Answer Engine Optimization: What AEO Means in Practice

Answer engine optimization is the practice of influencing what AI answer engines say about your brand when someone asks them a question. The engines in scope as of October 2026 are ChatGPT, Perplexity, Claude, Gemini, Microsoft Copilot and Google's AI Overviews and AI Mode. The outcome you want has three parts: the engine names you, it gets your facts right, and where it shows sources, it links to you.

GEO (generative engine optimization) is the same job under a different name. Both labels target the same engines, which fetch pages the same way and build answers the same way. Vendors sometimes split them, with AEO framed around question-and-answer formats and GEO around generative answers, but no engine documents treating the two differently. Our definition of answer engine optimization and the companion guide to generative engine optimization cover the terms in more depth.

What this means for you: ignore the label and ask what a given tactic actually changes. Most AEO advice on the market changes nothing the evidence can detect. A smaller set of checks changes whether an engine can see you at all.

An AI answer passes through four stages. Knowing which stage a tactic touches tells you whether it can work.

The parametric stage is fixed. OpenAI's model documentation lists a February 16, 2026 knowledge cutoff for the GPT-5.6 family and August 31, 2025 for the Instant path that ChatGPT also routes to. Content published after a cutoff cannot appear in an answer given without search.

Retrieval is discretionary. ChatGPT decides per query whether to search, and OpenAI says it rewrites the user's prompt into one or more targeted queries. Google may fan a question out into several related searches. Nobody outside the engine sees those queries.

The research lands hard on the last stage. A July 2026 critical survey of 45 GEO studies (arXiv 2607.14035) concluded that already-retrieved content can change its citation or use, "but no reviewed technique shows a stable, longitudinal, cross-platform causal effect on organic discoverability." Nearly every published AEO or GEO effect was measured on pages already inside the model's context. The real-world question, whether a page gets retrieved at all, sits upstream of those studies.

Three tactics dominate AEO checklists. Primary sources contradict all three as ways to get cited.

FAQ schema

Google states that structured data isn't required for its generative AI search features. Ahrefs ran a matched-control test on 1,885 pages that added schema and found AI Overview citations fell 4.6%, with no meaningful lift on any platform. Google also deprecated FAQ rich results in May 2026. One honest caveat: every page in the Ahrefs study was already heavily cited, so whether schema helps a page enter the candidate set is untested, not refuted. That still leaves no evidence for selling it as a citation lever.

Chunking content for LLMs

Google's guide to its generative AI features rejects chunking as a requirement and says there is no ideal page length. The academic chunking research measures how a retrieval system splits documents inside its own index. You don't set how ChatGPT segments your page, so rewriting your page into short blocks targets a variable you don't control.

Answer-first rewrites and AI-friendly style

Google says you don't need to write in a specific way for generative AI search, and you don't need to chase every long-tail variant of a question. C-SEO Bench, a NeurIPS 2025 benchmark, found most conversational SEO rewriting methods largely ineffective and often harmful to a document's ranking in the model's context. The July 2026 survey warns that citation-oriented rewrites can impair retrieval.

Two more you will hear

• llms.txt. Google says Search ignores it, and no other engine documents reading it.
• "Up to 40% more visibility." That figure traces to the 2023 GEO paper, which measured an engine its authors built, with sources already supplied in context.

What survives on the page is narrow. A specific fact stated plainly (a price, a count, a one-line definition) gives an engine a sentence it can quote. That is a reason to write clearly, not a formatting trick.

The levers that hold up are publisher-controlled and mostly subtractive. They remove blocks rather than add signals.

In your control

• Which AI crawlers your robots.txt allows, bot by bot.
• Whether your firewall or CDN actually lets those crawlers through.
• Whether your important content is in the raw HTML before JavaScript runs.
• Whether noindex, nosnippet, max-snippet:0 or data-nosnippet keep content out of answers.
• Whether your own site states your facts consistently.
• Whether you're mentioned on the third-party sites engines already cite. The evidence shows a strong association here, not a guaranteed effect.

Not in your control

• Whether an engine searches for a given question.
• Which queries it actually runs, and where your page lands in its candidate list.
• What the model learned in training, and when the next model ships.
• How big your brand already is. Large brands dominate citations in every study, and no tactic offsets that.

The next move is to spend effort only on the first list. If you're weighing whether the work justifies a budget, the page on whether AEO is worth it for a brand walks through that decision.

Work from the gates outward. Each step can silently cancel everything after it.

