Blog · AI
— AI··6 min read

Google AI Mode vs. AI Overviews in 2026: What Actually Changed (and What’s Merging)

Joona Heinonen· Choco Media · Rovaniemi

As of May 2026, the short answer is: they’re merging. Google announced it is folding AI Overviews and AI Mode into a single AI-powered search experience, so the distinction that SEO teams have spent two years debating is becoming less relevant by the month. But the merge doesn’t erase what each surface trained us to expect — it combines them, and understanding the two halves is still the fastest way to understand what the combined experience will reward.

Here’s the version we’d give a client who asked us to explain this in five minutes.

What Was AI Overviews, Exactly?

AI Overviews is the answer block that appears at the top of regular Google search results, synthesizing several sub-queries into one summary with citation chips linking to source pages. It launched in May 2024 and became Google’s default treatment for a huge share of informational queries — “how to,” “best,” comparison, and definitional searches. You didn’t opt into it; if your query matched the pattern, you saw it, sitting inside the search results page you already knew.

What Was AI Mode?

AI Mode was a separate, dedicated tab — a conversational search experience Google rolled out in the US from May 2025 and internationally from October 2025, built on the Gemini model family. Instead of one summary block, AI Mode breaks a query into as many as sixteen parallel sub-queries, synthesizes them into a longer conversational answer with inline citations, and remembers context across follow-up questions in the same session. It was built for multi-step research and comparison shopping — the kind of query where you’d normally open six tabs.

Why Does the Distinction Matter Less Now?

Google’s May 2026 announcement described the goal as “one seamless AI Search experience” that accepts more input types — text, images, PDFs, video, and even open Chrome tabs — and carries context across a session rather than resetting with every query. Practically, that means the sharp line between “you get a summary box” and “you get a conversation” is dissolving. The behaviors that made AI Mode distinct — query fan-out, multi-turn memory, deep synthesis — are becoming the default treatment across Google Search generally, not a separate destination you have to click into.

What Actually Changed for Citations

This is the part that should change how you think about ranking. Tracking of citation behavior inside AI Mode showed the share of citations going to traditional top-10 organic rankers fell from roughly 76% in mid-2025 to around 38% in early 2026. That’s a genuinely large shift, and it lines up with what query fan-out is designed to do: instead of citing whichever page ranks #1 for the literal query typed, the system pulls from whichever pages best answer each of the sub-questions it generated internally, even if those pages rank nowhere near the top for the original phrase.

Zero-click behavior inside AI Mode has also been tracked at around 93% — meaning the overwhelming majority of sessions end without a click to any external site at all. That number should be read carefully: it doesn’t mean citations are worthless, since brand exposure and the composition of the answer still shape what a user believes and does next. It means the click is no longer the right way to measure whether the exposure paid off.

Metric Mid-2025 Early 2026
Top-10 organic rankers’ share of AI Mode citations ~76% ~38%
Zero-click rate inside AI Mode sessions ~93%

These figures come from third-party tracking of Google’s AI surfaces published in 2026, not from our own client data, and Google doesn’t publish this breakdown itself — treat the exact percentages as directional rather than precise, but the direction is consistent across everything we’ve reviewed.

Does This Change What You Should Optimize For?

Less than you’d think, and that’s the reassuring part. The pages that keep getting cited across the fan-out process are the ones that answer one specific sub-question clearly, in a self-contained section, without requiring the reader to piece the answer together from three paragraphs of context. That is exactly the writing discipline behind everything we’ve argued for in writing FAQ sections that rank in AI Overviews: short, direct, quotable answers structured around the question a person would actually type or ask.

What does shift is the ranking-position assumption. If your mental model has been “get to position one and the citations follow,” the merge is a good moment to update it. Our review of why AI Overviews cite some sites and ignore others found the same pattern before this merge was even announced: citation and ranking position were already only loosely correlated, and that gap is widening, not closing.

A Practical Checklist for the Merged Experience

What This Looks Like for a Small Business Specifically

Enterprise SEO teams are reacting to this merge with dedicated AI-visibility dashboards and citation-tracking tools that cost more than most SME marketing budgets. You don’t need that to respond sensibly. The single highest-leverage move for a small or mid-size site is picking your ten highest-intent pages — the ones tied most directly to a purchase decision — and rewriting their structure so each subsection answers one question completely on its own. That’s a week of focused editorial work, not a new tool subscription, and it addresses the actual mechanism behind the citation-share shift described above: fan-out rewards self-contained answers regardless of how big the site publishing them is.

It’s also worth resisting the urge to panic-rewrite everything at once. The pages most exposed to this shift are informational and comparison content near the top of the funnel, where fan-out queries are most common. Product pages, pricing pages, and other bottom-funnel content are affected far less, since those queries tend to be specific enough that Google doesn’t need to fan out sub-questions to answer them.

Is “GEO” Still the Right Framing?

We think so, with one caveat. The term has always meant optimizing for how AI systems select and present information rather than how they rank blue links, and that distinction gets more relevant, not less, as Google’s own search results absorb more AI-native behavior. Our take in Is GEO Worth It in 2026? still holds: the practices that help you get cited overlap heavily with good classic SEO, so you’re rarely choosing one over the other. What’s changing is which surface delivers the exposure — a summary box, a conversational tab, or, increasingly, one merged experience that behaves like both depending on what you ask it.

The businesses that will handle this transition well are the ones who were already writing to answer real questions clearly, rather than writing to rank for a phrase. If that’s not quite where your content is yet, the fan-out data above is a good reason to make it a priority this quarter rather than next.

— Work with Choco Media

Want posts like this working for your business?

10–40 SEO + AI-optimised blog posts a month, researched, senior-edited and published straight to your site. Built to rank on Google and get cited by ChatGPT, Claude and Gemini.

See plans — from €199/mo →
No start-up fee · Price locked for 12 months · Cancel any time after
← All storiesNext story →
— Free tips, monthly

Get the playbook, for free.

One short letter a month — the prompts we use, the campaigns that worked, the AI tools worth the time. No sales pitch, just field notes.

— Want us to do it for you?

Hire the agency.

AI-accelerated content, paid media, brand and web — delivered by one small team that talks to itself. Currently taking on a handful of clients each quarter.

Book a call