Search Engine Optimization

How to get your business cited by ChatGPT, Perplexity and Google AI Overviews: a practical GEO guide

10 min readBy Davinder Kumar Meen

AI answer engines cite pages they can fetch, understand and quote. In October 2026 that means three things you control: let the right crawlers in, write pages with clear, sourced, quotable statements under a named author, and keep your structured data and business listings consistent with what the page says. Google's own guidance is that no special files or markup are needed for AI Overviews or AI Mode; the work is ordinary technical SEO done properly, plus a measurement routine that most businesses do not have yet.

Key takeaways

  • Google states there are "no additional requirements to appear in AI Overviews or AI Mode", and that you "don't need to create new machine readable files, AI text files, or markup" for them. Eligibility is the same as for a normal snippet: indexed, crawlable, snippet-eligible.
  • ChatGPT search, Perplexity and Claude each use a separate search crawler. Blocking a training crawler (GPTBot, ClaudeBot, Google-Extended) does not remove you from their search features; blocking the search crawler does.
  • In the Princeton/IIT study that coined "GEO" (KDD 2024), adding citations, quotations and statistics to a page raised its visibility in generative answers by roughly 30-40% on the paper's main metric. Keyword stuffing made it worse.
  • AI Overviews are growing on commercial queries: Semrush measured a 71% average increase on commercial-intent SERPs between November 2025 and April 2026, while transactional queries fell 5%. GA4, Search Console and a monthly prompt panel can measure your share today.

What "generative engine optimisation" means (and what it does not)

The term comes from a 2023 paper by Pranjal Aggarwal and colleagues (KDD 2024), which proposed GEO as "a black-box optimization framework for improving content visibility in generative engine responses" and found content changes could "boost visibility by up to 40%".

Two years on, the engines are mainstream: Google AI Overviews and AI Mode, ChatGPT with search, Perplexity, Microsoft Copilot, Gemini and Claude with web search. Every one of them does the same three things: fetch pages (via a crawler or a user-triggered fetch), select passages that answer the question, and attribute some of them with links.

GEO is therefore not a new discipline with secret markup. It is the overlap of technical SEO (can the engine get the page?), editorial quality (is there a sentence worth quoting?) and entity clarity (does the engine know who you are?). What follows is organised in that order.

Step 1: let the right crawlers in

Each vendor now documents separate user agents for training, search indexing and user-triggered fetches. The distinction matters because many sites blocked "AI bots" wholesale in 2023-2024 and quietly removed themselves from AI search answers.

Vendor Training crawler Search crawler User-triggered fetch Notes from the vendor's documentation
Google Google-Extended (robots.txt token, no separate user agent) Googlebot n/a Google-Extended "does not impact a site's inclusion in Google Search nor is it used as a ranking signal"; AI Overviews and AI Mode follow Googlebot's permissions
OpenAI GPTBot OAI-SearchBot ChatGPT-User "OAI-SearchBot is used to surface websites in search results in ChatGPT's search features"; sites that opt out "will not be shown in ChatGPT search answers, though can still appear as navigational links"; "Each setting is independent of the others"
Perplexity none documented PerplexityBot Perplexity-User PerplexityBot "is not used to crawl content for AI foundation models"; Perplexity-User "generally ignores robots.txt rules" because a user requested the page
Anthropic ClaudeBot Claude-SearchBot Claude-User All three honour robots.txt; Anthropic supports Crawl-delay and publishes crawler IPs

Practical rules:

  1. Allow every search crawler in robots.txt unless you have a legal reason not to. A business that wants enquiries wants to be in answers.
  2. Decide on training crawlers separately and write the decision down. Allowing them does not change your search visibility; blocking them is a policy choice, not an SEO one.
  3. Check that your CDN or bot-protection layer is not blocking these agents at the edge. Google's guidance specifically says to allow crawling "in robots.txt and through any CDN or hosting layer". A Cloudflare or WAF rule written in 2023 can override a permissive robots.txt.
  4. Confirm it with logs. Grep your CDN or server logs for OAI-SearchBot, PerplexityBot, Claude-SearchBot and Googlebot against the URLs you care about. If a page has not been fetched by any of them in 30 days, it is not going to be cited.

Google also notes that recrawling after a fix can take "several days to several months", so make these changes first.

Step 2: write pages that have something quotable in them

Answer engines extract passages. A page that circles its topic for 600 words before saying anything concrete gives them nothing to lift. The GEO study tested nine content changes on the same pages and measured visibility in generated answers. The ones that worked:

  • Cite sources, add quotations, add statistics: "a relative improvement of 30-40% on the Position-Adjusted Word Count metric and 15-30% on the Subjective Impression metric".
  • Fluency and easy-to-understand rewrites: a "visibility boost of 15-30%".
  • Authoritative tone on its own: "no significant improvement".
  • Keyword stuffing: fell below the unoptimised baseline (17.8 versus 19.5 on the paper's main metric). The authors conclude such methods "offer little to no improvement".

Translate that into an editing checklist for every important page:

  1. Open with the answer in one or two sentences, then explain. The first 60 words should survive being quoted on their own.
  2. Make claims specific and attributed. "Response time improved" is unquotable; "median response time fell from 4.1 s to 1.2 s after the migration (Lighthouse, mobile, 13 Oct 2026)" is.
  3. Put one fact per sentence where it matters. Engines quote sentences, not paragraphs.
  4. Use tables for comparisons and numbered lists for procedures. Both are extracted cleanly.
  5. Add a short FAQ with 40-60 word answers written as complete statements.
  6. Link to the primary source for every number you did not generate yourself.
  7. Date the page and show when it was last updated. Fast-moving topics are judged on freshness.

