SEO Isn't Dead. It Just Learned to Answer Questions
Trust Is the Deficit Everywhere
Broadcom spent three years pushing VMware customers into VCF-only bundles, watching SMB trust erode with every forced upgrade. Early September 2026, they rolled out vSphere Standard updates specifically to win that trust back. Not a new feature blitz. An admission that the old approach broke something, followed by an attempt to fix it the slow way.
Search engines are running the same repair job right now, just with less press coverage.
For years, ranking well meant gaming a system that rewarded keyword density and backlink volume over whether your page actually answered anything. Google's May 2026 AI-features guidance changed the terms. AEO and GEO — the disciplines built around getting cited inside AI Overviews, ChatGPT, and Perplexity — are not a separate track anymore. Google folded them directly into core SEO and said so in writing. There is no AI index sitting apart from the regular one.
That single decision matters more than it sounds. It means the tactics that used to work as a shortcut, the ones built for tricking a ranking algorithm, no longer have anywhere to hide. AI systems select sources they trust enough to quote by name. Trust, not trickery, is the actual currency now. VMware is learning that lesson with SMBs. Every website owner is learning it with Google.
One Discipline, Not Two Islands
The misconception going around marketing circles is that AEO replaced SEO the way streaming replaced video rental. New discipline, new rules, new specialists to hire. Google's own guidance says otherwise.
AI Overviews and chat assistants like ChatGPT and Perplexity are not pulling answers from some parallel AI-only index that lives apart from the search results you already rank in. They are pulling from the same crawled, indexed pages that show up in a normal Google search. If your page never got crawled and indexed properly for SEO, it does not exist for AEO either. There is nothing extra to plug into.
This kills the market for a specific kind of snake oil. For a while, agencies were selling llms.txt files and AI-only schema markup as the secret handshake to get cited in an AI Overview. Google's May 2026 guidance says standard crawlability and the same E-E-A-T signals that always mattered are enough. No separate submission process. No special file format that unlocks AI visibility.
Semrush tracked AI Overviews climbing from about 6.5% of queries in January 2025 to nearly 15.7% by November 2025. That growth did not require a new technical standard. It required content the algorithm already trusted enough to rank, now getting extracted and quoted instead of just linked.
Why Ranking First Isn't Enough Anymore
Ahrefs' 2026 numbers tell the uncomfortable part of this story. Top-page CTR dropped 58% on queries where an AI Overview appears. You can rank first and still watch your traffic disappear into a summary box nobody clicks past.
But Seer Interactive found the flip side in 2025: pages that got cited inside those AI answers saw CTR climb 38%. Same environment, opposite outcome, depending entirely on whether the AI decided your page was worth quoting by name.
This is the part that breaks people's mental model. Ranking and citation used to be the same reward for the same work. Now they are two different judgments happening on the same page. Google's crawler can decide you deserve position one. The AI layer sitting on top of that same index can look at your page, decide it is not extractable enough to quote, and cite a competitor sitting in position four instead.
The Princeton GEO research backs up why that split happens. Pages with quotations, real statistics, and cited sources saw citation rates lift 30-41%. Not ranking signals. Extraction signals. A page written in dense paragraphs with no clean, quotable claim can rank fine and still get skipped every time the AI needs a sentence to lift.
What to Fix on Your Site This Quarter
Skip the llms.txt file. Skip the AI-only schema markup someone is trying to sell you this quarter. Google's May 2026 guidance was direct about this: standard crawlability and the same E-E-A-T signals that always mattered are what get you cited. There is no special submission form for the AI layer.
What actually moves the needle is rewriting your existing pages so a sentence can be lifted cleanly out of them. Go through your top ten pages and find every claim sitting there unsupported. Add a real statistic with a source attached. Add a quotation from someone with actual authority on the topic, attributed correctly. The Princeton GEO research found this lifts citation rates 30-41%, and that gap is not small.
Then build answer blocks on purpose. Pick the exact question your page answers, write two or three sentences that answer it completely on their own, without needing the paragraph before or after for context. That block is what the AI copies. If your answer is buried inside six sentences of throat-clearing, it gets skipped for a competitor who put the answer in sentence one.
None of this requires new tools. It requires going back through content you already published and making it quotable instead of just readable.