Inside Publicis Groupe’s closed-door Cannes AI push

The strategy has two halves. The first, intended to buzz rivals and pierce Riviera pompousness, involves a “BS Bot” app that festival-goers can use to translate AI pitches into plain English.

The second will take place away from prying eyes and ears. The company’s executives are set to hold over 30 closed-door meetings with incumbent clients at its Riviera base this week, spearheaded by Scott Hagedorn, Publicis Groupe’s global chief solutions architect.

That push is focused on CPG, automotive, financial services, pharma and luxury clients.  Hagedorn and Serrano didn’t share which advertisers had been invited, but Publicis’ current roster includes carmaker Stellantis, premium Italian food brand Barilla and pharma giant Pfizer. Publicis unveiled a bespoke generative platform build for the latter back in February.

Over eighteen months on from the ChatGPT big bang, clients aren’t just using AI as a hygiene question to sift agencies apart in pitches. One media pitch consultant, who asked to remain anonymous, said that serious discussions around AI are featuring in “every single agency meeting right now and naturally agencies are flexing their capabilities.”

“Marketer sentiment is skeptical and cautious, especially in media,” they added.

Another pitch consultant, who asked not to be named, told Digiday that although the agency group won previous pitches on the strength of its data and tech offering, rivals are beginning to catch up.

“The other agencies have come up, they have gotten it together. There’s now a group of agencies that are all selling at the same level, [with] very comparable tech solutions, and very comparable tactics and capabilities,” they said. According to a Forrester survey released this month, 60% of U.S. agencies are using generative AI in some capacity. In other words, the field is getting more crowded.

One point of difference Publicis is leaning on is its data ownership. “There’s been way too much focus on the creative output capability of AI and not necessarily building these connected data models that can power new sets of applications,” Hagedorn added.

He said the company planned to highlight a suite of 15 ready-to-market applications developed for brands within those sectors, using both generative and non-generative AI, “designed for outcomes and not outputs.”

One such solution, Hagedorn said, had been developed to use AI to connect shopper IDs and purchasing habits with live stock data drawn from major supermarket clients, enabling CPG brands to understand swings in demand. For example, if a competitor’s product was leaving shelves faster than their own, they might be able to respond with promotional activity. “Those two things have never been tied together before,” Hagedorn said. “That’s how a retailer would work. They [CPG brands] have never been able to move at that speed. But increasingly, that’s what they need to do to be competitive.”

Hagedorn is keen to highlight the depth of Publicis’ owned data, drawn from Epsilon, Sapient, CitrusAd and Profitero. “Our apps are real, because we’ve trained them on the best data that we own,” he said.

Ownership of datasets is a flashpoint for clients concerned with the potential copyright or reputational risks associated with AI. “We think clients need to be asking: if they are renting, or if their provider or partner is renting the data, are they also renting the intelligence? Because without the data, none of this stuff works,” said Hagedorn.

“Every holding group is on this accelerated drive around AI,” said Ryan Kangisser, managing partner, strategy, at MediaSense. However, Publicis has particular reason to pay attention to its incumbent brand clients. The French firm has been on a years-long tear in the pitch room, adding new clients at the expense of its peers. But organic growth depends on retained clients deciding to spend more with the company than in previous years. Furthermore, if it can persuade clients to make new investments in its AI solutions now, a divorce between client and agency will be less likely in the future.

Additionally, existing clients are a more amenable audience for AI and tech services than fresh ones, the second consultant explained. Tech service fees are often negotiated down by newer clients, they said, meaning major agencies might win an account based on AI and tech expertise without making much revenue from that area. On new accounts, they added, “most tech fees get negated.”

By contrast, older client relationships coming up for renegotiation hold more revenue potential. The consultant said: “There’s a lot more flexibility at that point to upsell. Their [Publicis’] best audience for selling all of this AI stuff is, of course, their current clients.”