Search has changed. Many people now get answers from ChatGPT, Perplexity, Gemini and Google’s AI Overviews before they click any blue link. If your page is not the one those systems quote or cite, you lose that visit, even if you still rank. This guide explains what AEO and GEO mean, how AI search engines choose sources, and what you can do this week to get your site cited more often.
What Is AEO and GEO?
AEO (Answer Engine Optimization) means structuring content so a search or answer engine can pull a direct, correct answer from it. It started with featured snippets, People Also Ask boxes and voice assistants. The goal is to be the answer, not just a result.
GEO (Generative Engine Optimization) means making your content easy for large language model (LLM) systems to find, understand, trust and cite when they write an answer. ChatGPT search, Perplexity, Gemini, Copilot and Google AI Overviews and AI Mode all fall in this group.
AEO vs GEO vs classic SEO
- Classic SEO aims to rank a page in the list of results and win the click. Signals: relevance, links, technical health, page experience.
- AEO aims to win the extracted answer: a snippet, a PAA answer, a spoken reply. Signals: a clear question, a short direct answer, clean structure.
- GEO aims to be retrieved and cited inside a generated answer. Signals: crawl access for AI systems, clear facts and entities, trust and authorship, and content that adds something the model can’t get elsewhere.
These are not three separate jobs. GEO sits on top of SEO. Most AI search tools retrieve pages from a web index (their own, Bing’s or Google’s) before they write an answer. If your page is not crawlable, indexed and relevant, no AI tool will cite it. So treat AEO and GEO as an extra layer on good SEO, not a replacement.
How AI Search Engines Pick Their Sources
Each product works a little differently, and none publish their full ranking logic. But the general pattern is well understood:
- Query understanding. The system rewrites your question, often into several sub-queries (Google calls this “query fan-out” in AI Mode).
- Retrieval. It fetches candidate pages from a search index. ChatGPT search relies heavily on Bing plus its own crawler, Perplexity runs its own index and crawler, and Gemini and AI Overviews use Google’s index.
- Passage selection. It picks the specific passages that answer each sub-query best. This is passage-level, not page-level. One clear paragraph can earn a citation even if the rest of the page is average.
- Synthesis and citation. The model writes an answer and links the sources it relied on, usually favouring ones that look trustworthy, specific and current.
What this means in practice: you need to be reachable (crawlers allowed), retrievable (indexed and relevant), extractable (clean passages that answer one thing well) and trustworthy (a real author, accurate facts, a known entity).

Practical Steps to Get Cited by AI Search
1. Write clear answers first, details second
Put a direct answer of 40–60 words right under each question-style heading. Then expand with steps, examples and caveats. Models lift short, self-contained passages. A paragraph that starts with “It depends” and gets to the point three sentences later is hard to quote.
- Use question headings that match how people ask (“How long does X take?”).
- Define terms in one sentence: “X is a …”.
- Use numbered steps for processes and tables or lists for comparisons.
- Keep one idea per paragraph so each passage stands on its own.
2. Add FAQ sections and structured data
Schema markup does not guarantee a citation, but it removes ambiguity about what your page is and who made it. Use Article (with author and dates), Organization or Person, Product, HowTo-style steps in plain HTML, and FAQPage where you genuinely answer common questions. Google limited FAQ rich results to a small set of sites in 2023, so don’t expect the visual result, but the clear question-and-answer structure still helps machines read your page.
You can build valid JSON-LD in seconds with our free FAQ Schema Generator. Paste your questions and answers, copy the code, and test it in Google’s Rich Results Test.
3. Publish an llms.txt file
llms.txt is a proposed standard: a Markdown file at your site root (/llms.txt) that gives AI systems a short summary of your site and links to your most important pages. It is not an official ranking signal, and adoption by the big AI companies is uneven. It takes ten minutes, though, and gives any tool that reads it a clean map of your best content.
Create one with our llms.txt Generator. Add your site name, a one-line description and your key pages grouped by section, then upload the file to your root folder.
4. Let the right AI crawlers in
This is the most common silent problem I find in audits. A security plugin, CDN bot setting or old robots.txt rule blocks AI crawlers, and the site owner has no idea. The main user agents to know:
- OAI-SearchBot – OpenAI’s crawler for ChatGPT search results. Block it and you can’t appear as a ChatGPT search source.
- ChatGPT-User – fetches pages when a user asks ChatGPT to look at a link.
- GPTBot – OpenAI’s crawler for model training. You can block this one and still allow search.
- PerplexityBot and Perplexity-User – Perplexity’s index and live fetches.
- Google-Extended – a robots.txt token that controls use of your content for Gemini training and grounding. It does not affect Google Search or AI Overviews, which use Googlebot.
- ClaudeBot, Claude-SearchBot, Applebot-Extended, Bingbot – also worth checking.
Decide on a policy: many publishers allow search crawlers and block training crawlers. Then check what your robots.txt really says with our AI Crawler and Robots.txt Tester. Paste your robots.txt and a path, and it shows which AI bots are allowed or blocked. Also check your CDN or firewall (Cloudflare has an AI-bot block toggle), because robots.txt can say “allow” while the firewall says no.
5. Show real E-E-A-T signals
Experience, Expertise, Authoritativeness and Trust are Google’s quality framework, and AI systems reward the same things because they reduce the risk of citing something wrong.
- Put a named author on every article, with a short bio and a link to an author or About page.
