SEO
Google didn't make AI search the default. It did something quieter
At I/O 2026, the headline traveled fast and survived scrutiny badly. Here's what actually changed on 19 May — and why the real shift is both smaller than the panic and more consequential than the relief.
Google didn't make AI search the default. It did something quieter — and more permanent.
At I/O 2026, the headline traveled fast and survived scrutiny badly. Here's what actually changed on 19 May — and why the real shift is both smaller than the panic and more consequential than the relief.
The default that actually changed
The word "default" is doing two very different jobs in the coverage, and conflating them is where practitioners go wrong.
What did NOT change
The default experience — what loads when you type a query — has not flipped to AI Mode. The ten blue links did not die on 19 May.
AI Mode as the default experience waits on one thing: monetisation parity. The experience flips when AI Mode generates ad revenue comparable to what organic results produce today — not before.
What DID change
The default model — the intelligence sitting behind AI Mode for a billion-plus monthly users — has changed. Gemini 3.5 Flash is now the default model powering AI Mode, globally.
Google also announced the biggest rebuild of the search box in over 25 years — reasoning across text, images, files, video, and open Chrome tabs.
That's a commercial gate, not a technical one. The direction of travel is settled. Only the pacing is open.
The classic results page survives — but increasingly as the place you land before the conversation pulls you somewhere you never have to leave. Follow-up questions now live inside the results surface. A user moves from query to research to intent without a referral click leaving Google.
Why "it's not the default yet" is the wrong thing to be relieved about
If your read on I/O 2026 is "we have time," consider what the data was already saying before the announcement.
1B+
Monthly AI Mode Users
Up from roughly 100 million in late 2025 — a 10x jump in six months.
2x
Query Growth Per Quarter
Queries have more than doubled every quarter since AI Mode launched.
6mo
Time to Scale
The behaviour change is not waiting for Google to make it official. Users already are.
The question was never "when does AI Mode become the default?" It's "what percentage of my category's queries are already resolved without a click, and is that share rising?" For most of us, it is — and it was rising before May.
Two consequences that should reorganise how you work
Consequence 1: Ranking #1 no longer means appearing in the answer
The retrieval layer that assembles an AI response does not simply read off the organic SERP. You can hold position one and be absent from the synthesis a user actually reads.
Rank tracking as a primary KPI is measuring a surface that fewer of your buyers see. The ranked list and the cited answer are now two different things.
Consequence 2: The measurement model is structurally broken
Background information agents and agentic checkout — both confirmed at I/O — create outcomes with no session, no UTM, and no trackable click path.
You cannot fix attribution for a journey that never touches your analytics. You can only build a different way of knowing whether you're present where decisions get made.
You cannot fix attribution for a journey that never generates a click. The correct response is to build a different way of measuring presence entirely.
What practitioners should actually do this quarter
Stop optimising exclusively for the ranked list and start optimising for retrieval and citation. The unit of visibility is shifting — and the inputs are different.
We've been running live programmes for clients since mid-2023, arguing that the unit of visibility was moving from the ranked list to the answer — a year before I/O 2026 made it the obvious read. The announcement didn't change our position. It caught up to it.
Optimise for retrieval and citation
Optimise for retrieval and citation
The unit of visibility is shifting from "where do I rank for this keyword" to "am I cited when a model answers this question, and across which related questions." Those are different problems with different inputs.
Move monitoring from prompt-level to topic-level
Watching a handful of prompts tells you almost nothing durable — phrasings churn, and a model's answer to one wording rarely predicts the next. Track visibility at the level of a topic, across the cluster of questions a buyer actually asks. That's the only signal stable enough to act on.
Treat structured data as table stakes
Clean, machine-readable content is no longer a nice-to-have. The retrieval layer rewards content it can parse, attribute, and trust. Ambiguity that a human reader forgives, a model silently drops
Accept that some conversions happen where you cannot see them
Measure presence — citation share, answer inclusion, topic-level visibility — as a leading indicator. Treat it as the closest honest proxy you have for influence over decisions you'll never get a UTM for.
The actual headline
Google didn't make AI search the default on 19 May. It made the default model intelligent enough that the experience flip is now a matter of accounting, rebuilt the front door to assume conversation, and let the user behaviour run ahead of the official switch.
The change is already underway
User behaviour has moved. A billion-plus monthly users are already learning to ask, follow up, and decide inside the answer — with or without an official default flip.
The official default is a lagging formality
When monetisation parity arrives, the switch flips. That's a question of when, not whether. Planning around the announcement date is planning around the wrong event.
Winners stopped waiting
The practitioners who win are the ones who built monitoring for a world that runs on answers, not links — before that world was confirmed from a stage. We stopped waiting a year ago.
The quiet version of the announcement is the one worth planning around.
We've been helping clients navigate the changing world of SEO & AI Search
This is what our agency was built for - to be at the top of the game by anticipating, researching and implementing strategies to ensure our clients retain SEO visibility and embrace the new marketing opportunities of AI search.