1. Check robots.txt per bot. Rules are independent. OpenAI's crawler page says "Each setting is independent of the others": GPTBot governs training, while OAI-SearchBot controls ChatGPT search. A site can opt out of training and stay in search:

User-agent: OAI-SearchBot
Allow: /

User-agent: GPTBot
Disallow: /

Do the same review for PerplexityBot, Claude-SearchBot and Meta-WebIndexer. There is no robots.txt token for Google AI Overviews; Googlebot rules and Search Console are the controls there.

2. Check the network layer. OpenAI's help centre says eligibility means allowing OAI-SearchBot and confirming your host or CDN allows OpenAI's published IP ranges. A firewall that returns 403 to a crawler your robots.txt allows removes you without warning. From September 15, 2026, Cloudflare blocks its Training and Agent crawler categories by default on ad-displaying pages for new domains and existing free-tier customers, so you can lose access without editing anything. The AI Crawler Checker requests a page as each AI crawler and shows which ones get turned away.

3. Check rendering. Vercel's crawler study found ChatGPT and Claude crawlers fetch JavaScript files (11.50% and 23.84% of requests) but don't execute them. Compare your raw HTML with the rendered page. Pricing, product names and key claims that appear only after hydration are invisible to those bots. Googlebot does render. See code-level issues that stop AI indexing for the full list.

4. Check index and snippet eligibility. Google says a page must be indexed and eligible to show with a snippet to appear as a supporting link in AI Overviews or AI Mode, with no additional technical requirements. Look for stray noindex or snippet directives inherited from templates. Check Google Search Console and Bing Webmaster Tools separately, since Copilot draws on Bing's index.

5. State your facts plainly and consistently. Put your prices, plan limits and one-line definition in fetchable HTML, and make the pricing page, about page and footer agree.

6. Measure each engine separately. Engines behave very differently, so one blended number hides the detail you need. The free checker on this site shows whether engines name you for three buyer questions, as a snapshot.

These errors show up repeatedly, and most of them cost more than doing nothing.

• A blanket AI block. Disallowing every AI user agent to stop training also removes ChatGPT search eligibility. Split training bots from search bots.
• Allowing GPTBot to get found. GPTBot governs training only. Allowing it does nothing for ChatGPT search.
• Ripping out existing schema. Overcorrection is also unsupported. Google says existing markup does no harm. It just isn't an AI lever.
• Tracking an "AI rank". Google's generative AI report in Search Console exposes impressions only, with no position data. Any tool reporting your position in Google AI has nothing to validate it against. The same problem undermines most brand AI visibility leaderboards.
• Mass-publishing thin pages. In OtterlyAI's 2026 experiment, Google deindexed two new sites carrying roughly 1,000 AI-generated posts each, with no manual action shown, and Google's AI surfaces then almost never cited them.
• Hidden text or prompt injection. Research on LLM search agents shows planted instructions can sway answers once a page is retrieved. It is illegitimate, it breaks platform policies, and it puts your brand at risk. Never use it, and watch third-party pages for it.

When you evaluate vendors, ask which of their recommendations sit in the access layer and which sit in the rejected list. Our overview of AEO tools gives a framework for that conversation.

Is AEO different from SEO? Less than the label suggests. Google says there are no additional requirements or special optimizations to appear in AI Overviews or AI Mode beyond being indexed and eligible for a snippet. Ahrefs found 38% of AI Overview citations came from pages ranking in the top 10, so organic rank still matters but is far from the whole story. Other engines use their own indexes and crawlers, which is why per-bot access checks matter beyond classic SEO.

Should I remove FAQ schema I already have? No. Google says existing structured data does no harm, and removing it costs time for no documented gain. The point is narrower: schema is not a lever for AI answers, so don't add it expecting citations. Keep question-and-answer content where it helps readers, because the deprecation affected FAQ markup as a Google rich result, not the value of answering real questions on the page.

Does blocking GPTBot remove my site from ChatGPT search? No. OpenAI documents GPTBot as its training crawler and OAI-SearchBot as the crawler that surfaces sites in ChatGPT search, and states that each setting is independent. You can disallow GPTBot to keep content out of training while allowing OAI-SearchBot. Sites opted out of OAI-SearchBot won't be shown in ChatGPT search answers, though OpenAI says they can still appear as navigational links.

How long does AEO take to show results? Nobody can promise a timeline or an outcome. Access fixes take effect once a crawler next fetches your pages, but being fetchable only makes you eligible. Answers given without web search draw on training data frozen at a cutoff, so changes there wait for a new model generation. The honest measure of progress is which blocks you've removed, then what each engine says when you check again.

Can I track my position in AI answers? Not in any validated way. Google's Search Console report for AI features shows impressions only, with no clicks or position data, and Google says third-party tools don't have access to its internal ranking data. What you can track is whether an engine names you for a set of questions, whether it cites you, and whether the facts it states are correct, measured per engine.