Adweek published a report from the ADWEEK House Possible Group Chat, a session co-hosted with Mutinex that brought together marketing teams from Hershey, Breeze Airways, and PubMatic to compare notes on agentic AI in practice.
The headline finding
One standout example: media-mix modeling that used to take months of data preparation now takes hours. That kind of compression is not a pitch deck claim. That is what panelists reported from live deployments.
What the panelists actually said
The session focused on agentic AI systems designed to automate workflows, support measurement, and connect large datasets to creative and strategic tasks. Three themes came up repeatedly:
- Data infrastructure first. Panelists were consistent that AI tools layered onto messy data produce messy outputs. Cleansing and standardization come before any agent deployment.
- Repetitive tasks out, strategic work in. AI-assisted campaign workflows are being used to remove manual execution steps, freeing teams for higher-level planning and creative judgment.
- Human oversight stays. The session reinforced that cultural understanding and emotional intelligence remain guardrails. No one in the room was describing fully autonomous systems running without review.
The operator takeaway
If you run marketing for a small team or as a solo operator, the same logic applies at smaller scale. Before you bolt an AI agent onto your reporting or campaign workflow, sort out whether your underlying data is clean enough to trust. If it isn’t, the agent just automates bad inputs faster.
The Mutinex partnership was cited as one example of how brands are integrating AI into analytics and workflow management. No pricing details were shared in the report.
