On 15 May 2026, Google's Search Central team did something it had refused to do for two years: it published an official guide on how to optimize for generative AI features in Search. Titled "Optimizing your website for generative AI features on Google Search," the document is short — and the parts that say "you do not need to do this" are louder than the parts that say "you should do this." For an industry that spent 24 months selling AEO and GEO as separate disciplines with separate retainers, it lands like a fire alarm.
The headline framing from Google: AEO (Answer Engine Optimization) and GEO (Generative Engine Optimization) are not new disciplines. They are still SEO. AI Overviews and AI Mode read the same web that Search reads. Optimizing for one is optimizing for the other. The methodology you already had — if you actually had one — still works.
1. The four myths Google explicitly killed
Read the guide and you will see four claims that have been on every AEO/GEO consulting slide-deck for the last year get explicitly rejected:
- llms.txt is not needed. Google does not look at
llms.txtfor AI Overviews. The file does not hurt you, but it does nothing for visibility. - Content chunking is not required. Google's systems read the whole page, understand multiple topics on it, and select the relevant segment. Splitting a long article into 30 micro-pages is not the answer.
- "AI-friendly rewriting" is not a thing. Google can read normal, well-written prose. You do not need to bullet-ize every paragraph or insert "the answer is" stems.
- No special schema for AI. Existing structured data (FAQ, HowTo, Article, Product) is the only schema set Google uses for AI Search. Anyone selling you a "GEO schema package" is selling air.
2. The three things the guide quietly confirms you must do
Now the part the consultants will not lead with — the three things Google effectively validated:
- Non-commodity, first-hand content. Google contrasts a generic "7 Tips for First-Time Homebuyers" against "Why We Waived the Inspection and Saved Money — A Look Inside the Sewer Line." The first is invisible to AI; the second gets cited. If your content is information any LLM already knows, you will not get surfaced.
- Multimodal coverage. Images and short video clips embedded in your pages are increasingly being surfaced as the primary citation in AI Overviews. A page with one good diagram now outperforms a page with three paragraphs.
- Technical hygiene. Indexable, crawlable, semantic HTML, fast load times, mobile-passable. None of this is glamorous, and all of it is non-negotiable. A page that AI cannot fetch cannot be cited, regardless of how well-written it is.
3. What this means for budgets this quarter
If you are a marketing team that signed an "AEO retainer" in 2024 or 2025, this is the week to renegotiate. Most of those retainers were billing for work — llms.txt installs, schema packages, content chunking — that Google has now publicly said does nothing. The work that does matter (genuine first-hand content, multimodal production, technical foundations) is the work an SEO retainer should already cover.
If you are an agency that sells those retainers, this is the week to rewrite your pitch. The fact that AEO and GEO are SEO is not a death sentence — it is a clarification. The agencies that sell deep, first-hand content production will win this cycle. The agencies that sold "AI plumbing" services will not.
4. The PHP / Nginx engineering checklist
For the developers reading: there is still work to do, just different work than the GEO consultants told you.
- Confirm your
robots.txtdoes not blockGooglebot,Google-Extended, or the AI Mode user-agent. - Server-render anything you want cited. AI crawlers do not execute JavaScript reliably. A React SPA without SSR is functionally invisible.
- Drop Time To First Byte under 200ms — Nginx with a hot cache and HTTP/2 Push will get most teams there.
- Ship structured data (FAQ, HowTo, Article, Product) on every relevant template. Existing schemas, not novel ones.
- Build a real
sitemap.xmland ping it on every meaningful content update.
5. The May 2026 algorithm update sitting underneath all of this
Don't read the guide in isolation. The same week, Google's May 2026 core algorithm update started rolling out, and the signal it is sending is unsubtle: E-E-A-T weight is up, and pages without a clear named author, credible publisher, or first-hand experience signal are losing both Search and AI visibility. The two events are linked. The guide is the "what to do." The algorithm update is the "what happens if you don't."
My Take
This guide is the single most useful SEO document Google has shipped since the Helpful Content Update. Not because of what it says you should do — most of that was already obvious. Because of what it explicitly says you should not do, which has been wasting consulting hours across the industry for two years.
The play for the rest of 2026 is brutally simple: stop building AI-specific infrastructure and start producing content only your company could have written. The technical floor is real but small. The content floor is real and most teams are nowhere near it. The win is in the gap.
Sources
- Optimizing your website for generative AI features on Google Search — Google Search Central
- Google's New AI Search Guide Calls AEO And GEO 'Still SEO' — Search Engine Journal
- Google Publishes How To Optimize For Generative AI With Mythbusting — Search Engine Roundtable
- Google's New AI Optimization Guide Just Killed 4 GEO Myths — Averi
- Google's Latest Algorithm Update: E-E-A-T, AI SEO — QC Fixer