Google Just Confirmed GEO — But Strategy Still Wins
Key Takeaways
Google's May 15 guide says AEO and GEO are "still SEO" — and names five specific tactics (llms.txt, content chunking, AI-only rewrites, mention-buying, special schema) as unnecessary for its AI surfaces.
The single biggest visibility lever Google identifies is non-commodity content — original, first-hand, expert-grade work that can't be synthesized from somewhere else.
The reason your competitors get cited, and you don't, is rarely about size. It's about structural and entity evidence that AI engines can find, trust, and reuse.
68% of AI citations come from third-party sources. Your owned content alone is structurally insufficient.
The post-May 2026 playbook: rank in Search, structure for extraction, build third-party evidence, and make every post sound like something only you could have written.
The title on Google’s latest insights on AI is plain enough: "Optimizing your website for generative AI features on Google Search."
But inside, Google does something it has avoided for two years. It addresses the AEO and GEO terminology debate directly and says, on the record:
"From Google Search's perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO."
Yeah, you can bet that the internet split in half the moment that came out.
One camp took their obvious victory lap — we told you GEO was hype.
The other camp scrambled to defend their retainer decks — Google doesn't speak for ChatGPT.
But the reality is that both takes are missing what really happened.
What Google did was retire the hacks. It didn’t do away with the discipline.
The work I've been talking about here for months now — where AEO and GEO aren't competing strategies but the same craft applied across different surfaces — is more important now than it was a week ago.
In short, the companies that sold $79/month "GEO file generation" subscriptions just got publicly de-shoed.
Why Isn't My Content Showing Up in AI Search?
This is the question my clients are asking — even when they phrase it differently.
They're watching competitors get quoted in ChatGPT answers, cited in Perplexity, surfaced in Google AI Overviews.
Some have spent on GEO. The dashboard shows traffic flatlining. The instinct sits somewhere between did I waste that budget? and am I about to be invisible?
The reason your content is invisible in AI search is almost never because you skipped a specific AI optimization hack. It's because — at the structural level — AI engines can't easily retrieve, trust, or summarize what you've published.
That's a craft problem, not a tooling problem. And it's a reason to stop panicking about AI search and start fixing the foundation.
What Did Google's New AI Search Guide Say About AEO and GEO?
Short answer: Google said that optimizing for its generative AI features — AI Overviews and AI Mode — is still SEO.
The differences AEO and GEO consultants have been selling as separate disciplines do not exist from Google's perspective. The same core ranking and quality systems that decide what appears in regular Search decide what appears in AI answers.
The reason this matters has nothing to do with the terminology. It has to do with what Google also said in the same guide.
Google's AI features are powered by two underlying mechanisms:
Retrieval-Augmented Generation (RAG). Google's AI grounds its answers by pulling relevant pages from the Search index. No organic ranking, no AI citation.
Query fan-out. A single user query triggers 8–12 concurrent sub-queries (hundreds in Deep Search), and the model synthesizes a unified answer from all of them.
The strategic implication is unambiguous: you cannot be cited in AI answers without first being indexed and ranked.
The Five GEO "Hacks" Google Officially Killed
Inside the article’s "Mythbusting" section, Google names specific tactics and says you don't need them for its AI surfaces:
This is the most direct repudiation Google has ever issued of the AEO/GEO services industry. Vendors who priced "AI optimization audits" as a separate product are now selling against Google's own published documentation.
What Google Says You Should Be Doing
The guide's "do these things" list collapses to three real directives:
Create non-commodity content
Google explicitly contrasts commodity content (the generic "7 tips for first-time homebuyers" piece) with non-commodity content (a documented case, original data, a perspective that can't be synthesized).
The March 2026 Core Update re-weighted Information Gain — meaning Google now has algorithmic means to detect whether your page adds anything new to what already ranks.
Maintain technical clarity
Same ol, same ol. Indexable, crawlable, snippet-eligible pages with strong page experience are non-negotiable.
Avoid scaled variation
Pumping out near-duplicate pages targeting every long-tail variation now violates Google's scaled content abuse policy. The "more pages = more visibility" playbook rarely worked in the traditional age; it definitely won’t work now.
