How I Do SEO / GEO / AEO for AI Products (2026)

GEO (Generative Engine Optimization) and AEO (Answer Engine Optimization) are still SEO with extra layers — answer-first content, dense structured data, and being in each AI engine's index. Here is the framework I use across Shahriar Labs products in 2026.

The two audiences

Classic search crawlers (Google, Bing) and AI answer engines (ChatGPT, Claude, Gemini, Perplexity, Copilot) both rank you. They overlap but differ. SEO ranks you in a list of links. GEO/AEO gets you cited inside the AI's generated answer.

The priority order (highest ROI first)

Priority	Action	Why
1	Indexing reach	If you are not in Google + Bing + Brave, you are invisible to half of AI engines
2	Structured data	FAQPage on every page = ~3.2× AI Overview citation
3	Answer-first content	Direct answer in first 40-60 words
4	Titles + meta	Consistent, keyword-first, ≤60 chars
5	Content moat	Guides, tutorials, error-fix hubs
6	Off-site	Listicles, Wikidata, social presence

Where each AI engine grounds

• ChatGPT → Bing index
• Claude (incl. Claude Code) → Brave index
• Gemini → Google index
• Perplexity → Mix of Bing + Google + own index

Wrong index = invisible to that AI. Register everywhere.

Structured data that works

Schema	Pages
Person + ProfilePage	Entity / about page
Organization	Brand page
Article + BreadcrumbList	Every blog post
FAQPage	Every page that can carry it
SoftwareApplication	Every product page
HowTo	Tutorials with steps
ItemList	Content hubs

Use @id to link them. A unified @graph on the home page consolidates the brand.

I built a full open-source skill that wraps this playbook: seo-master-skill.

Answer-first writing (the AEO move)

Lead with the answer in the first 40-60 words. Definition-first: "X is Y that does Z."

Don't:

"In today's fast-moving world of AI engineering, many founders ask themselves..."

Do:

"LetX is a real-time collaborative LaTeX editor with offline-first editing and browser-side compile."

Hard rules learned the hard way

• Never fabricate aggregateRating — Google manual action risk.
• Never copy a competitor's stat. Every claim must be yours and true.
• Titles ≤ 60 chars or Google truncates.
• FAQ schema requires the FAQ to be visible on the page.
• Code-split heavy content (Three.js, etc) so initial JS stays small.
• IndexNow does not notify Google — only Bing/Yandex/DuckDuckGo.

The 2026 changes

• AI Mode in Google search is now mainstream; queries shift from keywords to questions.
• May 2026 core update demotes shallow answer-bait. Depth wins.
• llms.txt is optional, low priority — Google says it is not needed for AI search.

The compounding moves

• Internal links — every entity mention links to its canonical page.
• sameAs schema — Person + Organization sameAs lists every social and product domain.
• Off-site presence — Wikidata entry, dev.to crossposts with rel=canonical.

The framework, condensed

1. Be in every index (Google, Bing, Brave, IndexNow ping).
2. Ship JSON-LD on every page.
3. Lead with the answer.
4. Tight titles, no fake numbers.
5. Real depth, not Q&A spam.
6. Compound by linking.

If you want this as a skill your AI coding agent can apply automatically, use seo-master-skill. Internal links: My AI Agent Skills Stack.

FAQ

Q: Is GEO different from SEO? A: It is SEO with extra surfaces. The foundation is the same; the additions are answer-first writing and dense structured data.

Q: What's the single highest-ROI move? A: Get indexed in Bing. ChatGPT grounds on Bing; missing from Bing = invisible to ChatGPT.

Q: Does llms.txt matter? A: Low priority. Add it if cheap; never before content, schema, indexing.

---

Written by Shihab Shahriar Antor. I build seo-master-skill and ship products at Shahriar Labs. Hire me.