AI Search Is Changing SEO: A Practical AEO/GEO Playbook
A hands-on playbook for making your content findable and citable by AI answer engines — ChatGPT, Perplexity, Google AI Overviews — beyond the basics of what SEO is and why it matters.
Traditional SEO gets you ranked. AI answer engines need something extra: they read a page once, extract an answer, and either cite you or quietly rewrite your content in their own words without a click-through. This playbook is the tactical follow-up to the fundamentals — concrete changes you can make this quarter to be the source AI systems actually pull from.
Why classic SEO is not enough anymore
Search engines rank pages; AI answer engines synthesize an answer from several pages and show it directly, often with no click required. Ranking #1 no longer guarantees visibility if an AI Overview or a ChatGPT answer satisfies the user first. The target has shifted from "get clicked" to "get cited, and get clicked when the answer needs more depth than a summary can give."
What AI crawlers actually need from a page
- A direct, extractable answer near the top — not buried after three paragraphs of scene-setting.
- Clean semantic HTML and server-rendered content — most AI crawlers do not execute JavaScript reliably.
- Explicit structured data (Schema.org JSON-LD) that removes ambiguity about what the page is and who wrote it.
- A clear, named, credentialed author — E-E-A-T signals matter more, not less, when a model has to decide whether to trust a claim.
- Content that is genuinely citable — specific numbers, named examples, and claims a model can quote without hedging.
The AEO/GEO playbook
- Lead with the answer. Put a 2–3 sentence direct answer immediately under the H1, then expand — this is the block most likely to get lifted into an AI summary.
- Structure for extraction. Use real headings, short paragraphs, and lists/tables for anything comparative — models parse structure far more reliably than prose.
- Ship Schema.org markup on every content page: `Article`, `FAQPage`, `BreadcrumbList`, and `Organization`/`Person` for authorship.
- Allow the AI crawlers. Check `robots.txt` for `GPTBot`, `PerplexityBot`, `ClaudeBot` and similar — blocking them by default (a common CMS setting) makes you invisible to citation.
- Publish `llms.txt` — an emerging convention that gives AI systems a clean, curated map of your most important pages.
- Make claims specific and citable. "Fast" is not citable; "LCP under 2.5 seconds" is.
- Refresh dated content. Models weight and prefer recently updated pages when a query is time-sensitive — keep `dateModified` accurate.
How to know if it is working
Traditional analytics under-report AI visibility because a citation often does not produce a click. Watch referral traffic from `chat.openai.com`, `perplexity.ai` and similar sources in your analytics, and periodically ask the AI tools your target queries directly to see whether — and how — you are cited.
How this fits our SEO foundation
This playbook builds on the fundamentals in our guide to what SEO means and why you need it — read that first if you need the foundation before the tactics.
The bottom line
AEO and GEO are not a replacement for SEO — they are what SEO looks like once part of your audience reads your content secondhand, through a model. Structure for extraction, back every claim with structured data and a real author, and you win both the classic ranking and the AI citation.
Frequently asked questions
What is llms.txt and do we actually need one?
llms.txt is an emerging, community-proposed convention — a plain-text file at your site root that gives AI systems a curated map of your most important pages, similar in spirit to a sitemap but written for models rather than crawlers. No major AI vendor has confirmed it as a required signal yet, but it costs little to publish and gives you a direct way to point models at your best content.
Should I block AI crawlers like GPTBot to protect my content?
That trade-off depends on your goals. Blocking GPTBot, PerplexityBot or ClaudeBot in robots.txt keeps your content out of model training and citation entirely — which protects it, but also makes you invisible when someone asks an AI tool a question your page could answer. For most businesses trying to be found, allowing the crawlers is the right default.
Can I measure how often AI tools cite my site?
Not precisely — most AI products do not expose citation logs the way search consoles expose ranking data. The closest proxies are referral traffic from domains like chat.openai.com or perplexity.ai in your analytics, and manually asking the tools your target queries to see whether your page shows up as a source.