Danny Sullivan, speaking at Search Central Live Canada 2026, defined non-commodity content by three mandatory ingredients: a specific case behind the piece, first-hand experience from someone with a verifiable track record, and original data that doesn't exist elsewhere in the same form.
Why Does AI Cite My Competitors but Not My Brand?
Not because they're bigger. Because they have more discoverable, structured, third-party evidence for AI to find, trust, and reuse.
This is the most common question my clients ask right now, and they usually have a specific competitor in mind. The honest answer is structural, not random.
When ChatGPT, Perplexity, Claude, or Google AI cites a competitor, three things had to be possible:
The model could retrieve their content — indexed, crawlable, technically accessible.
The model could trust it — third-party sources corroborated the brand's expertise.
The model could summarize it cleanly — structure made the content extractable.
If any one of those three breaks, you disappear from the answer. And here's the part most brands miss: high Google rankings do not guarantee AI citations.
Brands that dominate traditional SERPs can be completely absent from AI answers when their structural and entity signals don't hold up.
Here are the four reasons that show up most often in my client audits.
1. They Have More Independent Third-Party Mentions
This is the big one. Per Erlin AI's 2026 analysis of more than 500 brands, 68% of AI citations come from third-party sources — only 32% come from brand-owned websites.
Translation: if all of your authority lives on your own site, AI has nowhere else to verify you.
A brand cited in a G2 review, a Reddit thread, and two industry publications hits 78% average AI citation coverage. A brand cited only through its own content: 18%.
Authentic PR, partner content, podcast appearances, and earned media are not awareness plays anymore. They're GEO investments — and they're what AI citation means for your business.
Source: Erlin AI, 2026 analysis of 500+ brands
2. Their Content Is Structurally Built to Be Cited
ChatGPT cites approximately 15% of the pages it retrieves. The pages that win citation share have three traits:
Answer-first formatting. A self-contained, quotable answer in the first 1–2 sentences under every H2. The first 30% of a page accounts for nearly twice as many citations as the final 30%.
FAQ schema. Pages with 3+ JSON-LD FAQ nodes are cited at 41% in Perplexity vs. 24% for controls — a roughly 40% weighting boost in ChatGPT.
Verifiable data points. Perplexity specifically favors content with unique, citable facts. Marketing fluff gets ignored.
Format matters too. Listicles account for 35.6% of all AI citations — not because lists are intrinsically better, but because they package answers in a structure AI can extract directly.
3. Their Entity Confidence Is Higher
AI engines build entity models of brands — structured representations of who you are and what you stand for — and they use those models to decide whether you're a safe choice to cite.
The signal AI looks for is called entity confidence: cross-platform consistency that AI uses to decide who to cite.
If your brand name, expertise area, and core positioning are consistent across your website, LinkedIn, Wikidata, Crunchbase, and third-party publications, your entity confidence rises. If it's inconsistent or thin, you're invisible — even when your content is good.
This is the part nobody is talking about enough: voice consistency across platforms is itself an AI search signal. It trains the models to associate your name with specific expertise.
When your voice drifts every time a different freelancer writes for you, your entity confidence drifts with it.
4. They Show Up in "Best" and "Alternatives" Lists
AI engines heavily use category lists ("best X tools," "top providers," "alternatives to Y") because the structure makes recommendations directly extractable. If a competitor appears in five credible category lists and you appear in zero, AI has a structurally easy reason to exclude you.
This is where unglamorous outreach work — getting included in industry roundups, comparison pieces, and category reviews — punches well above its weight for citation visibility.
What Does ChatGPT, Perplexity, and Google AI Cite?
Original, expert-grade, voice-distinct content that's structurally built to be extracted. The three platforms differ in how they retrieve, but they converge on the same underlying answer.
Google AI Overviews + AI Mode
Google's AI surfaces operate on top of the existing Search index via RAG and query fan-out. That means a page that doesn't rank in organic Search will not appear in AI Overviews. Period.
The signals that differentiate AI citation include answer-focused structure (each section makes complete sense in isolation), question-based H2s, contextual completeness (answer the main question and the obvious follow-ups), and non-commodity depth.
One non-obvious data point: YouTube holds a 200x citation advantage over every other video source in Google AI Overviews. Original video is one of the most underused citation surfaces for B2B brands.
