The Great Unbundling: Why Ad Agencies Are Abandoning Data Ownership for an AI-First Future

For over a decade, the advertising industry’s dominant narrative was one of consolidation. Agency holding companies spent billions in a frantic "arms race" to acquire data-tech stacks, DSPs, and identity providers. The goal was simple: build a proprietary ecosystem where the agency owned the pipes, the data, and the intelligence.

However, at this year’s Programmatic IO in New York City, that narrative took a sharp, almost contrarian turn. A panel of industry heavyweights—including Sir Martin Sorrell (S4 Capital), Bob Lord (Horizon Media), Maggie Summers (dentsu), and Shane McAndrew (WPP)—signaled that the era of the "all-encompassing agency stack" is ending. In its place, agencies are pivoting toward a model defined by neutrality, client data sovereignty, and the transition from "managed services" to "software-as-a-service" (SaaS) providers.

The Chronology of the Shift: From Asset Acquisition to Agentic Intelligence

To understand where the industry is going, one must look at where it has been. The 2010s were defined by the pursuit of proprietary data. The high-water mark of this strategy arrived in 2019, when Publicis Groupe acquired Epsilon for $4.4 billion, a move designed to unify identity and data across the entire holding company. At the time, analysts like Terence Kawaja hailed this as a "win" for the agency model, suggesting that owning the "truth" was the only way to remain competitive against the encroachment of Big Tech.

Yet, by late 2026, the sentiment has inverted. The panelists at Programmatic IO argued that owning data has become a liability rather than an asset. The emergence of agentic AI—autonomous systems capable of executing complex marketing tasks—has changed the calculus. Agencies no longer want to be the "custodians" of client data; they want to be the "architects" of the intelligence layers built on top of that data.

The Ownership Dilemma: Why Agencies Are Shedding Data

The central theme of the Programmatic IO discussion was the inherent conflict of interest in agency-owned data. Sir Martin Sorrell, ever the industry provocateur, cut to the heart of the matter: "Ownership of data brings problems." When an agency owns the data, the measurement, and the media-buying engine, they are effectively "grading their own homework."

The "Black Box" Problem

Advertisers are increasingly wary of the "black box" nature of AI-driven campaign optimization. With dentsu projecting that algorithms will drive 75% of global ad spend by 2028, the pressure for transparency has reached a breaking point. Agencies that insist on using their own proprietary data sets are finding that clients are pushing back, fearing that the agency’s recommendations are biased toward its own ecosystem.

The LiveRamp Precedent

This skepticism was on full display when the topic of Publicis’s planned acquisition of LiveRamp arose. When asked how an agency should advise a client regarding a data partner that is being swallowed by a competitor, Sorrell did not mince words: "We’d recommend clients work with the best possible alternative." This sentiment highlights a growing industry consensus: neutrality is the new competitive advantage.

Data Sovereignty: The Horizon Media Blueprint

Bob Lord of Horizon Media offered a blueprint for how agencies can navigate this shift. Horizon’s internal AI platform, blu. AI, is built on the principle that the client must retain ownership of both the underlying data and the specific models trained upon it.

"The client has to retain that data ownership, because then, if you ever have a breakup, the agency can’t take that data with them," Lord explained. He noted that in his career, he has witnessed scenarios where agencies "held data hostage" or, worse, applied the insights gained from one client to the campaign of a direct competitor. By formalizing client ownership, agencies are protecting their brand partners and, in turn, building long-term trust that is often absent in transactional relationships.

AI and the Disruption of the DSP

The rise of AI is not just changing data ownership; it is fundamentally altering the role of the Demand-Side Platform (DSP). Shane McAndrew of WPP noted that the interaction between machines is accelerating at a pace that renders current human-led buying processes obsolete.

"We’re starting to see it’s less human and more machine," McAndrew said. "We can go agent-to-agent and get more intelligence than we’ve ever gotten out of those DSPs."

The Threat to the Middleman

The implication for the programmatic ecosystem is profound. As AI agents become better at ingesting signals directly from supply-side platforms (SSPs) and publisher data, the "data marketplaces" embedded within DSPs face an existential threat. Supply-chain optimization is no longer just a manual task performed by analysts; it is an automated outcome driven by AI agents that can find efficiencies in milliseconds.

The End of the FTE Model: A New Compensation Paradigm

Perhaps the most significant implication of this shift is the potential collapse of the Full-Time Equivalent (FTE) compensation model. For decades, agencies have charged clients based on the number of hours or heads assigned to an account. In an AI-first world, this model is fundamentally misaligned.

The "Efficiency Trap"

"We’re not going to cost-cut our way to growth," McAndrew warned. He argued that while agencies must automate low-value, high-volume tasks, they must resist the urge to pass those savings to clients via headcount reduction. Instead, the focus should shift to "delivering the right set of intelligence."

Outcome-Based Pricing and SaaS

Sorrell advocated for a total shift to outcomes-based pricing, where agencies are compensated based on the business results—sales, brand lift, or market share—they deliver, rather than the effort expended.

Simultaneously, Bob Lord proposed that agencies should lean into their roles as software providers. By licensing AI tools directly to clients, agencies can offer a tangible, "hands-on-keyboard" experience. This changes the agency from a service provider to a technology partner, allowing for recurring revenue streams that aren’t tethered to the traditional, and increasingly inefficient, FTE model.

Implications for the Future: The Agency as Validator

As the industry looks toward 2030, the role of the agency will likely undergo a final, radical transformation. Sorrell predicted that the enterprise agency model is facing an "existential threat" from Big Tech platforms—like Meta—that are building end-to-end creative, planning, and distribution engines.

In this future, the agency’s role will not be to perform the "grunt work" of media buying, but to serve as the validator. As AI handles the execution, the agency becomes the guardian of brand safety, the strategist of creative excellence, and the auditor of the machine’s output.

The Hybrid Path Forward

Dentsu’s Maggie Summers articulated the most realistic transition path: a hybrid model. "It’s a mix of paying for the intelligence… and then having a model for outcomes so that we share and appreciate the value of that outcome-driven model."

The agencies that thrive in this environment will be those that embrace this dual-track strategy:

1. Infrastructure Neutrality: Moving away from hoarding data and toward managing it for the client’s benefit.
2. Value-Based Economics: Moving away from headcount-based billing and toward performance-based incentives and software licensing.

Conclusion: The Maturity of the Industry

The shift away from data ownership marks a "coming of age" for the advertising agency sector. By relinquishing the desire to own the "pipes," agencies are freeing themselves to focus on what they were originally built to do: provide the creative and strategic intelligence that machines, however smart, cannot replicate.

The "Frankenstacks" of the late 2010s are being dismantled in favor of agile, agentic, and transparent ecosystems. While the transition from FTE models to outcomes-based pricing will be fraught with organizational inertia, the logic of the market is clear. In the age of AI, the agency that owns the client’s trust will be far more valuable than the agency that merely owns the data.