Set the product names aside and this year’s Google Marketing Live made a single argument: the results page is being rebuilt from the ground up around Gemini, and everything downstream — ad formats, feeds, measurement — has to move with it.
The behaviour it’s catching up to is already here. Most people treat Google like an assistant now: they ask a whole question and expect a composed answer, not a page of links to pick through. The announcements mostly just make that the default rather than the exception.
For brands, the real story isn’t paid versus organic — it’s that both now feed the same engine. Gemini assembles its answers from whatever it can read: your product feed, your editorial, your ad creative. Treat those as one connected source of truth and you’re the brand it surfaces; leave them fragmented and you’re invisible. Quieter, but just as consequential, AI Mode is pulling record numbers of people into a signed-in Google — finally handing it the kind of first-party audience graph Meta has traded on for a decade.
The announcements that actually move budget
- Ads built for the conversation. A new class of formats made for AI Mode — each assembled in real time around the query so it reads as part of the answer rather than an interruption. The native, checkout-style Direct Offers placement is widening its rollout.
- Gemini 3.5 Flash as the default. By Google’s own numbers AI Mode passed a billion monthly users inside its first year, with query volume climbing sharply each quarter — and the search box got its biggest rethink in 25 years: multimodal, intent-aware, and quick to guess where a query is heading.
- Personal Intelligence, worldwide. Connecting Gmail, Photos, and Calendar to AI Mode is now live across close to 200 countries. More signed-in users means more context for Google — and a far larger addressable audience living inside search itself.
- Ask Advisor. A Gemini agent that reaches across Google Ads, Analytics, Merchant Center, and the Marketing Platform, stitching together signals that used to sit in separate tools.
- Meridian, inside Analytics 360. Google’s marketing-mix model now ships within the measurement stack, next to a new long-term-conversions metric. The message is blunt: incrementality is the lens now, and platform-reported ROAS is no longer the headline number.
- Asset Studio and Demand Gen. Gemini-made images, video, and copy flow straight into Asset Studio, while Demand Gen leans harder into creators and YouTube to fold discovery-style buying into performance media.
- A commerce protocol for agents. AP2, the Universal Commerce Protocol, and the new Universal Cart are spreading into more categories — with richer product attributes written specifically for AI agents to parse.
What this actually asks of your brand
In paid search the centre of gravity is moving toward intent modelling, autonomous bidding, and incrementality. Five places where the work pays for itself:
- Own your first-party data. With more people signed in, your own signals — CRM lists, customer match, on-site behaviour — become the single most valuable targeting input you have.
- Treat the feed as infrastructure. Structured product data is the language agents read when they compose an answer. Merchant Center has quietly turned into a strategic asset rather than a back-office chore.
- Move faster on creative. AI has removed the bottleneck on producing creative; the edge now belongs to whoever can test and learn quickest inside a coherent, recognisable brand system.
- Grow up on measurement. Incrementality tests and mix modelling now outrank last-click. Teams still defending pure in-platform ROAS will keep losing the room in budget conversations.
- Put guardrails on the AI. Prompting standards, brand-safety rules, and human oversight of automation are becoming genuine operating disciplines — not slideware.
How we’re responding across Zuti accounts
None of this is theoretical for us. Here’s what’s already changing in the day-to-day work:
- Auditing every client feed against the new conversational attribute standards, so products turn up in real questions — not only exact keyword matches.
- Bringing paid and organic content into line so each reinforces the same answers Gemini is being asked to give.
- Rebuilding reporting around incrementality — preparing Meridian-ready inputs and retiring siloed platform ROAS as the headline metric.
- Adding AI-visibility measures — mention rate and share of voice across AI surfaces — to the standard SEO scorecard.
- Rewriting content for extractability, favouring formats an AI will happily cite over pages that merely rank.