By Seb Joseph  •  July 18, 2025  •

Ivy Liu

These conversations tend to follow a predictable arc. Marketers first want to know if AI can actually make more content, faster and for less? Then, they’re asking how that production process mesh with data and media to deliver something that resembles personalization at scale.

“We call this the production backbone,” said Sadoun. And it’s something he said the group will try and innovate around early next year. He explained why: “This [production]  is a very big source of growth for the market in general and for us in particular if we do this right.”

He’s not alone in the calculation.

Meta has made its ambitions clear, aggressively pitching AI-generated tools to small and mid-sized advertisers. Granted, it’s less about disintermediating the holdcos and more about shorting up Meta’s long tail. But holding companies are watching closely, and eyeing similar gains among their enterprise clients. 

Sadoun pointed to “mid-single digit” creative growth in Q3, largely driven by increased demand for production and scope expansions. It’s a signal that clients are beginning to take the AI-enabled creative layer more seriously — not as a novelty but as infrastructure. In fact, more than a fifth (22%) of video ad creative was either built or enhanced using generative AI last year — a figure expected to nearly double (39%) by next year, per the IAB. 

That shift comes with a familiar pressure. As Sadoun explained: “We need to distinguish production studios using AI to deliver more content, faster, more agile, cheaper and of great quality.” 

Whether that’s possible is an open question. But the incentives are lining up. Production — long the overlooked and overromanticized part of the stack — is now seen as a lever for growth. Sadoun made the upside clear: creative agencies and production studios make up roughly 20% of Publicis’ revenue today. Get this right, and that number could climb. 

“The media landscape has been changing way faster than the creative landscape, and so agencies and clients had to adapt to this new model that we just presented, which created more opportunities,” said Sadoun. “On creative, although there’s been digitalization and socialization, we haven’t seen so much transformation versus what we have seen in media.”

The subtext: just as automation reshaped media buying — and forced CMOs to rethink their agency relationships — AI could do the same for creative.

Omnicom chief technology officer Paolo Yuvienco flicked at the point during his own briefing with analysts earlier this week. Like Publicis, the American holdco is racing to make AI a net positive by weaving it into production and planning. For instance, it uses a myriad of AI agents to simulate focus groups, tweak campaigns in real time and surface signals from shifting sales trends. 

“This isn’t just about efficiency,” said Yuvienco. “AI is allowing our creative teams to explore more unchartered creative territories, and that is really expanding the aperture of our creativity that we already believe we have an unfair share of here at Omnicom.”

The message from the holdcos is becoming clear: the old model — planning and creative silos, media scale as differentiator — is running out of road. That transition was already underway. AI just accelerated the collapse. 

“With more platforms and increased demand for audience-specific personalisation, the need for more creative assets has never been greater,” said Barney Worfolk-Smith, chief growth officer at creative data provider DAIVID. “If those assets can be produced at a fraction of the cost, the C-suite will be quick to chase the savings.”

https://digiday.com/?p=583550