Laura Higgins runs brand and innovation at Dollar Shave Club. She spent years at Procter & Gamble, which is to say she knows exactly what an agency is for. Asked by Digiday this month how much of the brand’s advertising is now made internally, she gave a number: ninety percent. Asked whether artificial intelligence might push that figure higher, she answered, “I hope so.”

Elsewhere she has been blunter still, describing generative AI as a way to bypass legacy agency models. Clients have thought this for years. What has changed is that one of them is now willing to say it into a recorder, with the receipts to back it up.

The receipts look like this. Dollar Shave Club’s Independence Day campaign, built around America’s 250th anniversary, was written by a small internal team using Claude and the AI video tool Higgsfield. The brief went to a first draft in two or three days. A finished thirty-second cut took about a week. The whole campaign, across multiple asset sizes, was live inside a month. The marketing department numbers twenty people and buys its own media. “There’s no way in the old world we would be able to go from idea to launch in a month,” Higgins said.

What she actually said

Read the quote carefully, because the useful part is the word bypass. Higgins is not predicting that agencies disappear. She says the opposite, and she still hires them, going outside when her team runs short of capacity or when an idea needs what she calls external chops. Dollar Shave Club’s earlier AI satire spot, the one in which a fictional razor executive proposes replacing everybody with AI, was made with a creative boutique called Too Short for Modeling.

The threat is narrower and more dangerous than extinction. AI is prising apart two things agencies have always sold together: the idea, and the making of the thing. Ideas were always the glamorous half. Execution was the half that paid for the building. Higgins’s own summary of where AI fits in her workflow is a single clause: the team uses it for the execution. Strip execution out of an agency’s revenue and what remains is a smaller, cleverer, far less profitable business.

The reason this is landing in 2026 rather than 2020 is capacity. In-house teams have existed for a decade, and they were always throttled by the same constraint: you can only make so much with the people you have. AI removed the throttle. That is why a twenty-person department can now outrun operations many times its size, agency or corporate. The advantage has quietly moved to whoever is structurally fastest, and small has become an asset rather than a limit.

“There’s no way in the old world we would be able to go from idea to launch in a month.”
Laura Higgins, Chief Brand and Innovation Officer, Dollar Shave Club

Answer one: get bigger

On 26 November 2025, Omnicom completed its acquisition of Interpublic Group. The combined company became the largest advertising holding company in the world by revenue, overtaking Publicis and pushing WPP into third. Omnicom’s 2025 worldwide revenue came in at $17.3 billion, up 10.1 percent.

The bill for getting there is in the filings. Severance, real-estate repositioning, contract cancellations, disposals and acquisition costs took $2.1 billion off 2025 operating income, worth $8.50 a share. More than four thousand roles went, along with several agency brands. The cost-synergy target was then doubled to $1.5 billion over thirty months, roughly a billion of which is expected to come from further job cuts, and analysts have projected up to three thousand more losses across 2026 and 2027. Chief financial officer Phil Angelastro described the logic without decoration: “you couldn’t keep two of everything.”

Source: Omnicom Group SEC filings and Q1 2026 investor presentation.

The strategy is working on its own terms. Omnicom’s adjusted EBITA margin on core operations widened to 14.8 percent in the first quarter of 2026 from 12.4 percent, driven mainly by the synergies. Organic growth came in at 3.9 percent, and the company spent $2.8 billion buying back its own shares in the same three months. Consolidation is a margin strategy, and by that measure it is delivering. The awkward question is whether a margin strategy answers a demand problem. Nothing in the merger changes what Laura Higgins does on a Tuesday.

Answer two: get cheaper

WPP took the same medicine without the acquisition. Revenue fell 5.4 percent globally across 2025 and the group has set out to strip £500 million of annualised cost from the business in 2026. Cutting is the reflex of an industry that has decided its problem is expense rather than relevance. It is also the only lever available to a company whose clients are leaving.

Answer three: get better

Publicis has spent the past decade doing the opposite of both, and it is the only one of the three that can point to a scoreboard. The group grew organically by 5.6 percent across 2025, finishing the year at 5.9 percent in the fourth quarter, and guided to between 4 and 5 percent for 2026. It has now outperformed the industry for six consecutive years and posted twenty straight quarters of outgrowing its peers, having invested around €14 billion in data and technology over the past ten years and built its AI platform, Marcel, back in 2017.

Publicis pulled in $10.4 billion of net new business media billings in 2025. WPP lost $1.8 billion. Winning the work, rather than owning the most of it, is the metric that separates them.

Chief executive Arthur Sadoun has been unsparing about the alternative. Asked about his rivals’ restructuring, he called Publicis’s approach “the polar opposite of that,” insisting that staff are the company’s greatest asset rather than a debt, and that he would rather acquire capabilities than buy back shares. On the wider noise around the merger, he was harsher: the only thing that demonstrates the strength of a model is organic growth, and “the rest is poetry.”

“The only thing that matters is organic growth. The rest is poetry.”
Arthur Sadoun, Chairman and CEO, Publicis Groupe

The scoreboard

Four strategies, one threat. Sources: company filings, earnings calls and investor presentations.

Where the squeeze actually falls

The industry’s own framing of this moment is a story about size, and size is a distraction. Publicis has argued that the number of global players capable of servicing the largest clients has fallen from six in 2020 to three today, which sounds like a case for consolidation until you notice who is growing. Omnicom bought its way to the top of the table and is growing at 3.9 percent. Publicis stayed put and is growing faster.

The pressure lands hardest on the agencies with neither advantage. A holding company can absorb the loss of execution revenue by taking cost out at group level. A boutique with a genuinely distinctive idea can survive on ideas alone, which is exactly why Dollar Shave Club still calls Too Short for Modeling. The businesses with the least room are the mid-sized shops whose value proposition was competent execution at a reasonable price. That was a real service for thirty years. It is now a feature of a software subscription.

There is a countervailing risk, and the smarter in-house teams know it. Matt Owens, chief design and innovation officer at the brand studio Athletics, put it carefully: a small team accelerating with AI needs enough talent to avoid producing the slop that audiences have already learned to recognise. Higgins agrees, and says her team will not take the AI route on every campaign. Speed is worth nothing if the work is forgettable, and the thing agencies were always genuinely good at was making the work unforgettable.

Which leaves the question the whole industry is now betting on, from four different directions. If ideas are all an agency can sell, the agency business is a smaller industry than it was. Omnicom has bet that scale can carry it through the shrinkage. WPP has bet that it can cut its way to the other side. Publicis has bet that capability is what clients will still pay for. And Laura Higgins has bet that she does not need any of them for the ninety percent, which is the only bet on this list that has already paid out.