UATOPIA Panel: Running a UA Team When Agents Do the Tasks

Adopting AI for UA teams changes the work of managers, creative specialists and the people who build their tools. Samantha Benjamin of Supersonic, Zeynep Tasoz Cataner of Zynga and Thibault Quillard of Madbox join Amit Carmi, CEO of Sett, to discuss the operating decisions behind that change.

Samantha describes how Supersonic uses automation for parts of campaign optimization, with people retaining oversight and responsibility for risk. Her examples separate maintaining a more established campaign from making the strategic decisions that come with a launch, when the data is still developing.

Zeynep discusses aligning creative and UA teams around shared performance goals and helping people adopt new ways of working. Thibault brings the hiring perspective: curiosity, hands-on experimentation and the ability to turn an idea into a useful first version. The discussion keeps returning to what a team learns from the work it produces.

The later questions move from individual tool use to systems a whole department can rely on. The speakers describe collaboration with R&D, the importance of accurate data and the process of taking a working prototype into production. They also discuss how managers can stay close enough to the tools to guide their teams and choose what to build next.

What you’ll take away

• Decide which operational tasks can be automated and where people retain strategic ownership.
• Align creative work with the performance goals of the UA team.
• Hire and develop people who combine curiosity with the ability to build and test.
• Plan for the infrastructure, data quality and engineering work behind a shared AI workflow.

Jump to the discussion

• 00:37 · Which decisions should AI own?
• 05:06 · Giving creative and UA teams shared goals
• 09:20 · The skills and attitudes the speakers hire for
• 20:57 · Getting from an idea to a first creative draft
• 34:22 · Supporting builders inside the team
• 35:09 · Turning prototypes into production systems
• 40:12 · How AI changes the work of managers