This post was created in partnership with Flam
Is AI bringing us closer to a brand marketing world where the audience can act inside the content instead of just watching it?
During a Brandweek 2026 fireside chat co-hosted with Flam, Ryan Joe, editor-in-chief of ADWEEK, sat down with Kumar Shah, Flam’s head of sales, and Melissa Berger, chief solutions officer at Digitas. They discussed whether a new AI interactive content format could turn the one-way ad-watching experience into responsive content.
Ads powered by audiences
Shah kicked off the discussion by describing how Flam’s AI content platform, fresh off a newly announced round of Series B funding, is bringing responsive ads into the marketplace. “We want to modernize the interaction that users have with ads and make it a two-way street,” he said.
With Flicks, one of Flam’s core AI tools, brands can build interactive ads that use audience metrics and behaviors to determine what they see next. “In real time, this ad can learn about your intent across a particular product, and then deliver ads based on who you are, what you chose, and why you selected it,” explained Shah.
Digitas’ Berger wondered whether this technology could help marketers finally achieve the coveted “personalization at scale” that’s dominated discussions for the last 20 years—but made the point that it has to be done well because “you risk alienating with a bad experience.”
To answer that question, Shah explained how Flam Flicks actually works. The user uploads brand guidelines and a raw asset, and gives it a prompt. The AI then generates the creative. Within the ad experience, the user engages with multiple variants, and their choices and selections, along with the brand’s existing customer data, help determine what they will see next. Shah noted that if all goes well, a 10-second ad can be built in 30 minutes or less.
“Speed to market is a key KPI right now,” said Berger, acknowledging that teams are always looking for ways to get ads out more quickly.
Leaning into the learning curves
The discussion turned to potential fears that brands may have about giving LLMs too much control—a very common topic for AI technologies today.
“Actually ensuring the data is clean and being used properly is very difficult,” shared Berger, “so I don’t think it’s an easy problem to solve.”
