The Agency Existential Crisis: How AI is Forcing a Reckoning for Global Consultancies

By Editorial Staff
October 7, 2026

The traditional agency model—once the bedrock of marketing, strategy, and business growth—is facing an existential threat that is no longer theoretical. As generative AI models reach a level of sophistication that allows them to synthesize vast quantities of data into actionable business intelligence in seconds, the "generalist" pitch that has long defined giants like McKinsey and top-tier advertising holding companies is rapidly losing its luster.

At Advertising Week in New York City this week, industry leaders gathered to address a sobering reality: if a machine can perform the duties of a second-year analyst or a mid-level account manager for a fraction of the cost, what is the value proposition of the human agency?

The End of the "Generalist" Era

For decades, the standard agency pitch has been one of breadth: "We have seen everything, and here is a deck to prove it." This reliance on human synthesis—taking disparate market signals and packaging them into a PowerPoint presentation—is now being disrupted by agentic AI.

Brian O’Kelley, CEO and co-founder of Apostra, was blunt during a panel discussion. "If you look at what [AI models like] Fable says versus what your second-year analyst says, it’s kind of the same," O’Kelley noted. "If your job is to make PowerPoints, I’m not really sure that’s going to be super valuable in a world where in, like, 20 seconds, Fable can make me look really smart."

O’Kelley’s perspective carries weight, particularly given his own recent career pivot. His firm, formerly known as the digital carbon measurement company Scope3, rebranded to Apostra last month, signaling a complete shift toward agentic media buying. This 180-degree turn reflects a broader industry recognition: the sustainability of the agency business model now depends entirely on its ability to integrate, rather than merely advise on, artificial intelligence.

Chronology of the Shift: From Consulting to AI Integration

The transition of the agency world from creative service providers to technological hubs has been accelerating since 2019. To understand the current tension, one must look at the recent timeline of institutional evolution:

• 2019: Publicis Groupe acquires Epsilon for $3.95 billion, a foundational move toward becoming a data-first organization rather than a pure creative shop.
• 2022: Brian O’Kelley launches Scope3, focusing on supply chain emissions—a niche that proved the importance of specialized data measurement in a fragmented digital landscape.
• 2025–2026: The explosion of agentic AI models (such as Fable) creates a commoditization of basic research and strategic planning, rendering standard generalist billable hours less defensible.
• September 2026: Scope3 officially rebrands to Apostra, signaling an industry-wide pivot toward automated media orchestration.
• October 2026: Publicis supplements its Epsilon-centered data play with a $2.2 billion acquisition of LiveRamp, effectively doubling down on its proprietary data collaboration capabilities.

The Holdco Response: Building Proprietary Moats

While individual agencies grapple with the fear of obsolescence, the major holding companies (Holdcos) are attempting to build "moats" through technology and proprietary data stacks. They are no longer content to act as intermediaries; they are seeking to become the infrastructure upon which marketing is built.

WPP has leaned heavily into "WPP Open," an AI marketing platform designed to be accessible to clients directly, effectively democratizing their own strategic tools. Publicis, through its Epsilon and LiveRamp integration, is constructing a closed-loop ad network that prioritizes data identity over creative messaging. Meanwhile, Omnicom is doubling down on "Omni," an Acxiom-fueled AI platform designed to cut out the "middleman" of traditional ad-tech vendors.

However, O’Kelley warned that these large-scale shifts are not a panacea. "Every agency, and this includes Horizon, has to make some pointed choices," he said during the panel. The implication is that even the largest firms cannot be everything to everyone without losing the precision that AI now demands.

Official Responses: The "Overarching Brain" Strategy

Dane Kunkel, SVP of performance and transformation at Horizon Media, stood as the counter-point to O’Kelley’s skepticism. Sitting alongside him on the panel, Kunkel acknowledged the pressure but framed it as an opportunity for transformation rather than a death knell.

"Our department was spun up to help our clients transform," Kunkel explained. According to Kunkel, the agency’s role is shifting toward becoming the "overarching brain" of a client’s operations. This involves using AI to handle the tactical execution of media buying and data processing, while human experts focus on cross-departmental integration—connecting finance, operations, and marketing.

"We’re pitching ourselves as new growth partners," Kunkel said. "We’re trying to change the way we’re serving clients and moving away from full-time equivalent (FTE) models so we can drive outsized value." This shift away from billing based on headcount is perhaps the most significant change in the agency business model in a century.

Abhay Singhal, CEO and co-founder of InMobi, offered a supportive, if cautionary, perspective. "I actually think agencies are in a phenomenal place, as long as they’re able to reimagine what they do," Singhal noted, pointing to the etymological roots of the word "agency"—the act of acting on behalf of a client. In the age of AI, the agency is not just a creative advisor; it is the operator of the client’s autonomous business agents.

Implications: The Rise of the "Forward Deployed Engineer"

If the agency of the future is to survive, it must change its talent profile. A recurring suggestion at Advertising Week was the adoption of the "forward deployed engineer" model. Popularized by firms like Palantir, these engineers do not sit in a remote office; they are embedded within the client’s organization to customize AI systems in real-time.

Singhal went as far as to suggest that every media planner should effectively be rebranded as a forward-deployed engineer. Yet, this creates a significant economic friction point.

"You can’t just go to a media planner and say, ‘Go build database integrations with a bunch of internal systems, and also think about security and scale,’" O’Kelley countered. He argued that agencies simply do not have the capital to compete with tech giants like Meta or Google for top-tier AI engineering talent. While agencies are currently outsourcing their engineering needs to consultancies in India, this remains a stop-gap measure that fails to capture the core intellectual property of the AI revolution.

The Path Forward: Deep, Not Wide

The consensus among industry observers is that the "generalist" model is dead. For an agency to survive the next five years, it must embrace a strategy of radical verticalization.

"AI is pretty damn smart about almost everything," O’Kelley noted. "It’s better to be the best at one little thing with agents that help you do it really well than to be a generalist."

Agencies must now choose their battlegrounds. Whether it is becoming the premier firm for pharmaceutical marketing, mastering the nuances of linear TV in a digital-first world, or building the most robust AI loops for creative optimization, the future belongs to the specialist.

The era of the "deck-building" agency is coming to a close. The era of the "AI-orchestrating" agency has begun. For those that can pivot from selling broad-based advice to providing hyper-specialized, agentic-driven value, the future is not necessarily one of decline, but of profound evolution. For those who cling to the generalist model, however, the encroaching shadow of automated intelligence is likely to prove insurmountable.