Lina Hope
By Sam Bradley
June 17, 2026
For years, the advertising industry has treated "synthetic audiences"—digital twins created through artificial intelligence—as a fascinating novelty, useful for high-level strategy and focus-group simulations. However, a firm line remained drawn in the sand: media buyers were largely unwilling to trust these AI-generated personas with the high-stakes, real-world mechanics of programmatic bidding.
That barrier is now officially crumbling. WPP Media, the powerhouse media arm of the global advertising giant, has announced a strategic partnership with the technology firm Givsly, signaling a decisive move to integrate synthetic data directly into the programmatic buying workflow. By leveraging AI to process unconventional data signals—including philanthropic donation history—WPP is attempting to pivot from traditional, behavior-based targeting toward a more nuanced, values-driven approach to consumer connection.
The Core Innovation: Mapping Values to Media Spend
At its foundation, the partnership between WPP Media and Givsly represents a shift in how advertising agencies perceive "the consumer." While traditional programmatic buying relies heavily on cookies, IP addresses, and browsing history, the WPP-Givsly model utilizes a sophisticated synthesis of third-party datasets to build a more holistic picture of a person’s potential motivations.
Givsly’s engine, powered by the Claude large language model (LLM), operates by overlaying U.S. zip code data with national census figures, election trends, and, crucially, proprietary data regarding non-profit donations. By analyzing the frequency and nature of charitable giving within specific geographic regions, the platform creates a proxy for both disposable income and social values.
"We’re looking for discoverability and increased reach," explained Tara Sadlak, group director of media delivery at WPP Media. The system allows planners to move beyond the superficial tracking of "open web interactions," which Sadlak argues are no longer sufficient to build genuine brand resonance. Instead, the agency is attempting to reach audiences where their ethical drivers and real-world actions intersect.
A Chronology of the Shift Toward Synthetic Intelligence
The journey to this moment has been iterative, marked by both skepticism and gradual experimentation across the industry:
- The Early Exploration (2023–2024): Agencies like Code & Theory and Marcus Thomas began experimenting with AI personas to replace traditional focus groups. The goal was speed—generating qualitative insights in minutes that previously took weeks to gather.
- The Industry Pivot (2025): The trend gained institutional legitimacy when Dentsu entered a partnership with the startup Evidenza, using synthetic audiences to inform high-level media planning. Simultaneously, major publishers like The Times and News UK began transforming their own first-party data into synthetic cohorts for advertisers, proving that the technology could work safely within walled gardens.
- The WPP Breakthrough (2026): By mid-2026, WPP Media moved the needle further, transitioning from using synthetic data for planning to using it for active, automated bidding in Connected TV (CTV) and online video environments.
Supporting Data and Performance Metrics
The efficacy of this transition is supported by early-stage testing conducted by WPP Media. Across four separate campaigns for brands within the beauty and fashion sectors, the agency observed a 2% lift in video completion rates. While a 2% increase may appear incremental to a casual observer, in the hyper-competitive programmatic landscape, such a gain represents a significant efficiency improvement in engagement.
Givsly CEO and founder Chad Hickey notes that the speed of this technology is its primary differentiator. "We can spit out an audience in 30 seconds," Hickey says, describing a process that would have traditionally required extensive data mining and third-party audience modeling.
Forrester analyst Jay Pattisall, however, offers a grounded perspective on these results. "Synthetic audiences are only as good as the data that compiles them," Pattisall noted. He emphasizes that the success of the WPP initiative is inextricably linked to the agency’s ability to "enrich and model" these personas by layering in their own first-party client data—likely utilizing clean-room environments like InfoSum to ensure privacy compliance while maintaining data depth.
Official Responses and Strategic Rationale
The impetus for this move is rooted in the shifting nature of consumer loyalty. According to data from GWI, approximately 86% of consumers report that shared values are a primary driver when choosing between competing brands.
"Ethical drivers aren’t always present in a standard brand brief," Sadlak observed. By integrating Givsly’s data, WPP is attempting to solve a perennial problem in media buying: how to find people who don’t just click on ads, but who align with a brand’s mission.
WPP Media has confirmed that it intends to scale this solution across its entire U.S. client roster. The agency’s confidence in the system suggests that the "hallucination" risks associated with LLMs—long the primary deterrent for agencies—have been sufficiently mitigated by the structure of the Givsly platform, which tethers its outputs to verified geographic and donation data rather than allowing the AI to generate entirely speculative profiles.
The Broader Implications for Media Buying
The integration of synthetic audiences into the programmatic ecosystem carries profound implications for the future of digital advertising:
1. The Death of the "Standard" Demographic
The ability to build cohorts based on values—such as "Californian women interested in wellness" who also show specific patterns of community engagement—marks the end of the "age and gender" era of media buying. If agencies can successfully model values, demographic targeting will eventually be viewed as a blunt instrument.
2. The Rise of "Synthetic" Political Advertising
While Givsly currently excludes political donation data to maintain a non-partisan stance, the appetite for this capability is immense. As Hickey revealed, political agencies are already knocking on their door, looking for the ability to identify micro-segments of voters based on nuanced, conflicting values—such as religious attendance coupled with support for specific social policies. This capability could fundamentally alter the landscape of political campaign spending, making micro-targeting more efficient and potentially more intrusive.
3. Privacy and the "First-Party" Necessity
The success of this model underscores the necessity of first-party data. As third-party cookies disappear, agencies that possess deep, proprietary data silos—and the tools to match them—will dominate. The synthetic audience is not a replacement for real data; it is an amplification tool. Without the "anchor" of real-world client data, the synthetic model remains a theoretical construct.
4. Ethical Safeguards and Regulatory Scrutiny
As agencies move further into synthetic territory, they will face increased scrutiny regarding transparency. If an ad is served to an individual because an AI guessed they held specific values based on the donation habits of their neighbors, the industry may face new questions regarding privacy, manipulation, and the ethical use of predictive modeling.
Conclusion: The New Frontier
WPP’s partnership with Givsly represents a shift from "guessing" to "modeling." By treating synthetic audiences as a legitimate component of the programmatic stack, WPP is signaling that the era of being "leery" of AI is over.
The industry is moving toward a future where media buying is less about tracking the user’s history and more about predicting the user’s future alignment with a brand’s values. As this technology matures, the definition of an "audience" will likely become more fluid, more dynamic, and increasingly indistinguishable from the real humans they are intended to mirror. The question moving forward is not whether synthetic audiences work, but how far the industry is willing to push the boundaries of this digital reflection before the consumer feels the weight of the invisible hand.
Rifan Muazin
Siti Muinah