None of this licenses invented statistics: the engines that reward sourced numbers also surface the sources.

Step 3: make your entity unambiguous

Engines decide whether to cite you partly on whether they can resolve who you are. For a service business the signals are mundane and often missing:

  • An Organization (or ProfessionalService) schema node on every page with the legal name, founding date, addresses, phone, and sameAs links to LinkedIn and the directories where you genuinely have a profile.
  • A named author on every article, with a Person node, a job title, an author page and a matching LinkedIn profile. Anonymous "admin" bylines give AI engines nothing to attribute.
  • Consistent facts everywhere. If the home page says 50 staff and the services page says 100, an engine can quote either.
  • Google Business Profile and, for stores, Merchant Center kept current. Google lists both as things to maintain for AI features.
  • Structured data that matches the visible text. Google's guidance is explicit: make sure "structured data matches the visible text on the page". Marking up an FAQ the page does not show is a quick way to lose the snippet.

Step 4: what about llms.txt and "AI text files"?

The llms.txt proposal (Jeremy Howard, first published September 2024, revised August 2026) describes a Markdown file at /llms.txt with an H1, a summary blockquote and curated links so that an agent can read the important parts of a site without parsing HTML. Several AI vendors publish one for their developer documentation.

Two cautions. The proposal's own page makes no claim that any search engine or assistant reads these files when ranking or citing, and Google says you "don't need to create new machine readable files, AI text files, or markup" to appear in AI Overviews or AI Mode. Treat llms.txt as cheap hygiene for the tools that do read it, keep it in sync with the real pages, and spend the effort on the steps above.

Step 5: measure it, monthly

Most businesses cannot say whether AI engines send them anything. Four low-cost instruments:

  1. GA4 channel. GA4's default channel grouping now includes an "AI Assistant" channel; check Reports > Acquisition > Traffic acquisition. Add a custom channel group that matches referrers such as chatgpt.com, perplexity.ai, copilot.microsoft.com, claude.ai and gemini.google.com so you can see landing pages, not just totals.
  2. Search Console. Google reports AI Overviews and AI Mode appearances inside the normal Performance report under the "Web" search type; there is no separate AI report. Track impressions and clicks on the pages you rewrote.
  3. A citation panel. Write 20 prompts your buyers would ask (ten informational, ten "which company should I use for X in Y"), run them on the same day each month in ChatGPT with search, Perplexity, Gemini and Google AI Mode, and record which URLs are cited and where. An hour's work, and the only direct measure of visibility you will get.
  4. Crawler logs. The grep from Step 1, repeated monthly, tells you whether new pages are being fetched at all.

What to expect

Set expectations with the data that exists. Semrush's July 2026 study of 600,000+ US keywords found AI Overviews "grew an average of 71% across search engine results pages (SERPs) with commercial intent" over six months, while "the appearance of AI Overviews for transactional-intent queries decreased by 5%", and that Google Ads and AI Overviews "now appear together on the same SERP roughly 2x more often than they did a year ago". Research-stage queries ("how do I choose a Shopify agency") are increasingly answered by AI; "buy now" queries less so. Informational and comparison pages are where citations are won.

A realistic 90-day plan: fix crawler access in week one, rewrite the ten pages that matter with the checklist, add authors and organisation schema, publish an FAQ page with real answers, and run the citation panel from month one for a baseline. Citations follow recrawl, which Google warns can take weeks.

FAQ

Do I need special markup to appear in Google AI Overviews?

No. Google's documentation states there are "no additional requirements to appear in AI Overviews or AI Mode" and "no special schema.org structured data that you need to add". The page must be indexed, crawlable by Googlebot and eligible for a snippet, and existing structured data must match the visible content.

Will blocking GPTBot stop my site appearing in ChatGPT search?

No. OpenAI documents GPTBot (training) and OAI-SearchBot (search) as independent settings. A site can disallow GPTBot and still be shown in ChatGPT search answers as long as OAI-SearchBot is allowed. Blocking OAI-SearchBot removes the site from search answers, though it "can still appear as navigational links".

Does keyword density matter for AI answers?

The GEO study found keyword stuffing reduced visibility below the unoptimised baseline, while adding citations, quotations and statistics improved it by 30-40% on the study's main metric. Clear, sourced, specific sentences are what engines extract.

How do I know if an AI engine is sending me traffic?

Check the "AI Assistant" channel in GA4's traffic acquisition report and build a custom channel group for AI referrer domains. Search Console includes AI Overviews and AI Mode impressions under the "Web" search type. A monthly panel of 20 prompts run across four engines gives you citation data directly.

Is llms.txt worth adding?

It is cheap and harmless if kept current, but neither Google nor the proposal itself claims it affects citations or rankings. Fix crawler access, content and entity signals first.

Impinge Solutions builds and modernises websites and online stores and offers an AI search visibility audit. If you want the crawler check and citation panel run on your site, request an estimate.

Sources

Are you ready to partner with us?

Our good-natured team is waiting to hear your business requirements.

Book a free project call

Stay informed with our newsletter