Google Officially Confirms: Stop Chasing GEO, AEO and AIO — Just Write for Humans

Insight drawn from Google’s Search Off the Record podcast, cross-referenced with the latest independent research through April 2026

There’s something quietly reassuring — and slightly maddening — about the fact that Google’s most senior search representatives just spent the better part of an hour on a podcast telling SEOs that the answer to all the AI search disruption is essentially: nothing has changed.

Danny Sullivan, Google’s Search Liaison, and John Mueller sat down on Search Off the Record to address the avalanche of questions arriving about GEO, AEO, AIO, and every other acronym the industry has conjured since AI Overviews went mainstream. The conversation cuts through the noise in a way that most SEO commentary doesn’t, because these are the people inside the machine.

Here is what they actually said — and what the data says alongside it.

Summary

• Google’s core ranking principle remains: reward content that human beings find satisfying, full stop.
• GEO, AEO, AIO, and similar acronyms are subsets of SEO, not replacement disciplines.
• The more you optimise for a specific AI format, the more fragile your position becomes when that format evolves.
• Original, authentic, multimodal content is the only durable strategy — it was already the right answer before AI search existed.
• Structured data helps but is not a magic ticket into AI responses.
• Success measurement needs to shift from volume of clicks to quality of conversions.
• AI Overview queries are generating 35% more organic clicks when your brand is cited in the AIO itself.

The Acronym Problem Nobody Wants to Admit

Sullivan opened with something that felt refreshingly honest. He joked about the sheer proliferation of new terms — GEO (Generative Engine Optimisation), AEO (Answer Engine Optimisation), AIO, and whatever else the industry coins next — and suggested we might as well call it “LNOPO” for how many letters we’re burning through.

The frustration behind the joke is real. Every new format spawns a cottage industry of “you must do X now” advice, courses, consultants, and toolsets. And the Google team’s response — after consulting their engineers and writing up a public blog post — was essentially: you don’t need to do any of it differently.

“Everything we do and all the things we try to improve, it’s all about how do we reward content that human beings find satisfying — and say ‘that was what I was looking for. That’s what I needed.'”

— Danny Sullivan, Google Search Liaison, Search Off the Record

That’s not a throwaway line. It is the clearest articulation of Google’s ranking intent that Sullivan has offered in years.

Timeline: How We Got Here

(See the interactive diagram above)

The pattern Sullivan described is not new. Back when Alta Vista, WebCrawler, and Open Text all had quirks of their own, publishers were creating six different versions of content for six different search engines. The nuances didn’t last. The search engines converged toward the same goal — serving user intent — and the bespoke optimisations became irrelevant. His point is that history is repeating itself with AI formats.

What the Data Actually Shows (Accurate, Sourced)

The Google guidance is philosophical. The data is brutal. Here is where they intersect.

Click-through rate collapse under AI Overviews:

Ahrefs re-ran its study using December 2025 data and found that the presence of an AI Overview correlates with a 58% lower average clickthrough rate for the top-ranking page.

Seer Interactive, tracking 3,119 search terms across 42 client organisations and 25.1 million organic impressions from June 2024 through September 2025, found organic CTR plummeted from 1.41% to 0.64% for queries with AI Overviews present.

Pew Research Center, tracking 68,000 real search queries, found that users clicked on results 8% of the time when AI Overviews appeared, versus 15% without them — a 46.7% relative reduction.

Similarweb data shows zero-click searches rising from 56% in 2024 to 69% by May 2025, meaning nearly seven in ten searches now conclude on Google’s results page without a click to any external site.

But being cited changes everything:

Brands cited in AI Overviews earn 35% more organic clicks and 91% more paid clicks compared to those not cited.

AI Overview prevalence:

Semrush analysis of over 10 million keywords found AI Overviews peaked at appearing for nearly 25% of all queries in July 2025 before settling to around 15.69% in November 2025, with volatility suggesting Google is continuously refining the feature.

As of January 2026, AI Overviews appear in 25.8% of all US searches, with informational queries triggering them 39.4% of the time, while e-commerce queries trigger them only 4% of the time.

This last statistic matters enormously for Shaiful’s client work — transactional and local queries (restaurants, clothing manufacturing) remain relatively insulated.

The “North Star” Principle Explained

Sullivan used the phrase “north star” to describe the fixed point that all of Google’s ranking systems — traditional and AI — are calibrated toward. It is worth stating clearly because the SEO industry has a tendency to treat each format update as a new destination rather than a different vehicle heading to the same place.

The north star is: does the content satisfy the human being who asked the question?

That’s it. Not: did you include FAQ schema? Not: did you use conversational H2s optimised for voice? Not: did you split your content into AIO-friendly atomic answers?

Mueller added a useful practical framing around this. He noted that CMS platforms like WordPress now handle most technical SEO concerns automatically — the gap between “technically accessible content” and “content Google can understand” has closed for the average site. What remains is the content quality question, which is entirely a human judgement.

Compare: Old SEO vs AI-Era SEO Thinking

The Commodity Content Warning

One of the most practically useful moments in the podcast was Sullivan’s discussion of commodity content — factual, undifferentiated information that any competent source could produce. He used the annual “what time is the Super Bowl?” frenzy as the example: thousands of sites writing long posts just to answer a simple factual query, until Google could pull the answer from a data feed and serve it directly.

