The Rise of the Agentic Ad Buyer: How AI is Reshaping Connected TV

The landscape of digital advertising is undergoing a tectonic shift, one driven not by a new platform or a change in consumer viewing habits, but by the quiet emergence of "agentic AI." While artificial intelligence has been the centerpiece of tech discourse for years, the practical application of AI agents—autonomous systems capable of executing complex, multi-step tasks across disparate software—is only now reaching the high-stakes world of connected TV (CTV) ad buying.

For years, "buying ads" meant human traders manually navigating demand-side platforms (DSPs), analyzing spreadsheets, and executing bids. Today, that model is being challenged by a new workflow that leverages the Model Context Protocol (MCP) to allow AI agents to "talk" to ad tech platforms directly. As this technology matures, it is moving from the realm of hypothetical R&D into the hands of agencies like InterMedia, signaling a fundamental transformation in what it means to be a media buyer.

The Chronology of an AI Revolution

To understand the speed of this evolution, one must look at the recent timeline of the underlying infrastructure. The concept of the AI agent is not new, but its ability to interface with professional-grade software was severely limited until late 2024.

• Late 2024: Anthropic introduces the Model Context Protocol (MCP), an open-source standard designed to allow AI agents to securely connect to external data sources and applications. This provided the "handshake" necessary for AI to interact with business tools.
• 2025: A coalition of ad tech firms rallies around the development of the "Ad Context Protocol," an extension built on MCP specifically designed to translate the nuances of ad demand into language that LLMs can process and execute.
• Mid-2026: Vibe.co, a self-serve CTV platform recently acquired by Walmart, rolls out deep integrations with Claude. Simultaneously, Olyzon, a CTV ad-decisioning startup, launches its agentic orchestration layer during Advertising Week New York.
• Present Day: Independent agencies like InterMedia have begun migrating the majority of their biddable campaigns to these agentic workflows, marking the first real-world shift from human-executed bidding to human-supervised, AI-orchestrated media buying.

Supporting Data: Why Now?

The transition to agentic buying is not merely a trend; it is a response to the crushing complexity of modern digital marketing. As the number of streaming platforms, publisher inventory sources, and display channels has ballooned, the cognitive load on human media buyers has become unsustainable.

According to Arthur Querou, Co-Founder and CEO of Vibe.co, the adoption of MCP-integrated tools has been swift. More than 1,000 brands have already tapped into Vibe’s MCP-enabled environment, with a significant percentage of total ad spend on the platform now flowing through agentic channels.

"SMBs use these tools because they have limited resources," Querou explains. "They want the sophistication of an enterprise-level trading desk, but they need the efficiency that only an AI agent can provide."

The technical distinction between these integrations is telling. While Vibe leverages the conversational flexibility of MCP for Claude-based dashboards, Olyzon has prioritized deterministic API connections. Jules Minvielle, CEO of Olyzon, notes that while agents are brilliant at synthesis, the "last mile" of ad buying—executing a bid—requires the absolute consistency of a hard-coded API. "An agent talking to an agent is not the same as having our agents writing what’s needed directly into the API," Minvielle says. This hybrid approach ensures that while the strategy is AI-driven, the execution is mathematically verifiable.

Official Perspectives: The View from the Frontline

For agencies like InterMedia, the shift has been immediate and profound. David Nyurenberg, SVP of Digital, and Sydney Brower, Director of Digital Media and Strategy, represent the new breed of "AI-augmented" media professionals.

During a demonstration, Nyurenberg prompted an AI agent to analyze the performance of CTV campaigns across major publishers like Peacock and Warner Bros. Discovery. Within seconds, the agent provided a granular report—not just showing spend, but identifying growth trends in specific inventory segments and suggesting optimizations.

"It changes what it means to be an expert," Nyurenberg explains. "Instead of being an expert in how to navigate a platform, how to push the right buttons, and knowing where to click, you become an expert in being judicious about what changes to make. You are no longer a mechanic; you are an architect."

Brower adds that the efficiency gain is palpable. By centralizing reporting and decision-making within a single dashboard—Claude’s interface, in this instance—campaign managers regain "time and brain power" that was previously lost to toggling between tabs and cleaning up data. "It’s about elevating the work," she says. "We can now devote our energy to deeper, client-driven strategy rather than manual labor."

However, both remain cautious. Brower uses a vivid analogy: "You can trust AI to fly a plane, but I still want a pilot in the cockpit to watch it." The "human-in-the-loop" model remains the gold standard for risk management. Before an agent executes a bid change, Brower verifies the parameters within the DSP dashboard. This skepticism is well-founded, given the recent, well-publicized failures of consumer-facing AI agents that have hallucinated or exposed sensitive user data.

The Implications for the Industry

The rise of agentic ad buying creates a series of complex questions for the future of the advertising ecosystem.

The Question of Ownership

As the lines between agency tools, SaaS platforms, and AI models blur, a sense of digital ambiguity emerges. If an agency uses an AI agent built by a third-party platform (like Olyzon) to manage ads on a third-party publisher (like Vibe), who "owns" the agentic decision? This echoes the modern dilemma of digital asset ownership—if an AI agent makes a $50,000 mistake, is it a software bug, an agency error, or a prompt-engineering failure?

Standardization Hurdles

Despite the promise of MCP, the industry remains fragmented. DSPs often use conflicting terminology for identical metrics. Currently, human agents must perform "manual standardization" to ensure that the AI is comparing apples to apples. Until the industry agrees on a universal data language, these AI agents will continue to require a human translator to bridge the gap between platforms.

The Evolution of the "Agency"

The most significant implication is the existential question facing agency holding companies. For decades, agencies have built their value proposition on access to proprietary tools and massive, manual teams. In an agentic world, these advantages diminish.

Nyurenberg predicts that while large holding companies will attempt to build proprietary agents, many will fall back on "stitched-together" tech from SaaS partners. The agency of the future will be a boutique of "AI-enabled marketers"—professionals who understand the interplay of brand strategy and agentic logic.

Jules Minvielle offers a provocative forecast: "In 18 months, every single media buyer in the world will be equipped with a platform like ours." Whether this leads to a commoditization of media buying or a new renaissance of creative strategy remains to be seen.

What is certain is that the "button-pusher" era of digital advertising is coming to a close. The future belongs to those who can effectively command the agents that are now, quite literally, taking the wheel of the global media machine. The technology may be nascent, but its trajectory is clear: ad buying is becoming an exercise in oversight, orchestration, and high-level judgment. The cockpit is crowded, but the pilot is finally—and perhaps for the first time—looking at the horizon instead of the instrument panel.