ChatGPT (Search-Enabled)
ChatGPT triggers live web retrieval on roughly 76% of interactions. In that mode, it relies heavily on Bing index ranking.
A page that doesn't rank in Bing won't be cited. But when a page does rank and its title matches ChatGPT's internal sub-questions, citation rate jumps above 88%.
The structural signals that move the needle: TL;DR summaries, FAQ schema, inline citations, and front-loaded answers.
Third-party authority comes mostly from Wikipedia, LinkedIn, Reddit, and editorial domains. Journalism accounts for 27% of all AI citations — and rises to 49% on time-sensitive queries.
Worth flagging: ChatGPT's citation behavior shifted with GPT-5.4. Reddit citation share dropped from ~60% to 10% in six weeks last fall after a single Google parameter change. Forbes, PR Newswire, and Medium absorbed the gap.
Citation share is volatile. Build authority across multiple credible third-party surfaces — don't bet your visibility on one.
Perplexity AI
Perplexity is the most explicit and trackable of the three. Every source is numbered and linked. Average citations per answer: 5+. The factors that determine selection:
Crawler access — PerplexityBot reads static HTML at 94% parse rate, JavaScript-rendered content at 23%.
Direct, front-loaded answers near the top of the page.
Current, verifiable facts (AI citations average 25.7% newer content than traditional search).
3+ JSON-LD FAQ nodes — shortens time-to-first-citation by about six hours.
8+ structured brand attributes — brands with 8+ get cited 4.3x more than brands with fewer than 3.
Third-party source authority across G2, Reddit, and industry publications.
If your Perplexity citation rate is zero, start with the crawler check. You'd be surprised how often the answer is "the bot can't read the page."
Source: Erlin AI Perplexity SEO research, 2026
The Post-May 2026 GEO Playbook
If Google's guide ended the era of GEO hacks, what replaces them?
Three shifts. None of them are new. All of them are now table stakes — and together they're the six-part AI-smart content framework I use with every client.
Shift 1 — Structure for Extraction
Stop optimizing only for "readability." Start optimizing for extractability. The pages that get cited share six structural elements that give every page the strongest chance of being cited:
A TL;DR block in the first 100 words of every substantive post
Answer-first H2s where each section opens with a 1–2 sentence direct answer
At least one "quotable claim" sentence per section — self-contained, citable
A minimum of 3 JSON-LD FAQ nodes per post
An explicit definition of your core concept near the top
Original media (image, video, chart) that can be retrieved separately
Notice how this post is structured. The TL;DR up top. The answer-first opener under each H2. The quotable claims. The reason every section here looks the way it does is because the post is itself an example of the work.
Shift 2 — Signal Authority with On-Page Evidence
Non-commodity content has three ingredients Google can verify:
A named author with external presence (LinkedIn, bylines, speaking)
Inline data from original research, client work, or first-hand observation
A specific case behind the piece — not "what brands should do," but "what happened when we did this"
The part most teams underinvest in is cross-platform entity consistency. Your name, expertise area, and positioning should be aligned across your site, LinkedIn, podcast appearances, and anywhere else you publish. Your internal links are the map AI uses to navigate that consistency on your own site. Broken or inconsistent internal navigation = a brand whose own ecosystem can't corroborate itself.
Shift 3 — Keep the Voice Human
This is the shift I care about most, because it's the one nobody else is selling and the one AI tools cannot generate.
Brands with distinctive personalities see 20% higher customer retention. Human-generated content with genuine personality gets 5.44x more traffic than AI-generated content.
And 83% of people can detect AI-generated content — and when they detect it, they disengage.
Source: HeyThereHumanoid, 2026 AI brand voice study
The good news: AI engines have started building this into their own evaluation. Voice consistency across platforms trains the models' entity models. A distinctive, consistent voice doesn't just earn human trust. It earns AI confidence.
The frame I use with clients: AI should amplify your voice, not replace it. Use AI for structure, research, and first drafts. Then make every published piece sound like something only you — or your client — could have written, based on direct experience.
If a GPT prompt can assemble it, so can everyone else.
What This All Means for Your Brand and Content Moving Forward
Google didn't kill GEO. It exposed the people who sold GEO as a separate product.
