What you bring to Google's AI is increasingly what determines performance. That was the consistent message across Google Marketing Live 2026 through announcements on search, measurement, and commerce alike. First-party data, clean briefs, and measurement infrastructure that connects media activity to real business outcomes are the foundation the platform is being built around.
Let’s take a look at the announcements that we were most excited about: measurement, AI Max, YouTube as a performance channel, and agentic commerce.
The biggest structural announcement of the keynote was the integration of Meridian directly into Google Analytics 360. Google is placing marketing mix modeling inside the same interface that houses campaign reporting, audience data, and attribution analysis, making MMM a true practitioner tool.
The integration is in the paid Google Analytics 360 tier, with support for first-party and cross-channel data and a built-in budget optimizer for scenario planning. What's available inside Google Analytics is structured reporting and light visualizations, which is a meaningful step up for teams that haven't had MMM visibility inside their reporting environment, but should not be considered a replacement for a full Meridian build. Teams with more complex modeling needs, custom data inputs, or deeper scenario planning requirements will still get substantially more from the standalone offering. That said, even at that scope, having model outputs and campaign data in the same place can change how budget conversations happen.
We've spent the last year building and deploying Meridian for clients across industries, and what we've seen consistently is that having the model is rarely the bottleneck. The harder work is organizational: getting clean, unified data in, and building the internal alignment to act on what surfaces. That part doesn't change with the Google Analytics 360 integration. What changes is how much friction stands between the model and a budget decision. We cover this more in-depth in our blog around the actionability gap for MMM.
Google also introduced Qualified Future Conversions, a Gemini-powered metric that connects current ad engagement to predicted future purchasing behavior using signals like branded search activity. QFCs are in restricted pilot with broader beta access expected later this year. Google has indicated they'll eventually feed back into Meridian, giving the model more timely demand signals rather than relying solely on historical data.
AI Max moved out of beta at GML, which is a milestone, but the more consequential announcement for day-to-day campaign management was AI Brief. It lets advertisers describe their brand in natural language — what to say, what to avoid, which audiences to prioritize, how the tone should feel — and Google uses that as context to shape how AI Max generates copy and matches queries.
AI Mode search queries are running three times longer than traditional searches, which means ad copy written for short keywords will increasingly underperform in this environment. The advertisers best positioned for that shift are the ones who can give the AI a high-quality brief rather than a tightly controlled keyword list.
For teams already familiar with AI Max, if you haven't already audited your negative keyword lists, URL exclusions, and brand guidelines, those are the inputs that determine how well AI Brief can work. The existing guardrails in AI Max — brand exclusions, URL exclusions, the ability to opt out of Text Customization — still apply, and they remain the levers advertisers control directly.
We've written previously about getting AI Max set up and tested well, and the testing framework there holds. What's new is that the brief itself is now the starting point, and the quality of what you put into it matters more than it did before.
Google also introduced Campaign Type Attribution at GML, giving advertisers a cleaner way to understand YouTube's contribution within a broader media mix rather than evaluating it against other channels using a single attribution model. The announcement also included a measurement partnership with TransUnion that brings multi-touch attribution data into the picture for YouTube campaigns.
For teams running Demand Gen, Campaign Type Attribution changes what's possible in the measurement conversation. YouTube can now be evaluated on terms that reflect how video actually influences purchase behavior, rather than being squeezed into a last-click or linear framework where it was always going to look underpowered.
Demand Gen also expanded to Google Maps at GML, which extends the placement surface and introduces a new context for performance measurement. The Maps extension is still early, but it adds another channel to account for in cross-channel modeling.
The agentic commerce announcements — Universal Commerce Protocol (UCP) expanded to new verticals, the Universal Cart rolling out across Search, Gemini, and Maps — represent infrastructure being put in place rather than a feature available today.
UCP-powered checkout is live for some retailers and rolling out to more markets this summer, but it's a good time to start preparing. What determines whether a brand shows up when an agent is doing the browsing is product feed quality: Merchant Center completeness, accurate product data, and visibility across AI-powered shopping surfaces.
For more on agentic commerce, Adswerve’s Kaysen Jacobelli has been tracking this shift since before GML, and she and our partners at Lamark have the more detailed read on what it means for retail brands.
Google Marketing Live tends to produce long to-do lists. The announcements this year are all pointing at the same underlying requirement. The value you get from AI Max, from Meridian inside Google Analytics, from agentic commerce, all runs through the same foundation: quality first-party data, clean measurement infrastructure, and the internal alignment to act on what your models surface.
If you're trying to figure out where to start or want to pressure-test your current setup against where the platform is heading, we'd love to talk.