Pedro Dias outlines practical AEO strategy as Google backs SEO fundamentals

Reality check: AEO is still SEO

The cleanest read on the current AEO debate is that the market is racing ahead of itself while Google keeps pointing back to the same old discipline. Pedro Dias is getting attention because his approach looks like the first practical version of AI search visibility that does more than feed the latest hype cycle.

That matters because Google has now said, plainly, that optimizing for generative AI features in Search is still SEO. Its guidance also tells site owners to ignore tactics like chunking content, unnecessary AI text files such as llms.txt, and inauthentic mentions. In other words, the company is not rewarding a new bag of tricks. It is telling publishers to get the basics right and make their sites easier for Search to trust, understand, and surface.

What Google is really backing

Google’s message has been consistent across its AI search guidance. In May 2025, it said the same underlying advice that has always applied to Search also applies to AI search experiences like AI Overviews and AI Mode. The priorities were not exotic: unique, satisfying content, strong page experience, crawlability, and structured data that actually matches what users can see on the page.

That framing matters because it strips the mystique out of “AEO” and “GEO.” AI features like AI Overviews and AI Mode are part of Search, not a separate universe with separate rules. If a site is thin, confusing, slow, or structured in a way that does not reflect the visible page, AI does not suddenly become more forgiving just because the interface looks different.

Google reinforced that stance again in May 2026 with an official guide saying that for Google Search, optimizing for generative AI features is still SEO. The company also published a separate guide, AI Features and Your Website, to explain how AI features work from a site owner’s point of view and how content can be included in those experiences.

Why the market keeps reaching for shortcuts

The problem is that the AEO and GEO market is crowded with easy-to-sell shortcuts. In 2025 and 2026, marketers have been promoting schema, prompt tracking, and llms.txt as if each one is a visibility lever with clean cause and effect. That is attractive because it sounds measurable, and measurable feels safer than doing the harder work of improving the actual site.

But Google’s own guidance pushes back on several of those claims. It has specifically told site owners to ignore chunking, unnecessary AI text files, and inauthentic mentions, which cuts against the idea that visibility can be gamed with standalone formatting tricks. The reality check is simple: if a tactic only exists to please a model and does not make the page better for people or clearer for crawlers, it is probably not a durable strategy.

That is why Pedro Dias’ name is getting traction in the broader debate. He is being discussed alongside arguments about whether schema truly helps AI systems understand content, or whether it mostly helps search features parse and classify pages. That distinction matters, because a tool that clarifies content is useful, but a tool that is treated like a magic key is just more vendor theater.

What a practical AEO strategy actually changes

If a team adopted the practical version of this strategy tomorrow, the first change would be in content structure. Pages would need to answer a real query cleanly, with visible sections that map to actual user needs rather than to invented AI prompts. The goal is not to create more content for a model to chew on. The goal is to make each page easier to scan, easier to trust, and easier to connect to a specific intent.

That means writing with fewer decorative flourishes and more explicit framing. A good AEO page is not a pile of paragraphs that happen to mention a topic many times. It is a well-organized resource with direct answers, supporting detail, and a clear relationship between the headline, the body copy, and any structured data on the page.

The second change would be in source authority. Google’s guidance around unique, satisfying content is not just a content brief, it is an authority test. Teams would need to lean on first-hand expertise, original analysis, and visible evidence that the page is worth surfacing, because generic rewrites and scaled content are exactly the kind of material that tends to vanish into the background.

The third change would be in information architecture. Sites would need clearer sectioning, stronger internal links, and a cleaner crawl path so that important pages are not buried under layers of fluff. Crawlability is not a side issue here. It is part of the core logic of whether Google can interpret the site well enough to use it in traditional search and in AI experiences.

A practical rollout would look like this:

• Tighten page templates so headings describe actual subtopics, not marketing language.

• Align structured data with what is visibly on the page, instead of adding markup that overpromises.

• Reduce duplicate or near-duplicate pages that dilute authority across the site.

• Build internal links around topic clusters that reflect how users move through the subject, not how the CMS happened to be set up.

• Remove content and markup that exists only to signal “AI readiness” without adding real value.

Where schema fits, and where it does not

Schema is still useful, but it is not a cheat code. Microsoft’s vendor guidance says schema helps search engines and AI systems understand content, which is true enough, but that is not the same as guaranteeing inclusion in AI answers or forcing a model to cite your page. The current conversation gets sloppy when people treat schema like a direct switch for visibility.

The better way to think about it is as a support layer. Schema can help define entities, relationships, and page types, but it works best when the page itself is already clear. If the visible content is muddled, the markup is not going to rescue it. If the visible content is strong, schema can help reinforce that structure for both search systems and AI experiences.

What Google’s AI rollout signals

Google’s broader AI push only strengthens this reading. In May 2026, the company said AI Mode had launched in the United States one year earlier, and it announced that Search was being upgraded with Gemini 3.5 Flash as the default model in AI Mode for everyone globally. That is a serious expansion of AI inside Search, but it is not a signal that the fundamentals got thrown out.

If anything, the rollout shows the opposite. Google is scaling AI features while still insisting that the right response is strong SEO. That is the key takeaway from this moment: the winners will not be the teams chasing every new acronym, but the teams that make their content more legible, more credible, and easier to retrieve across a search system that is becoming more AI-driven without becoming less dependent on solid SEO.