Skip to content
New this week26 prompts for Sora, Veo and Runway video workflows.
Explore video prompts→

Search AIPrompt123

Search prompts, categories, platforms and repositories

SEO, AEO & GEO

How to get cited by ChatGPT, Perplexity and AI Overviews

Site Admin·· 4 min read·Updated 14 November 2026

The short answer

To get cited by AI search engines, publish clear pages with a 40-60 word direct answer, question-style headings, structured data, original data and visible author expertise. Make claims easy to quote and verify, keep entity information consistent, and allow AI crawlers in robots.txt.

Google ranks a page and sends a click. An answer engine reads several pages, synthesises a response and may cite a few. The unit of competition is the passage, not the page. A long article with one excellent paragraph can be quoted while the rest is ignored. A short page with a clean definition can beat a comprehensive guide if the definition is easier to extract.

That changes the optimisation target: write self-contained passages that answer a question completely, with enough context to be understood out of order.

The structure that gets quoted

  1. A 40-60 word direct answer at the top. Not a teaser. The actual answer.
  2. Question-style H2s. Mirror how people ask.
  3. Definitions before detail. "What is X?" then "How does X work?"
  4. Lists and tables for steps and comparisons. These are easy to quote.
  5. A unique data point or example the model cannot get elsewhere.
  6. An FAQ block with concise answers.
  7. Visible author, date and sources.

This is exactly the structure we use on prompt pages and category hubs.

Entity consistency

Answer engines build an entity graph. Your brand, author names, product names and URLs should be consistent across your site, social profiles, GitHub and directories. Use the same name and description. Link to your profiles with sameAs in Organization schema. Inconsistent naming makes it harder for a system to attribute a claim to you.

For a company, publish an About page with founding details, location, leadership and a clear description of what you do. For an author, publish a bio with credentials and links.

Original evidence beats summary content

A model can summarise the top ten results without your help. What it cannot invent is your data: a survey, a benchmark, a teardown, a copy-count statistic, a methodology note. Original evidence gives an answer engine a reason to cite you rather than a competitor.

If you publish a statistic, explain the method, the sample and the date. Add a short methodology section and a last-updated line. This is also what E-E-A-T reviewers look for.

Technical checklist

  • Allow AI crawlers. GPTBot, Google-Extended, PerplexityBot, ClaudeBot and others should be allowed if you want citations. Keep a human-readable policy.
  • Server-render the content. If the answer only exists after client-side JavaScript, many crawlers will miss it.
  • Use semantic HTML. One H1, logical H2/H3, real lists and tables, descriptive link text.
  • Add JSON-LD. Organization, WebSite with SearchAction, BreadcrumbList, Article, FAQPage, HowTo and relevant domain types.
  • Publish llms.txt. A concise map of your most useful pages for language models.
  • Keep pages fast. Crawlers have budgets, and users have less patience.
  • Avoid noindex on pages you want cited. Check pagination and filter pages for accidental noindex.

AEO: answer engine optimisation

AEO overlaps with GEO but focuses on the answer itself. Write the answer a voice assistant could read aloud without context. Use the question as the heading, then answer in two or three sentences. Put the detail below for humans who want it.

For comparisons, use a table with a clear verdict first. For how-tos, use numbered steps with one action per step. For definitions, use the pattern "X is … It matters because …".

What not to do

  • Do not stuff keywords. It does not help extraction and makes passages harder to quote.
  • Do not publish thin programmatic pages. If the only difference is a city name, it will not be cited and may be ignored or penalised.
  • Do not fake expertise. Invented credentials and citations are increasingly easy to detect and damaging to trust.
  • Do not block AI crawlers and expect citations. Pick a policy and be consistent.

Measuring GEO

Track branded search volume, referral traffic from AI assistants where available, and manual citation checks for your priority questions. Ask the same question across ChatGPT, Perplexity, Gemini and Google AI Overviews monthly and log whether you are cited, which page, and what passage was used. Treat it as a qualitative research process, not a precise ranking.

You can read our own llms.txt and methodology note for an example of the transparency pattern.

The bottom line

Write for a smart reader who is in a hurry, then make the page easy for a machine to parse and quote. Original evidence, clear structure and consistent entities do most of the work. Everything else is maintenance.

Read the SEO and marketing prompts to put this into practice.

Copy-ready prompts featured in this guide

Copy any prompt below or open its full page with interactive variable fields and platform tabs.

Site Admin

Author & Platform Administrator

Administrator and lead editor at AIPrompt123, curating high-performance AI prompts, SEO/AEO/GEO strategies, and developer repositories.

Verified Knowledge · FAQ

Questions about this guide

Short answers to what readers ask most about this topic.

Keep reading

Roundups2 min read

New prompts and trending repos: week 46

This week: a system prompt for research agents, a safer browser automation pattern, and three repositories gaining stars fast.

Site Admin·