Google Marketing Live 2026: sorting real features from announcement effects

A good media buyer does not get excited about a product announcement: they sort what they hear. Available now, due in six months, or laboratory concept.

1. Decoding the roadmap from Mountain View

Every year, Google Marketing Live plays the same score: a polished keynote, marketing badges that change — this edition under the “Gemini Advantage” banner — and an implicit promise that digital acquisition will be fundamentally different within eighteen months. Most advertisers leave with their eyes full of slides and a hazy backlog of tests.

At Bright the stance is different. A good media buyer does not get excited about a product announcement. They sort what they hear by a simple grid: what is available and testable now, what will be within six months, and what belongs in the lab? That filter avoids two symmetrical mistakes: ignoring what really changes, and spending operational time on vapourware. It is that same sorting that guides our recommendations in our clients’ media plans.

Here is our technical reading of GML 2026, theme by theme.

2. AI & Search: AI Max extends, AI Brief arrives

Coming soon — AI Max extends to Shopping and Travel

AI Max, until now confined to standard Search campaigns, extends to Shopping and Travel campaigns. That is concrete, and it is the logical next step in the gradual automation of Search. For e-commerce advertisers who already work their product feeds rigorously, extending broad query matching to Shopping inventories is a change to anticipate from Q3. The risk of off-target delivery is the same as in Search: without quality assets and solid audience signals, the algorithm optimises into a void.

“…AI Max for Search and Performance Max, letting Google identify opportunities that traditional keyword-only strategies might have missed.”

The same logic of finding opportunities beyond keywords now applies to Shopping — with the same demands on asset and product feed quality. Google showed us the future of Search

Coming soon — “AI Brief”: the semantic guardrail, at last concrete

This is the real good surprise of GML. AI Brief is a free text field built directly into the campaign interface, letting the advertiser instruct the AI “in their own words”: brand positioning, tone, excluded audiences, semantic territories to avoid. It is not a technological revolution, but it is a real step towards algorithmic transparency. Until now, guiding Google’s AI meant playing with indirect levers — assets, audience signals, keyword exclusions. Being able to impose semantic guardrails in natural language changes the granularity of control. To test as a priority on accounts where brand safety is at stake.

Teased / futuristic — “Ads in AI Mode”

The idea: inserting formatted ads within Gemini’s conversational answers, in Google Search’s AI Mode interface. The concept is structurally interesting for the future of generative Search. The reality today: the volume of qualified commercial intent in those conversational flows remains marginal. The formats are not standardised, the performance metrics unknown. To watch over 12 to 18 months, not to budget for now.

3. Demand Gen: a wider network, not a format revolution

A reminder / nothing new

Demand Gen extending to Google Maps, and the stronger integration of product feeds and creator assets (UGC, influencer content), are presented as novelties. In reality, this is an extension of the delivery surface for a format that has existed since 2023. Demand Gen remains what it was: a social-like prospecting campaign, mixing YouTube, Discover and Gmail in a logic of reach and top-of-funnel engagement.

Bright’s view: Google continues to praise AI’s ability to “predict” intent before the click. That is a sales argument, not a guarantee of direct performance. Mixing static formats and YouTube video within a single conversion-driven, ROI-minded campaign remains a risky bet. YouTube is an excellent lever for recognition and consideration. Asking a Demand Gen campaign to rival Search or Shopping on immediate cost per acquisition is to misread the format’s place in the funnel. Use it for what it does well.

4. Commerce: gentle disintermediation

A reminder / technical — Universal Commerce Protocol

The Universal Commerce Protocol (UCP) is the new interoperability protocol Google is pushing, to standardise and unify purchase flows between merchants and platforms. The ambition is structuring on paper: less friction in exchanging product data, greater coherence across Google’s surfaces. In practice it is an infrastructure change — useful, but it does not change the business equation in the short term.

Teased / futuristic — Native Checkout in Gemini

“Native Checkout” goes further: the user could complete a purchase directly in the Gemini chat interface, without ever visiting the merchant’s site. The feature is on a waitlist, with no rollout date.

Bright’s business analysis: the technical feat is real — removing the friction between intent and purchase mechanically lifts the conversion rate. But for DTC brands the calculation is more complex. Accepting Native Checkout means arbitrating between immediate sales volume and sovereignty over the customer relationship. You lose direct contact on the site: no native cross-sell, no first-party data collection in your own environment, no retargeting on a session you own. It is the same dilemma as Amazon Prime these past ten years, in a generative AI version. The decision belongs to each brand, by its data maturity and its margins.

5. Measurement: technical maintenance, data integration, and the MMM illusion

A reminder / technical — Google Tag Gateway

Migrating to the Google Tag Gateway is a technical necessity in the face of third-party cookies being gradually deprecated — not an option nor a novelty. Accounts that have not yet implemented it are already losing signal quality. That is maintenance, not innovation.

“you have a hand on the data you send to providers, since it goes to your server first.”

That is exactly the argument justifying the migration to the Tag Gateway today: taking back control of the data flow before it leaves for third parties. What server-side GTM brings

Coming soon — Data Manager in Google Analytics (GA360)

Historically housed in Google Ads, Data Manager moves into Google Analytics 360 to become a hub for ingesting cross-channel advertising signals — Meta, TikTok, CRM data. That is a structuring move for multi-platform advertisers who want to centralise their view of conversions without going through a third-party CDP. Access remains conditional on GA360, so it is reserved to accounts with a considerable analytics investment. For those who have it, implementation is a short-term priority.

“putting a marketing data warehouse in place opens the way to a finer, more objective measurement of performance.”

Data Manager is never more than one further brick in that architecture — centralising the signals does not replace the groundwork on sources and pipelines. Building a marketing data warehouse

Teased / large accounts — Meridian built into GA360

Google announces that Meridian — its open-source Marketing Mix Modeling model — will be built directly into GA360. The announcement is as political as it is technical: it is a direct answer to Robyn, Meta’s MMM. Google pairs it with the launch of Meridian Studio, a graphical interface meant to make MMM more accessible and to reduce the dependence on raw Python. The move is real and worth noting: lowering the technical barrier to entry was the main brake on adoption.

But the illusion stops at the interface. Having ourselves orchestrated the deployment of bespoke MMM projects for some of our clients at Bright, we know the tool does not make the model: a Bayesian framework always needs immaculate historical data, or the figures can be made to say anything. Studio becomes more accessible, but the rigour the data demands stays the same. For 90% of advertisers, the prerequisites — clean data over two to three years, granular spend channel by channel, a team able to interpret the outputs — remain the real bottleneck. For the others it is a serious avenue to assess, not a deployment to rush.

6. The informed advertiser’s roadmap

GML 2026 is neither a revolution to adopt wholesale nor a catalogue of vapourware to ignore. The useful signal is there — it just must not be drowned in the noise.

What you do now: Lock your structures around AI Max (Search, then Shopping) and deploy AI Brief on your brand-safety-sensitive campaigns. These are levers that are available, documented, and of immediate value to traffic quality.

What you structure over three to six months: Invest in Data Manager if you are on GA360 and operate across several paid channels. Centralising the conversion signals is the groundwork on which every future optimisation depends.

What you watch from a distance: Native Checkout, Ads in AI Mode and Meridian are potentially disruptive technologies. They deserve active watching, not a rushed deployment. The first brand to jump without a net pays for everyone else’s learning.

Maturity in acquisition is not about being first to test. It is about knowing exactly why you are testing.