“Every visit, every session, and every moment of attention creates a signal,” Reed said. “When we connect those signals, they become actionable intelligence that allows us to engage users more effectively and monetize those relationships faster and at much greater value.”
That need reflects a broader shift in the industry, according to David Clinch, co-founder of Media Growth Partners.
“Media needs new forms of revenue, but also needs better forms of revenue,” he said. “This kind of deal could be a better version of legacy data brokering, or it could offer a much greater scale of data ingestion than the media industry has accessed before.”
Speed, with engineers in the room
Palantir enables media and marketing companies to move faster than they could on their own, according to all five executives.
Stagwell had a working version of its agentic marketing system within 16 weeks, built during a trial that began on a handshake between Penn and Palantir CEO Alex Karp. Zeta rebuilt its entire data cloud on Foundry, Palantir’s commercial platform, after a test so convincing that Zeta skipped a competitive review, according to Steinberg.
“We tested its product and found it superior,” he said. “It was so good that we didn’t even go out and run a big bake-off.”
USA Today’s Reed made a similar case to investors.
“What may take us a considerable amount of time to build internally, we can do in a matter of weeks or months with Palantir,” he said.
Much of that speed comes from what Palantir calls its forward-deployed engineers, who embed with clients and build alongside their staff. All four firms used forward-deployed engineers. At Fox News Digital, those engineers shadowed the newsroom before building anything, according to Fox News Digital president and editor-in-chief Porter Berry.
“The forward-deployed engineer sits with the reporter who writes the piece and sees the editor who chooses the photo,” he said. “After observing that process, they then go and build digital versions of workflows.”
The resulting platform, called Newsroom, handles the dozen or so tasks that follow the writing of an article, such as suggesting video, links, photos, and tags. It now assists in the production of 95% of Fox News Digital’s articles, although nothing reader-facing is generated by AI.
“They didn’t just come in and advise,” Berry said. “They came in and built something with us that will endure over time.”
A layer above the models
Another key appeal of Palantir is its interoperability, which differentiates it from many of the other AI firms that have models of their own, according to Kahan.
At Stagwell, Foundry sits between its data and whichever AI model works best for a given task, letting it work across Azure, Google, and other clouds.
“We want to be agnostic to the LLM,” Kahan said. “That is what the ontology gives us: the ability to take that data, integrate it, and reason over it. If someone comes out with a new model, we switch the model.”
Similarly, Fox News Digital uses ChatGPT, Claude, Grok, Gemini, and Cursor, but those tools serve a different purpose than Palantir, according to Berry.
“What makes Palantir different from [other AI products], which are individual tools, is that we are able to industrialize our operation, the way we do our work,” he said.