The same logic now applies at scale to AI Overviews. If the information you produce could be synthesised from ten other public sources, an LLM can and will do exactly that — and you won’t get the click.

What AI cannot synthesise is your first-hand experience, your original research, your specific point of view. Sullivan was explicit: this is your genuine competitive advantage, and it was already the right strategy before anyone said “GEO.”

Tips: What to Actually Do

Focus on originality, not format optimisation. If your content has a perspective, an experience, or data that nobody else has, it is structurally protected from AI summarisation. If it doesn’t, no amount of schema or AEO tweaking will save it.

Go multimodal where it serves your audience. Sullivan described walking around Portland, filming geese, and getting an instant answer from the Google app. That search was a video query returning a text answer. The practical implication: if your content topic has a visual or instructional dimension, video and images aren’t nice-to-haves — they’re alternative search entry points.

Stop measuring clicks as your primary KPI. With AI Overview queries generating significantly fewer clicks, a traffic-first dashboard will make 2025–2026 look like a disaster even on a well-run site. The number to watch instead is conversion rate of the visitors who do arrive — Sullivan noted that visits from AI-format queries correlate with longer time on site, suggesting better contextual alignment before the click.

Structured data: use it, don’t obsess. The Google team included it in their recommendations, but specifically as a “think about this along the way” rather than a silver bullet. For sites already using Rank Math, the standard implementation is sufficient for most schema types.

Build a direct audience channel. Sullivan mentioned that his primary conversion goal when running his own sites was always capturing an email address — creating a continuing connection independent of search format changes. That logic has only strengthened.

Do not create a page for every sentence someone might search. This was said directly. Long conversational queries are now common because AI Overviews encourage exploratory dialogue. The answer is not to create a separate landing page for every permutation. Write comprehensively about your topic for humans; the systems will handle the matching.

Query Fan-Out: The Concept Worth Understanding

Mueller explained query fan-out clearly: when a user submits a query to an AI Overview, the system doesn’t just search for that exact phrase. It fans out into multiple related sub-queries, synthesises the results, and composes the answer. This is why you can rank #1 in organic results for a query and still not appear in the AIO for the same query — the AIO was effectively answering a different, broader question on your behalf.

Understanding this dissolves a lot of unnecessary anxiety. Optimising for the exact query becomes less important than being a comprehensive, trusted source on the broader topic.

Expert Opinions

Stuart Forrest, Global Director of SEO at Bauer Media, told the BBC that the industry is “moving into the era of lower clicks and lower referral traffic for publishers,” acknowledging it as a structural shift rather than a temporary disruption.

Lily Ray, VP of SEO Strategy at Amsive, has raised concerns about quality control within AI Overviews, noting repeated examples of spam, misinformation, and inaccurate results appearing in live responses since the feature launched.

SEO specialist Barry Adams has argued that AI Overviews are poor at citing appropriate sources, and that using LLMs for fact-dependent search responses is “fundamentally wrong.”

These perspectives matter because they frame where the gaps are — and gaps in AI accuracy are precisely where human expertise and original reporting have durable value.

External Resources

• Google Search Off the Record Podcast — the source conversation
• Google’s Search Central SEO Guide — official guidance referenced in the podcast
• Seer Interactive AIO CTR Study (September 2025) — 25M impression dataset
• Ahrefs AI Overviews CTR Update (December 2025) — 300,000 keyword study
• Semrush AI Overviews Research — 10M keyword analysis

FAQs

Does Google treat GEO and AEO as separate disciplines from SEO? No. Danny Sullivan explicitly said these are subsets of SEO — a specialisation, much like local SEO, not a replacement framework. Optimising for a specific AI format without the underlying content quality is optimising for a moving target.

Do AI Overviews hurt all sites equally? No. Industries like Science, Computers & Electronics, and People & Society see AI Overviews on over 17% of their keywords, while Real Estate, Shopping, and Arts & Entertainment see them on less than 3%. Transactional and local queries are significantly less exposed.

If I’m cited in an AI Overview, is that better than ranking #1 organically? Based on current data, citation in an AIO produces 35% more organic clicks than ranking without citation. But the path to citation is the same as traditional SEO — comprehensive, trustworthy, original content.

Should I rewrite all my content for AI formats? That’s exactly what Google says not to do. Dramatic shifts toward AI-specific formatting introduce instability. The formats will evolve; the content quality principle won’t.

Is structured data more important now with AI search? Helpful, yes. Essential, no. The Google team described it as worth thinking about alongside your other work — not as the determinant of AI visibility.

What does “multimodal” actually mean for my content strategy? Practically: if your topic can be shown as well as described, add video and images that are original to your site. These create additional search entry points beyond text queries and appear to perform well in AI-assisted search results.

What metric should I replace traffic with? Conversion rate of sessions, email captures, time on site, and direct audience growth (newsletter subscribers, return visitors). These metrics are less susceptible to AI Overview interference and reflect genuine business value.

The core message from Google is almost annoyingly simple: write content that genuinely helps real people, presented across the formats they actually use, and measured by what it achieves for your business rather than how many clicks it generates. The AI era changes the delivery mechanism. It doesn’t change what earns trust.