The agency holdcos have an AI story, but not an AI business model

It’s a tidy frame — a way to signal that generative tools won’t hollow out the model and EBITA can be protected.

Investors, however, aren’t buying it. 

WPP’s new strategy landed without a share price bump. Publicis’ continued growth in both revenue and profitability wasn’t enough to stop its own shares from falling. Dentsu’s sluggish results hit its market value in kind. Havas held steadier. Omnicom bounced after its results but on the back of merger mechanics and capital returns rather than any renewed confidence in the direction of travel. 

Strip away all the noise and the market’s message is consistent. The AI story, as currently told, isn’t landing. Some of that is execution. But mostly it’s something more damning: when every holdco is reaching for the same vocabulary, the same AI layers, the same agentic orchestration and the same human-supervised automation, the argument stops sounding like a strategy and starts sounding like a script.

A few soundbites from the earnings calls illustrate the point:

Publicis CEO Arthur Sadoun told investors AI is “not a headwind — for us, it is actually a strategic driver of growth and margin expansion.” 

His counterpart at Omnicom, John Wren, reached for the same register, describing the group as “the world’s learning marketing and sales company, built for intelligent growth in the next era.”

Havas boss Yannick Bolloré said his business had evolved as “an AI-driven organization fueled by human ingenuity, where technology amplifies human creativity rather than replacing it.” 

Dentsu Americas chairman Giulio Malegori said the group would “accelerate investment in the media strategy, while embedding an AI-enabled workforce with stronger data integration across planning, activation and analytics.” 

Cindy Rose, the CEO of WPP — arguably the bellwether of the holdcos — closed the loop. She said the group’s “new mission” is to be a trusted partner for the world’s leading brands in the age of AI. 

Five execs on their respective earnings calls this week each claimed that AI is an asset not a threat. Investors, reading all five simultaneously, are entitled to ask which one is right. Or whether any of them are. 

The business model problem

The holdco platforms need to deliver on the promises made — and so far, clients aren’t seeing it.

“The honeymoon around AI and agencies is essentially over,” said Gartner vp analyst Jay Wilson. “A year ago, agencies were selling clients on their AI platforms and saying this is going to create cost efficiencies and speed to market. CMOs said, ‘is it going to cost less?’ And agencies were saying, ‘well, not quite yet’. And now, as I talk to CMOs who are reviewing and renewing agency scopes of work, they’re still not seeing the cost savings in terms of fee.”

The problem isn’t only delivery. It’s that the business model needed to make good on those promises doesn’t exist yet. 

When AI automates creative production and removes the human from ad buying, the old time-and-materials model collapses and nothing obvious replaces it. Pay per asset breaks down when assets are generated at scale. Retainers feel arbitrary when output is no longer tied to hours. Subscription fees could struggle to capture the value of work that compounds over time. Outcome-based pricing is where the industry points when pressed – but it requires client data access and board-level trust that most agencies haven’t earned and can’t manufacture.  

As Bruce Biegel, senior managing partner at Winterberry Group put it: “The thing that comes up consistently when we talk to these companies is ‘how do they price whatever comes next?’”

Nobody knows. 

But whatever model wins, creative will be at the center of it — not as a production cost to be automated away but as the judgement layer that determines whether any of it works. What should exist. What a brand should stand for. Whether the work actually moves people. Those are not efficiency questions. They are value ones, and the industry has no agreed way of pricing them yet. 

That is precisely what makes the pricing question so consequential. The received wisdom is that creative eventually migrates into the media budget, becoming a percentage of spend rather than a line on a separate invoice. If it does, it would amount to more than a billing adjustment — a reordering of what agencies are for and where their leverage actually lies. 

Until the industry resolves this, the client frustration compounds, the business model stays broken and investors stay skeptical.

“Almost all the focus and all the noise [around AI] is about efficiency — and maybe a bit about effectiveness — but there’s very little about real value creation,” said Paul Frampton-Calero, CEO of independent media agency Goodway Group and the former boss of Havas Media Group U.K. “AI will give you faster and cheaper. But faster and cheaper doesn’t mean any better. You’ll just get there quicker with fewer humans in the loop.”

A reality check

It is a point that lands differently when it comes from someone, who has sat inside the system. And it is one that, in quieter moments, the holdco chiefs would likely not dispute. 

In fact, Wren has said publicly that everyone will end up with the same tools and that what matters is the talent behind them, which reading between the lines, is an admission that the platform race is already over as a differentiator. The architecture underneath those platforms bears that out. 

WPP’s new turnaround plan is the latest acknowledgement of that reality so far and, in its own way, the most revealing shift in this whole earnings season.

For years, WPP operated as a federation of brands held together by a shared balance sheet and not much else. What was announced last week was an attempt to move away from that. The group is collapsing into four operating units, media at the center with Open recast as connective tissue rather than a product in its own right. Most companies in trouble dress the diagnosis up. WPP, to its credit, named it plainly: too complex, inconsistently executed and carrying a balance sheet that needed fixing before anything else could happen. 

But knowing the diagnosis and acting on it are two different things. The space to move decisively – to absorb short-term pain, restructure operating models and rebuild client trust simultaneously – is precisely what running a public company inside a quarterly reporting cycle tends to suppress. And in the meantime, clients are drawing their own conclusions. So the script stays tidy, the narrative stays optimistic and the harder questions get deferred to a future that remains, for now, conveniently unpriced.

Ebiquity CEO Ruben Schreurs, whose firm advises major advertisers on media and agency performance, put a finer point on it:  “In my humble opinion, the key thing for the giants to achieve is clear differentiation and a fitting value proposition that serves the interests of their key clients. Ironically, for companies in the business of brand building and marketing, their own marketing and brand positioning has not been as strong as it could have been.”