Published June 22
Roy Armale first sketched out the idea for WPP Open, the holding company’s AI operating system, on an iPad in 2022. Four years later, Armale, now DEPT’s chief product officer, is building another AI operating system — this time with a different philosophy.
DEPT is launching Deptify, which aims to seamlessly plan, create, activate and measure campaigns. And it’s reporting early success.
Three of DEPT’s top six clients have already adopted the product, as well as clients that make up around 20% of the agency’s revenue, according to Armale. By the beginning of 2027, the agency expects that number to jump to 80%.
Several major holdcos have launched their own AI-based operating systems in the past year, including WPP, Publicis, Omnicom, Dentsu, Havas, Stagwell and Horizon.
Armale argues DEPT’s advantage lies in integrating with clients’ existing tech stacks. In theory, that could make it easier to reach both sides’ ultimate motivation: growth and bottom-line sales.
“When you’re working with clients that have 10 or 20,000 people or even 100,000 people to train on this stuff, it becomes virtually impossible,” Armale told The Current. “All we’re doing is orchestrating tech from the best partners that we can find.”
DEPT isn’t alone in this philosophy. Stagwell’s agentic system, which launched earlier this year, is also centered around using advertisers’ existing infrastructure to power their strategies.
Practically, this means a brand can continue using its existing tools, whether it be Asana or Workfront, Figma or Photoshop. And an AI assistant is there every step of the way for the whole project or campaign, preserving information and context that might otherwise be lost.
Recent research supports Armale’s rationale. An IAB survey found that only 30% of brands, agencies and publishers said that they have fully integrated AI into campaign workflows.
While agencies have made a major push into this new era, pouring over a billion dollars into AI overhauls, Armale argues they don’t have the capital to truly break through compared to tech giants that are spending billions on AI innovation.
“The reality of it is no matter how much you spend, you don’t have the money to do it,” Armale said. “By trying to replicate all of that stuff, we’re just going to be throwing money in the toilet.”
He also sees a major conflict of interest in agencies selling their own tech because it can change their business models and potentially create misaligned client goals.
“If you’re selling them tech, then your profit model is around tech sales, seats and consumption,” Armale said. “And it conflicts with a growth model because if they become more efficient and grow faster, you end up losing money. And I find that that conflict between the profit motive and the purpose motive.”
Agentic AI’s place in media planning and buying
Armale’s vision also reflects a broader challenge facing media buyers as agentic AI proliferates across advertising platforms.
Within planning and buying, agentic AI is “rapidly becoming table stakes,” according to Josh Lee, EDO’s chief technical officer and head of product. Still, agentic AI has yet to reach its full potential in streamlining the advertising pipeline.
“Each technology partner is building a planning agent, but only within their own ecosystem,” Lara Koenig, vice president of global strategy and partnerships at MiQ, told The Current. “A buyer is still having to go talk to six different platforms. [That’s] not delivering on the promise of unified campaign management and unified activation, because the platforms that are building these are still quite siloed.”
A major hindrance in this evolution for media buyers is a familiar one: walled gardens.
“I think the walled gardens are some of the biggest hurdles,” Tylynn Pettrey, senior vice president of analytics and AI at Chalice AI, told The Current. “They have their place and there’s lots of media spend that’s optimized well there. But the walled gardens do what’s best for the platform.”
“They don’t necessarily have the brand’s interest in mind, so it makes it challenging to take your orchestration above the platforms and make sure you have unified measurement and unified activation.”