- Show first-hand experience: your own screenshots, tests, numbers from your own accounts, and the steps you actually took.
- Cite primary sources (official documentation, standards, original studies) and link to them.
- Show published and updated dates, and actually update pages when facts change.
- Make it easy to reach you: a real contact page, business details and policies.
6. Make your entity clear
AI models think in entities: people, brands, products, places and how they relate. If the model can’t tell who you are, it is less likely to trust or name you.
- Use the same name, description and logo everywhere (site, social profiles, directories).
- Add Organization or Person schema with
sameAslinks to your real profiles. - Describe what you do in one plain sentence on your homepage and About page.
- Earn mentions on other reputable sites. Brand mentions, not just links, help models connect you with a topic. Guest posts on relevant sites are one way; our Write for Us page explains how we accept them here.
7. Add information that doesn’t exist elsewhere
If your article repeats the same ten points as every other page, the model has no reason to cite you in particular. Add original data, a worked example, a template, a tool, a clear opinion backed by experience, or a comparison nobody else made. Unique, specific content is what gets quoted.
8. Keep the technical basics tight
- Serve main content in the HTML. Many AI crawlers don’t run JavaScript well, so content that only appears after scripts run may be invisible to them.
- Keep pages fast and avoid interstitials that hide content.
- Write unique title tags and meta descriptions. Preview them with the SERP Snippet Preview tool.
- Keep your XML sitemap current and submit it to Google Search Console and Bing Webmaster Tools. Bing matters more now because ChatGPT search draws on it.
How to Measure AEO and GEO
Tracking is still immature, so use several rough signals together:
- Referral traffic in GA4 from chatgpt.com, perplexity.ai, gemini.google.com and copilot.microsoft.com.
- Search Console: AI Overviews and AI Mode clicks and impressions are counted inside the normal Web performance data, so watch impressions on question queries.
- Manual prompts: keep a list of 20–30 questions your customers ask and check monthly which sources each AI tool cites.
- Server logs: confirm OAI-SearchBot, PerplexityBot and others are actually visiting your key pages.
- Branded search: growing brand searches often follow more AI mentions.
Common AEO and GEO Mistakes
- Blocking AI search bots by accident through a plugin, CDN or a copy-pasted robots.txt.
- Blocking every AI bot on purpose and then wondering why you’re never cited. Blocking training is a choice; blocking search crawlers removes you from those answers.
- Burying the answer under long intros and stories.
- Treating llms.txt or schema as magic. They help machines read you. They don’t replace useful content or authority.
- Mass-producing thin AI content. Models and Google both discount pages that add nothing new.
- Anonymous content with no author, dates or sources.
- Ignoring Bing. Not being indexed in Bing can hurt ChatGPT visibility.
- Content hidden behind JavaScript, tabs that load late, or logins.
AEO and GEO Checklist
- Robots.txt and CDN allow the AI search crawlers you want (test with the AI Crawler and Robots.txt Tester).
- Site is verified and sitemaps submitted in Google Search Console and Bing Webmaster Tools.
- An llms.txt file is live at your root (build it with the llms.txt Generator).
- Every key page answers its main question in the first 40–60 words under a clear heading.
- Question-style H2/H3 headings with short, self-contained answers.
- FAQ section and valid JSON-LD on relevant pages (use the FAQ Schema Generator).
- Article, Organization or Person schema with author, dates and sameAs links.
- Named author, bio, About and Contact pages.
- At least one original element per article: data, example, template or tool.
- Main content visible in raw HTML without JavaScript.
- Pages updated when facts change, with a visible updated date.
- Monthly check of AI referrals and manual prompt tests.
Start with the crawl check and the answer-first rewrite of your top five pages. Those two steps usually make the biggest difference. If you want help auditing your site for AI search, get in touch.
Frequently Asked Questions
What is the difference between AEO and GEO?
AEO (Answer Engine Optimization) focuses on getting your content chosen as the direct answer in features like featured snippets, People Also Ask and voice results. GEO (Generative Engine Optimization) focuses on getting your content retrieved and cited inside AI-generated answers from tools like ChatGPT, Perplexity, Gemini and Google AI Overviews. They overlap a lot, and both depend on solid SEO.
Does AEO and GEO replace SEO?
No. Most AI search tools retrieve pages from a search index before writing an answer, so your pages still need to be crawlable, indexed and relevant. AEO and GEO are an extra layer on top of good SEO, not a replacement for it.
Does llms.txt help with ranking in AI search?
llms.txt is a proposed standard, not a confirmed ranking factor, and major AI companies have not committed to using it. It is quick to create and gives any system that reads it a clean summary of your important pages, so it is a low-cost step, but it won’t make up for weak content.
Should I block GPTBot and other AI crawlers?
That depends on your goals. You can block training crawlers such as GPTBot while still allowing search crawlers such as OAI-SearchBot, so you can still be cited in ChatGPT search. If you block the search crawlers, you will not appear as a source in those tools.
Does schema markup help get cited by AI?
Schema markup helps machines understand what your page is about, who wrote it and how it is structured. It does not guarantee a citation, but combined with clear answers and trust signals it makes your content easier to retrieve and quote.
How do I know if AI search engines cite my site?
Check GA4 for referral traffic from chatgpt.com, perplexity.ai and gemini.google.com, watch impressions on question queries in Search Console, review server logs for AI crawler visits, and regularly ask AI tools the questions your customers ask to see which sources they cite.
