The promise of AI-powered advertising has long hovered between pilot projects and PowerPoint decks. PubMatic and independent agency Butler/Till are now pushing that promise into live market reality.
The two companies have launched what they describe as a fully autonomous, end-to-end agentic advertising campaign across connected TV (CTV) for Clubtails, a flavored malt beverage brand from Geloso Beverage Group. Unlike earlier “AI-assisted” efforts, this campaign removes humans from day-to-day execution almost entirely—allowing software agents to interpret strategy, activate media, and optimize performance in real time.
For an industry that still relies heavily on manual workflows, fragmented platforms, and opaque optimization logic, the move marks a notable inflection point.
From AI-Assisted to Agent-Led Advertising
What makes this campaign different is not just automation—it’s autonomy.
Butler/Till submitted a natural-language media brief inside Claude, Anthropic’s generative AI platform. From there, PubMatic’s agentic application layer took over. Its AI agents interpreted the brief, built the media strategy, set up the campaign, activated inventory, and began continuous optimization across premium programmatic CTV supply.
No media plans built in spreadsheets. No manual trafficking. No human-in-the-loop pacing adjustments.
Instead, PubMatic’s agents handled strategy generation, activation, and optimization, making real-time decisions using live performance signals and supply-side data.
This isn’t theoretical. The campaign is live, delivering geo-targeted, premium CTV inventory with contextual alignment and age-appropriate controls—critical for alcohol advertising.
The Infrastructure Behind the Autonomy
Agentic advertising has been discussed widely, but few platforms can support it at production scale. PubMatic argues it can, thanks to what it calls a three-layer Advertising Intelligence architecture.
At the foundation is accelerated infrastructure powered by NVIDIA-enabled computing, designed for microsecond-level inferencing and high-throughput stream processing. This matters because true agentic execution requires real-time decision-making at the speed of programmatic auctions.
Above that sits a real-time signal intelligence layer, ingesting contextual, performance, and supply-side data continuously. On top is the agentic application layer—where autonomous agents interpret objectives, make trade-offs, and execute actions without human prompts.
The Clubtails campaign runs through PubMatic Activate, the company’s platform for agentic buying across its direct, premium supply footprint. Because Activate connects directly to publishers, the workflow also reduces intermediary costs—a practical benefit in a market increasingly focused on supply path optimization.
Why AdCP Matters
A key enabler of the campaign is AdCP (Ad Context Protocol), an open protocol developed by AgenticAdvertising.org. AdCP standardizes how AI agents communicate with advertising systems through transparent APIs.
In practical terms, AdCP allows Butler/Till to operate inside its preferred GenAI environment while still activating campaigns across external platforms like PubMatic. The protocol avoids the walled-garden problem that has plagued earlier automation tools.
This interoperability is crucial. As agencies experiment with multiple GenAI tools—Claude, ChatGPT, proprietary LLMs—the ability to plug into ad platforms without custom integrations could determine which systems scale and which stall.
PubMatic has positioned itself as a key contributor to this emerging ecosystem, having helped shape AdCP and published what it claims is the industry’s first specification for agent-to-agent communication in programmatic advertising.
Operational Efficiency, Measured in Seconds
One of the clearest benefits reported so far is speed.
According to PubMatic, the agentic workflow produced an optimized media plan in seconds, compressing what would normally take days or weeks of discovery, planning, and activation. That efficiency extends beyond planning into live optimization, where agents continuously adjust pacing, targeting, and performance across inventory sources.
For agencies under margin pressure, this matters. Labor-intensive media operations have become harder to sustain as fees compress and complexity grows—especially in CTV, where inventory fragmentation and measurement challenges persist.
Scott Ensign, Chief Strategy Officer at Butler/Till, framed the experiment as part of a deliberate innovation strategy rather than a one-off test.
The agency operates an internal Innovation Fund designed to bring emerging technologies to clients willing to move quickly. In this case, Clubtails became the proving ground for a model that could fundamentally change how agencies operate.
CTV as the Proving Ground
It’s not accidental that this experiment launched in connected TV.
CTV combines premium brand environments with programmatic complexity, making it an ideal test case for agentic systems. Inventory quality, audience controls, and contextual relevance are non-negotiable—especially for regulated categories like alcohol.
The campaign delivers geo-targeted premium CTV placements, matched with contextually aligned creative and age-gated audience controls. These requirements traditionally add friction to planning and execution, but agentic workflows are well-suited to managing such constraints dynamically.
If autonomous systems can perform reliably in CTV, they can likely extend into other high-value formats.
Financial Implications Beyond Efficiency
Beyond operational speed, PubMatic points to financial efficiencies driven by its direct supply path. By activating inventory through PubMatic’s premium publisher relationships, the agentic workflow reduces unnecessary hops in the supply chain.
This aligns with broader industry trends. Advertisers are demanding transparency, fewer intermediaries, and clearer ROI—pressures that have accelerated interest in SPO and direct paths to supply.
Agentic buying doesn’t just automate decisions; it can encode financial optimization directly into execution logic, balancing cost, quality, and performance continuously.
How This Compares to the Rest of the Market
Many ad tech vendors now claim AI-driven optimization, but most still rely on rules-based automation or human-supervised machine learning. Fully autonomous agent-led activation remains rare, especially at scale and with real brand dollars on the line.
Rivals in the DSP and SSP space are experimenting with copilots and AI assistants, but these typically augment human workflows rather than replace them. PubMatic’s approach goes further, shifting humans upstream to strategy and governance while delegating execution to machines.
That shift won’t appeal to every agency or brand. Control, accountability, and trust remain concerns—especially when AI systems make decisions autonomously. But for performance-driven teams seeking speed and efficiency, the appeal is clear.
What Comes Next
Full performance results from the Clubtails campaign are expected in Q1 2026, which will provide the clearest signal yet of whether agentic advertising delivers not just efficiency, but superior outcomes.
If the results hold, this activation could serve as a blueprint for how agencies, publishers, and platforms collaborate in an AI-first advertising ecosystem.
For now, PubMatic and Butler/Till have moved the conversation forward—from what agentic advertising could be, to what it looks like when deployed in the wild.
