Google has added agentic AI features and upper-funnel measurement capabilities to Meridian, its open-source marketing-mix-modeling platform, according to a Marketing Dive report.

The release marks the most significant expansion of Meridian since Google shipped the tool as an open-source project, broadening its scope from sales-driven attribution toward the brand-building activities that sit earlier in the customer journey.

What changed in Meridian?

The upgrade introduces two distinct additions:

• Agentic AI workflows that automate steps inside the modeling pipeline
• Upper-funnel measurement for campaigns focused on awareness and consideration rather than direct response

Meridian originally launched as a marketing-mix-modeling (MMM) framework, a statistical technique that uses aggregate data to estimate how spend across channels drives sales. The new release pushes the tool into territory where advertisers have historically relied on brand-lift studies or panels.

Why upper-funnel matters for advertisers

Lower-funnel channels such as paid search and retail media produce clean conversion signals that fit neatly into MMM equations. Television, audio, display, and influencer placements do not. Brand spend has long resisted attribution because its effects are diffuse and delayed.

Google's pitch is that the new upper-funnel capabilities close that gap inside Meridian itself, removing the need for advertisers to bolt on separate brand-lift studies. For media buyers managing portfolios that include both performance and brand budgets, the appeal is consolidation: one model, one set of assumptions, full funnel coverage.

What agentic AI adds

The agentic component targets analyst workflow. MMM work is labor-intensive: data cleaning, scenario planning, model validation, and post-hoc diagnosis can consume weeks of practitioner time. Agentic AI tooling inside Meridian is intended to automate the routine steps, leaving strategists free to interrogate the output rather than assemble the inputs.

The framing tracks a broader shift across the martech stack, where vendors from Salesforce to Adobe have spent 2024 and 2025 wrapping AI assistants around analytics products. Google is positioning Meridian as open-source in a category where competitors such as Meta's Robyn and smaller vendors sell proprietary alternatives.

Who should pay attention

Three groups have the most riding on the upgrade:

• Agencies and in-house teams running MMM regularly who want to reduce cycle time
• Brand advertisers with significant upper-funnel spend that resists attribution
• Open-source MMM practitioners already using Meridian or competitors and tracking feature parity

The open-source licensing remains a differentiator. Enterprises locked into closed measurement stacks will still see Meridian as a benchmarking option rather than a primary system, but the new features raise the floor for what an open-source MMM tool can claim to measure.

Open questions

Marketing Dive's coverage does not detail whether the upper-funnel module introduces new data inputs, new partner integrations, or new model architectures, or whether it relies on existing Meridian inputs reorganized for awareness-stage use. Adoption data, customer case studies, and pricing implications are also absent from the report.

Practitioners evaluating Meridian against incumbents such as Nielsen, IRI, or LiveRamp's measurement suite will want Google to publish technical documentation and validation benchmarks before treating the upgrade as on par with established brand-measurement products.