May 5, 2026
While the industry panics over Generative Engine Optimization (GEO), Moz’s Dr. Pete Meyers warned audiences at Brighton SEO 2026 that abandoning traditional SEO is a critical misallocation of resources.
Predicting a hybrid search landscape where 85% of intent stays on feature rich SERPs and only 15% routes to pure AI mode, Meyers argues that AI models rely heavily on search grounding. Because LLMs pull their answers from the traditional retrieval layer, classic signals, especially off page brand mentions, are now the load bearing pillars of AI visibility.
The takeaway? Stop optimizing for the AI interface, and get back to building the foundational brand authority that feeds it.
Why GEO chasers misallocated their year
Pete Meyers spent the last few years preparing Moz customers for an AI-shaped SERP. He sat down with us at Brighton SEO 2026 hours before his closing keynote, The Infinite Tail: Keyword Research for AI, with a forecast that the rest of the industry is going to find inconvenient.
His own abstract describes the territory as "the evolving landscape of hybrid search combining AI and organic results". This is what the conversation walked through.
Roughly 85% of search intent is going to land on a feature-rich SERP. About 15% is going to route to AI mode. Search grounding will keep the LLMs anchored to the same web that traditional SEO has been organising for 25 years.
The teams that pivoted their entire roadmap to GEO already mis-allocated the year.
That is the bet behind Google's Web Guide test. It is also the bet behind why Pete is still telling Moz customers to invest in things that look uncomfortably like SEO.
Why is Pete Meyers's forecast worth listening to?
Pete is the Principal Innovation Architect at Moz, the SEO-data company that sits inside Ziff Davis. Before that, he was on Moz's marketing side as a marketing scientist tracking Google algorithm changes, creator of MozCast, and a long run of Whiteboard Friday appearances. Several years ago, he moved closer to the product side, and his job description compressed.
That compression is the whole reason this conversation happened. Pete's customers used to be two or three years behind the curve, and Moz had time to drag them forward. Now the customers are calling Moz asking where do we need to be? and Pete admits the honest answer for most of the past year was "we're all a little lost right now."
The Pete Meyers worth listening to is the one whose product roadmap depends on guessing the next six months correctly when nobody else has. Moz's new Prompt Explorer is the marketing layer. The bet sits underneath it.
Why is search grounding pulling SEO back to the foundation?
The story Pete kept coming back to was the double whammy. LLMs that used to answer from training data are now grounding their answers in live search.
So when ChatGPT or Gemini responds to a query, the same retrieval layer that ranks a Google SERP is shaping which sources the model summarises. The SEO levers that were supposed to be made obsolete by AI search are now load-bearing for AI search.
Ahrefs ran the correlation across 75,000 brands tracked in Google AI Overviews. Brand mentions across the web correlate with AI visibility at 0.664. Branded anchors sit at 0.527. Domain rating at 0.326. Backlinks at 0.218.
Brand mentions outperform every traditional SEO signal Ahrefs measured (Ahrefs, May 2025; the pattern was confirmed across ChatGPT and AI Mode in Ahrefs's December 2025 follow-up). Pete framed the same shift in conversation: the visible move is from on-page to off-page, where other people talking about you starts to matter more than what you publish.
Google's Web Guide makes the foundational role of search even more obvious. Web Guide is a Search Labs experiment that uses a custom Gemini to organize results into AI-clustered topic groupings, powered by query fan-out: parallel related searches the system stitches together based on intent (Google Blog, July 2025). The presentation layer is generative. The retrieval layer is the same SERP everyone has been optimizing for.
The team that won Google rankings in 2024 has a head start on Web Guide. The team that pivoted to "ChatGPT-first" content production is now optimizing for the interface, not for the retrieval layer that decides what the interface shows.
Ready to start scaling your business?
Let's get started
Should SaaS teams stop investing in structured data?
This is the question every in-house SEO has been ducking since AI Overviews launched. Pete's answer in conversation was direct.
The receipts come from Mark Williams-Cook (replicating an earlier test by Richard Barrett), who built a fake company called DUCKYEA t-shirts and gave it a completely invented schema vocabulary: @type: MallardEnterprise, nestingGrounds, featherCode, region: South Pondshire.
No address anywhere in the visible page text. The whole address lived inside the bogus structured data. Then he asked ChatGPT and Perplexity for the company's address. Both LLMs returned "77 The Muddy Bank, South Pondshire, DK99 YEA, United Queendom" verbatim from the made-up schema (Williams-Cook, October 2025).
Williams-Cook's own takeaway: LLMs are reading whatever text sits inside the <script> tag and using it as content. The structural meaning of the schema is irrelevant; the bytes inside the tag get ingested as plain text.
That tracks with what he showed earlier at the training layer: "@type": "Organization" tokenises into the regular words "type" and "Organization" with no difference from any other text on the page. Julio C. Guevara replicated the same finding through a controlled two-page extraction test (Search Engine Roundtable, September 2025).
That does not mean drop schema. Williams-Cook himself says it directly: Google still uses its Knowledge Graph with its AI systems, structured data reduces ambiguity, rich features in traditional SERPs need it, and Pete's forecasted 85% feature-rich routing is exactly where rich features live.
It does mean the schema-as-AI-citation story that vendors keep selling is overstated. If you are choosing between fix the schema first or get other people talking about your brand first, Pete's data and Ahrefs's correlations both point the same direction.
What does "soft brand" mean for prompt tracking?
Pete's most useful new framing was the soft brand idea. The standard branded-vs-non-branded split assumes the prompt either contains your brand or it does not. Pete pointed at a third category that the binary misses.
Pete's "branded problem" framing means a problem that standard branded-prompt tracking would catch. The soft-brand category slips through, because the prompt reads as non-branded even when the answer comes back loaded with brand names.
The category is brand-biased even when the prompt is not. SEO tools, project management software, CRM platforms, and accounting software: an LLM responding to "best tool for X" returns brand names whether the user asked for them or not.
Pete warned against the inverse trap: if you put your brand in the prompt, the LLM will obediently put your brand in the answer, because LLMs are suggestible. Hundred percent branded visibility on branded prompts means the model is agreeing with you. Tracking it tracks the model's compliance, nothing more.
The diagnostic worth running: in your category, are LLMs already adding brands to non-branded prompts? If they are, off-page brand-building is the stronger play. If they are not, you are competing on entirely different signals.
Why does Pete forecast a roughly 15/85 hybrid SERP?
Pete delivered the hypothesis with the number flagged as a guess.
The number is a guess. The mechanism is not. Web Guide is already a working version of intent-based query routing inside Google Search Labs. The same query fan-out tech that powers AI Mode also powers Web Guide, which keeps the clickable web in front of the user.
AI Mode handles the queries LLMs do well: long-form synthesis, the "I'm 55, training for a marathon, and my doctor said ease up on protein" prompt that no human is ever going to write a static article for. Web Guide and feature-rich SERPs handle everything else.
The implication for the budget is the part that the agencies will not tell you. Pivoting the SEO team to GEO sounds like preparation. It is misallocation if 85% of intent stays on routes that reward search grounding.
Off-page brand-building, brand mentions on Reddit / Quora / G2 / YouTube, structured content that Web Guide can cluster: this is the same playbook that wins AI Mode citations. The 15% is real and worth tracking. It is a wedge of intent worth winning, while the foundation underneath it is what determines whether you ever get into the wedge in the first place.
I imagine a world in a year or so. A query comes in, and Google goes, “Hey, you know what, this needs how-to, or this needs a long structured answer, or this is local, or this is branded.” LLMs are really bad at navigational content. They just want to go to your website. Maybe you're gonna go straight to AI mode 15% of the time, just making up a number, but the other 85% you're gonna go to something that's more feature-rich, more organic.
- Dr. Pete Meyers
What does this mean for the next six months?
Pete's case is narrower than "SEO is unchanged". SEO has changed in real ways. His argument is that the work the SMB and agency teams should be doing is the work they were already supposed to be doing, and the AI-eats-search panic has them shopping for the wrong tools.
The shift is from publishing more content for the LLMs to get more places talking about you so the LLMs and the SERP both pick you up. Ahrefs's 0.664 correlation on brand mentions is the data point Pete did not name, but every part of the conversation kept pointing toward.
The job description for an SMB SEO lead in 2026 looks more like PR-meets-content than like keyword research.
The risk Pete flagged at the end of the conversation is the one that breaks the model. If publishers stop publishing because AI eats the clicks, the LLMs run out of corpus to ground on.
The wave of publisher copyright lawsuits hitting OpenAI, including Ziff Davis's April 2025 filing, is an early sign that the AI side of the industry knows this, too. Pete's bet is that some balance will get struck within a couple of years. His job, and ours, is to be ready for either outcome.
Head of SEO
As Head of SEO at Seeders, Hans helps companies scale smarter by combining search behaviour, paid performance, and conversion insights into systems that drive real growth. Recognised as a Top 20 Content Marketer by Small Business Trends, he brings a lean, experimental approach focused on building what actually works. Rather than relying on reports or one-off strategies, Hans focuses on creating marketing systems and tools that turn visibility into measurable business impact. His goal is simple: help clients move from scattered efforts to structured growth.
