Google is bringing agentic AI and upper-funnel measurement capabilities to Meridian, its open-source marketing mix model, in an update that also extends its GeoX incrementality tool globally. The changes push Meridian beyond reactive reporting toward a more dynamic, real-time model-building assistant - a shift that matters for marketers who need faster, more accurate cross-channel measurement without ceding control to black-box systems.
Agentic AI steps into model building
The new agentic AI features aim to help marketers audit data quality, resolve errors, and guide model construction in real time. Rather than simply producing a report, Meridian can now act as an interactive assistant during the modeling process. Google said the update also includes back-end improvements to speed and efficiency, described in a company blog post as "significant updates."
These changes extend Google's broader push to inject agentic capabilities across its advertising products. For measurement specifically, the goal is to turn interfaces from what the company calls a "reactive report card" into something more dynamic. Marketers working with AI for Marketing tools will recognize this pattern - assistants that don't just surface insights but actively help shape the analysis.
Upper-funnel channels get direct measurement
Meridian can now incorporate more brand-building signals directly into its models. This gives marketers a clearer view of upper-funnel channels like TV and out-of-home, which are traditionally harder to measure than digital performance media. The update means a TV campaign's contribution can sit alongside search and social data within the same model, rather than being estimated separately or ignored.
Google is also making Meridian GeoX, its publisher-agnostic incrementality solution for geographic experiments, available globally after a beta period. The company said incorporating GeoX insights can improve the overall accuracy of MMM results.
Broader measurement suite changes
Beyond Meridian, Google is integrating Data Manager directly into Google Analytics and Display & Video 360. Both platforms also received an enhanced conversions feature designed to improve ad relevance while securing matched customer data. A new Data Strength Uplift Metric in Google Ads will calculate the conversions recovered by a marketer's first-party data setup.
These updates arrive as interest in marketing mix modeling remains high, even after Google abandoned its plan to deprecate third-party cookies. Meridian was introduced two years ago as a privacy-minded alternative during the height of the cookie debate. Brands including Alaska Airlines have reported finding new revenue opportunities through the tool, though concerns about Google grading its own homework persist.
Why this matters for marketers
Marketers get two practical gains here: faster model iteration through agentic assistance, and the ability to measure upper-funnel channels that have long resisted clean attribution. If you're running TV or OOH alongside digital, Meridian now absorbs those signals directly rather than treating them as external assumptions. The GeoX global rollout also means geographic incrementality testing is available without platform lock-in. For teams already using MMMs, the agentic features could cut the time spent on data cleaning and model troubleshooting - work that typically falls to analytics teams before any strategic insight emerges.
Your membership also unlocks:
