in Search to be a supporting link. "There are no additional technical requirements." The crawler is the normal Googlebot. Overviews may use "query fan-out", running several related searches across subtopics, which lets them show a wider set of links than one classic search.

Citations. Overviews show supporting links. Pew Research analyzed 68,879 Google searches from 900 US adults in March 2025. Of those, 18% produced an AI summary. 88% of the summaries cited three or more sources. Users clicked a link inside the summary on 1% of visits, and clicked a regular result on 8% of visits with a summary, against 15% without one.

robots.txt. This is where Google differs most. There is no separate AI Overviews bot. Google's crawler docs say Google-Extended controls Gemini training and grounding, and that it "does not impact a site's inclusion in Google Search nor is it used as a ranking signal in Google Search." To limit what Overviews show from your page, Google points to nosnippet, data-nosnippet, max-snippet and noindex. Those controls apply to your regular search snippets too.

Measuring it. Google says traffic from AI features is "included in the overall search traffic in Search Console", under the Web search type. You get the clicks and impressions, but you cannot split out AI Overviews from regular results.

What the differences mean in practice

JavaScript. Vercel's study of AI crawlers found that OpenAI's and Perplexity's crawlers do not render JavaScript, while Google's crawling does. A page that builds its text in the browser can show up in AI Overviews and be invisible to ChatGPT and Perplexity. We explain how to check in AI crawlers don't run JavaScript.

Firewalls. OpenAI and Perplexity both ask you to allow their published IP ranges. Bot protection that blocks unknown crawlers can remove you from both while Google works fine.

Training and search are separate for OpenAI and Google. You can opt out of GPTBot and Google-Extended and keep full search visibility. With Perplexity there is nothing to opt out of, because its docs say neither bot trains models.

What to do for all three

1. Allow the search crawlers. Your robots.txt should allow OAI-SearchBot, PerplexityBot and Googlebot. Our robots.txt guide has a file you can copy.
2. Check the firewall. Test with each bot's user agent and confirm you get a 200, then allow the IP ranges OpenAI and Perplexity publish.
3. Put your text in the HTML. Fetch your key pages with curl and count the words. If your product description or pricing only appears after JavaScript runs, ChatGPT and Perplexity cannot use it.
4. Get indexed in Google. It is the only requirement for AI Overviews, and a good sign your pages are findable elsewhere.
5. Write pages AI answers can cite. Answer one question per page, state it early, and back it with numbers, quotes and named sources. The GEO paper found those additions raised visibility, and keyword stuffing lowered it.
6. Measure each engine separately. Filter utm_source=chatgpt.com and perplexity.ai referrals in your analytics, use Search Console for Google, and run the same set of buyer questions in all three each week.

Verdict

You do not need three strategies. The same three steps cover all of them: allow the crawlers, serve real HTML, and write pages worth citing. Then measure each engine on its own, because their numbers live in different places.

To see whether each AI crawler can reach and read your site today, run our free check. If you are still sorting out the terms, SEO vs GEO vs AEO explains how they overlap.

Want this done for your site?

Every day we make one change to your website that helps AI engines name you, and a person checks each one. Email me your website and the question your buyers ask AI.

Email Roshan to get started

Or check your site free: