Stagwell is embedding differential privacy technology from Harvard University’s OpenDP into its new AI marketing platform, which was co-developed with Palantir. The move represents a deliberate strategy to build privacy directly into the foundation of its AI tools from the start.
The privacy moat: The platform gives marketers a single dashboard to parse tens of millions of customer records, letting them find and build new audience segments in minutes. But while competitors like WPP and Publicis are also developing their own AI tools, Stagwell is betting that making privacy a core feature will give it a key advantage in a data-sensitive market.
Adding statistical noise: The partnership embeds OpenDP’s “differential privacy” framework, a method that adds controlled statistical noise to data to obscure individual identities. Also used by the U.S. Census Bureau, the technique makes it mathematically impossible to identify a person within a dataset while still allowing for accurate aggregate analysis, according to the company's announcement.
Built-in, not bolted-on: "We believe privacy must be built into our products and solutions from the start – not something you bolt on later," said Mark Penn, Chairman and CEO of Stagwell. "That's why we've teamed up with Harvard's OpenDP to uphold rigorous protection at every step."
An early version of the platform—reportedly called the "Audience Creative and Optimization System"—is already in use with some U.S. clients, with a wider rollout of the standalone product expected in the coming months. Its adoption will be a key test of whether a privacy-first approach can become the new standard in AI-driven marketing.
Stagwell’s move is part of a broader company strategy centered on the "Three E's of AI": enablement, efficiency, and engagement. The investment in new technology comes on the heels of a strong Q3 earnings report, where the company highlighted double-digit growth in its non-advocacy work and first announced the Palantir partnership to investors.
