The AI-Era SEO Playbook: AEO, GEO, and How to Get Cited by ChatGPT

Search changed. The question is no longer "how do I rank on Google" — it's "how do I show up in Google and in the answer ChatGPT, Perplexity, Claude, and Gemini serve to my future customer."

I run the SEO for several AI SaaS products. This is the playbook that works in 2026 — the part of the strategy that survives every model update and every algorithm change.

The three pillars#

1. Classic SEO is still the foundation#

None of the AI-era advice matters if Google can't crawl, render, and index your pages correctly. The basics still win:

• Technical SEO. Fast Core Web Vitals, clean crawl, proper canonical URLs, no orphan pages, structured data on every important page, working sitemap.xml, working robots.txt.
• Content quality. Original, expert-led content. Pages that answer questions humans actually ask. Internal linking between related pages so PageRank and topic authority flow where they should.
• Backlinks. High-authority, topically relevant links still move the needle. This is the slowest part of the strategy and the most important to do first.

If you're missing any of these three, fix them before you do anything else.

2. AEO — Answer Engine Optimization#

AEO is the practice of structuring content so AI assistants can use it as a trustworthy source. The mechanics:

• Schema.org JSON-LD. Real Person, Organization, Product, Service, FAQPage, Article, BlogPosting, BreadcrumbList. Every important page. Every time.
• Author and entity data. Linked author profiles, consistent NAP, organization schema linking back to the person.
• Quotable facts. Write sentences that stand on their own. AI systems pull snippets; make sure each section has at least one sentence that answers a question completely.
• FAQPage structured data. Add FAQPage JSON-LD wherever you answer real questions. Most AI systems use this directly.

3. GEO — Generative Engine Optimization#

GEO is newer and more specific: getting cited by the AI systems that generate answers. Tactics that work:

• Cite your sources. If your post references research or a public dataset, link to it. AI systems weight verifiable sources more heavily.
• Original data. Run a survey, publish a benchmark, share real numbers. Original research gets cited disproportionately.
• Clear author bios. AI systems increasingly cite byline. Make sure your author bios link to verifiable credentials.
• llms.txt. A machine-readable summary of your site placed at /llms.txt and /llms-full.txt. AI crawlers check this first.

The files you must publish#

Three files in 2026 are not optional:

/llms.txt#

A short, plain-text description of your site — who you are, what you do, where you are, what's on the site, and how to contact you. The llms.txt spec is a proposed standard that GPTBot, ClaudeBot, and other AI crawlers already respect. Keep it under a few hundred tokens.

/llms-full.txt#

The expanded version with full content excerpts. This is what AI systems actually read when they need to reason about your site.

/sitemap.xml#

A complete, accurate sitemap. Include every indexable URL. Don't include admin pages, API routes, or staging. Submit it to Google Search Console and Bing Webmaster Tools.

How AI crawlers actually work#

The four major AI crawlers to care about right now:

• GPTBot (OpenAI). Crawls to train models and to power ChatGPT browse. Block it from /admin and /api, allow it everywhere else.
• ClaudeBot (Anthropic). Crawls for Claude.ai. Same access policy.
• PerplexityBot. Powers Perplexity's index. Allow it on public content.
• Google-Extended. Controls whether Gemini uses your content for grounding. Allow it unless you have a strong reason not to.

In robots.txt this looks like:

User-agent: GPTBot
Allow: /
 
User-agent: ClaudeBot
Allow: /
 
User-agent: PerplexityBot
Allow: /
 
User-agent: Google-Extended
Allow: /
 
User-agent: AhrefsBot
Disallow: /

Block the SEO-spam crawlers (Ahrefs, Semrush, MJ12, DotBot) — they don't contribute to your rankings and they cost bandwidth.

Citations are the new ranking#

The single biggest shift in 2025–2026 is that AI search cites sources. Being cited by ChatGPT or Perplexity on a relevant query is worth more than ranking #3 in Google.

How to maximize citations:

1. Be the canonical answer. For your core topics, write the most accurate, most complete, most citable page on the internet. Original data helps. Clear author attribution helps. A trusted domain helps.
2. Match the question shape. AI systems prefer pages that mirror the question's structure. FAQ pages get cited for "what is X" questions. Comparison pages get cited for "X vs Y" questions.
3. Update aggressively. AI systems weight recency. Update your top pages quarterly and re-submit your sitemap.

What to measure#

Rankings still matter, but track these too:

• Branded search volume. Are more people Googling your brand?
• Citation count. When AI answers mention your brand or domain, that's a leading indicator of growth.
• Direct + referral traffic from AI sources. Perplexity and ChatGPT increasingly link back to sources.
• Real conversions. None of this matters if it doesn't move pipeline.

The 90-day plan#

If you're starting from scratch:

• Weeks 1–2. Fix technical SEO. Audit crawl, render, structured data, sitemap, robots. Fix everything blocking indexing.
• Weeks 3–6. Publish llms.txt, llms-full.txt, and ai.json. Add schema.org JSON-LD to every important page. Write or update author bios.
• Weeks 7–10. Publish 4–6 deep, original, well-cited pieces of content on your core topics. Include original data where possible.
• Weeks 11–13. Outreach. Get backlinks from real, topically-relevant sites. Pitch original research to journalists.

You'll start seeing AI citations within a quarter and Google ranking gains within two.

If your site isn't showing up in AI answers yet, book a strategy call — I help AI SaaS teams ship the full AEO/GEO stack end to end.