The work that matters now is the work that mattered before — done with sharper structural awareness and a clearer understanding of how AI engines build trust:
Rank in Search. The foundation is the foundation.
Structure for extraction. Make your content easy to retrieve, parse, and quote.
Build third-party evidence. 68% of citations live outside your site. Invest accordingly.
Sound like something only you could have written. Voice is the moat AI cannot copy.
If your KPI dashboard still only shows rank tracking and organic clicks, it needs a rebuild. The teams winning this next cycle are measuring eight GEO metrics that indicate whether AI is citing your brand — not just whether anyone is clicking through.
If you're looking at your content right now and wondering whether AI search is even seeing it — or whether your brand voice is sharp enough to be cited by a model that has already read everyone else's — that's the right question.
A practical place to start: my three-step system to check if AI is driving traffic to your site. Run it on your top five pages. The answer will tell you whether the rest of this post applies to you, urgently or eventually.
Or send me a message. The conversation is usually shorter than people expect.
Frequently Asked Questions
GEO, AEO, and Google's May 2026 AI Search Guide
Did Google's May 2026 update kill GEO and AEO?
No. On May 15, 2026, Google clarified that AEO and GEO are "still SEO" — meaning optimizing for AI Overviews and AI Mode uses the same core ranking systems as traditional Search. What Google retired were specific tactics: llms.txt files, content chunking, AI-only rewriting, and special schema. The discipline of optimizing for AI citation is more valuable than ever. The hacks are not.
What is non-commodity content?
Non-commodity content is Google's term, introduced in its May 2026 AI search guide, for original content built on direct experience, first-hand cases, or proprietary data that cannot be synthesized from sources already available online. Commodity content — generic "7 tips for X" pieces assembled from common sources — is now algorithmically discounted by Google's March 2026 Core Update under the Information Gain signal.
Do I need an llms.txt file to appear in AI search?
No, not for Google's AI search features. Google's May 2026 guide explicitly states that llms.txt files and other special AI markup are not required to appear in AI Overviews or AI Mode. The file may still offer marginal value for Anthropic's Claude and smaller open AI surfaces, but it is not a Google ranking factor and should not be sold as one.
Why does ChatGPT cite my competitors but not my brand?
Most often because your competitors have stronger third-party evidence — not because they're bigger. Research shows 68% of AI citations come from third-party sources like industry publications, G2, and Reddit, not brand-owned websites. Brands with 8+ structured attributes verifiable across the open web are cited 4.3x more by Perplexity than brands with fewer than 3. Owned content alone is structurally insufficient.
Can my content appear in AI Overviews without ranking in Google Search?
No. Google's AI Overviews and AI Mode operate on top of the existing Search index using retrieval-augmented generation (RAG) and query fan-out. A page that is not indexed and ranking in organic Search cannot be retrieved by Google's AI features. Strong traditional SEO is the precondition for AI Overview visibility — not a separate or parallel strategy.
Is brand voice a GEO ranking signal in 2026?
Yes, indirectly. AI engines build entity confidence models — structured representations of who a brand is and what it stands for — using cross-platform consistency in name, expertise, and positioning. A distinctive, consistent brand voice across your website, LinkedIn, and third-party publications increases entity confidence and makes AI engines more likely to cite you specifically. 83% of readers can also detect AI-generated content and disengage when they do.
How long does it take to get cited in AI search?
For Perplexity, pages with 3+ JSON-LD FAQ nodes can shorten time-to-first-citation by approximately six hours after publishing. For Google AI Overviews and ChatGPT, citations typically follow standard indexing timelines (days to weeks) once the page is ranking. Brand-level entity confidence — and the third-party evidence that supports it — takes months to accumulate and is the longer-term driver.
Written by
Brad Bartlett
Brad is a copywriter and content strategist who helps creators, brands, and organizations build content that's actually worth reading — and built to be found. He specializes in conversion-focused copy, brand voice, and SEO and AI search optimization, with a straightforward philosophy: great content has to be authentic before it can perform. He works comfortably across the AI content space, helping clients use the tools without losing the voice. Fiverr Pro vetted, 4.9 stars out of 5 across 1,600+ clients.