Loading page...
Loading page...
In May, Google published a guidance piece that quietly tried to put a lid on the AEO and GEO conversation. The argument: llms.txt is not necessary, content chunking is not required, AI-specific rewriting is not a real category, and special schema for AI search is not a thing. From Google's perspective, "Answer Engine Optimization" and "Generative Engine Optimization" are not separate disciplines. They are just SEO.
Meanwhile, in the same week, the actual numbers landed. ChatGPT now handles more than 2 billion queries daily. AI-referred sessions to websites grew 527 percent year over year through mid-2025 and have continued to climb. The data and the corporate guidance do not say quite the same thing.
Google is partly right. They are also partly defending their turf. Here is the framing we are using with clients, and the practical checklist that comes out of it.
The technical claim that AEO and GEO require an entirely new content discipline is overstated. A well-optimized SEO page (clear hierarchy, good structured data, fast load time, expert content, citation-worthy claims) does most of the work for AI search engines as well. The argument that you need to maintain a parallel set of "AI versions" of your content is, as Google says, mostly noise from people selling consulting services.
Specifically, the following claims are overhyped:
llms.txt as a required technical artifact. It is not required. It is not enforced. Most major LLM search systems do not read it. If your content is good and your structured data is clean, you do not need it.
Special schema markup for AI search. Standard Schema.org markup is fine. The "AEO schema" you may have seen in vendor pitch decks does not actually exist.
AI-specific content rewrites. A piece written for a human reader is, in most cases, also a piece that AI search engines will summarize and cite. Rewriting it twice is wasted effort.
To the extent that "AEO" was sold as a new SaaS category that you need a new vendor to handle, Google is right to push back. Most of that pitch was overhyped.
The framing that "it is all just SEO" misses three meaningful shifts that are happening at the same time, and they are not best characterized as "the same thing we have always done."
In traditional search, the user clicks. The site gets the visit. The conversion happens on the site.
In AI search, the user reads the AI's synthesized answer. The site may or may not get the click. The conversion has to happen earlier in the funnel or be triggered by the AI's citation rather than by direct traffic.
The implication is not that AEO is a new technical discipline. It is that the metrics you measure success against are different. "Organic traffic" used to be a sufficient proxy for content ROI. In an AI-search world, "citations in AI answers" is a real, separate metric. Tracking it requires different tooling, even if the content production process looks similar.
Traditional search ranks sources by relevance and authority. AI search synthesizes from sources and weights them by perceived trustworthiness, recency, and specificity. The result is that shorter, more authoritative pages can punch above their weight in AI search in ways they could not in traditional search.
This is not just a technical detail. It changes the content strategy in a real way. Specifically: a 400-word, highly specific, citation-dense piece that is exceptionally accurate on a narrow topic can show up in an AI answer where a 4,000-word "comprehensive guide" gets passed over.
If you are still writing the long-form-comprehensive-guide as your default content format because it ranked in 2019, you may be missing the cite-worthy shorter format that AI search is actively preferring.
Traditional search has one distribution surface: Google's results page. AI search has many: ChatGPT's answer, Gemini's answer, Perplexity's answer, Claude's web-aware answer, the AI Mode results inside Google itself, and the agentic search experiences inside other tools.
Each of these surfaces has its own quirks. Perplexity heavily weights recency. ChatGPT favors well-structured FAQ-like content. Gemini integrates more with Google's existing index than the others. Optimizing for all of them is not the same as optimizing for traditional search.
To Google's credit, optimizing for traditional search is still a strong baseline that helps with most AI surfaces. But "it is just SEO" undersells the fact that you now have five surfaces to think about rather than one.
Here is the actual checklist we run for clients trying to be visible in AI search, without buying the overhyped vendor pitches:
Standard SEO basics, done well. Clean structured data. Fast page speed. Mobile-friendly. Clear heading hierarchy. Internal linking. None of this is new. It is still the foundation.
Concrete claims with cited sources. AI search systems weight content that makes specific claims and backs them with citations. Vague content with no sources does not get cited. Specific content with credible sources does.
Author expertise made explicit. Bylines, bios, credentials, and clear attribution. AI search systems use these signals heavily. Anonymous content gets cited less.
Direct answers to specific questions, near the top of the page. AI search systems pull from content that is structured to answer a question directly. "What is X" content where the first paragraph actually defines X is more citable than content that buries the answer 800 words in.
Recency signals. A clear updated-at date, periodically refreshed content, references to current events. Most AI search systems weight recency. Evergreen content that is genuinely refreshed wins over evergreen content that has not been touched since 2022.
A way to measure citations. Tools like HubSpot AEO Sensor, Perplexity's source tracking, and manual ChatGPT queries against your brand. The goal is to know whether you are being cited, not just whether you are getting traffic.
That is the list. Six items. No special schema. No llms.txt. No new vendor.
Google's "it is just SEO" framing is partly a useful corrective against vendor hype and partly a defensive move from a company that is rapidly losing its position as the default discovery layer. Both can be true at the same time.
The practical implication for businesses: take the corrective seriously (do not spend money on AEO snake oil), but do not take the framing as a reason to ignore the structural shift. The structural shift is real, the metrics are different, the surfaces are different, and the content patterns that win in AI search are not always the same as the patterns that won in traditional search.
The studios that get this balance right are doing solid SEO basics, tracking AI citations as a separate metric, and tuning their content for the specific surfaces where their audience actually shows up. The studios that get it wrong are either ignoring AI search entirely or buying every AEO vendor pitch they see. Neither extreme is the answer. The middle is where the visibility lives.
Want help auditing your content for AI search visibility without the vendor hype? Book a free audit call and we will run the six-item checklist on your top 10 pages.
Sources:
Like what you are reading? Book a free intro chat and we will show you what AI can do for your business.