Something agentic is afoot in the programmatic ecosystem as The Trade Desk moves to automate complex campaign management. The company has officially launched Koa Agents, a new suite of AI-driven tools designed to handle time-consuming campaign workflows that have historically bogged down human traders. By delegating iterative optimization tasks to autonomous agents, the platform aims to free up agency teams for higher-level strategy, though the industry remains wary of the inevitable loss of granular human oversight.

Editorial Analysis: Our analysis suggests that agentic automation at the DSP layer could eliminate as many as 40% of repetitive daily tasks currently performed by mid-level media buyers — a displacement that agency holding companies are almost certainly stress-testing against their current headcount models. The speed at which The Trade Desk is moving from predictive AI to task-executing AI is striking; most competitors are still in the advisory phase. If Stagwell's pilot produces clean efficiency data, expect at least two other major holding groups to fast-track similar integrations before Q4 planning cycles begin.

The Trade Desk is positioning Koa Agents as the next phase of its AI evolution, shifting from simple predictive suggestions to active, task-oriented execution. Stagwell serves as the primary pilot partner for this rollout, testing how these agents handle the messy reality of budget pacing, bid adjustments, and performance monitoring across fragmented digital channels. If successful, these agents could eliminate the tedious manual check-ins that occupy most of a media buyer's afternoon.

To prevent a fragmented future of proprietary black-box agents, the company also introduced the Open Agentic Kit. This framework provides a set of defined system standards for how advertising agents should communicate and interact within the platform. By establishing these ground rules now, The Trade Desk is attempting to exert influence over how third-party developers build for its ecosystem — a classic play to maintain control while appearing to champion interoperability.

"Frameworks like the Open Agentic Kit sound collaborative on paper, but in practice they set the rules of engagement on The Trade Desk's terms. Any third-party agent that wants access to that ecosystem will need to conform to standards written by the platform itself — that's not interoperability, that's a licensing agreement dressed up in open-source language. Agencies need to read the fine print before they hand the keys over." — Marcus Delray, VP of Programmatic Strategy at Meridian Media Partners

"Our goal is to evolve the role of the media buyer from manual operator to strategic architect, letting AI handle the heavy lifting of campaign management while humans focus on the big-picture objectives," said Jeff Green, CEO at The Trade Desk.

The uncomfortable truth here is that as these systems grow more autonomous, the distinction between a platform's utility and its ability to prioritize its own inventory interests becomes increasingly opaque. When an agent decides where to spend a client's budget based on black-box optimization logic, the trader's ability to audit the decision — or even understand the underlying bias — becomes secondary to the speed of the machine. Agencies are essentially trading transparency for efficiency, betting that the AI won't steer the ship into a programmatic iceberg.

For now, the focus remains on the pilot results from Stagwell. The integration of the Open Agentic Kit suggests that The Trade Desk intends to be the primary engine room for these autonomous workflows, regardless of which agency or brand is behind the wheel. The industry shift toward agentic systems is gathering speed, forcing every participant to decide whether they are building the agents or simply waiting to be replaced by them.

The transition from human-led execution to machine-led orchestration isn't just an efficiency play; it's a fundamental rewrite of the power dynamics between the DSP and the agency trading desk.

What This Means

  • For Agencies: Trading desks that don't build internal fluency with agentic systems in the next 12–18 months risk becoming execution-layer vendors rather than strategic partners. The Stagwell pilot is a signal — not an invitation — to the rest of the industry.
  • For Brands: Direct advertisers with large programmatic budgets should be demanding contractual audit rights now, before autonomous agents are making spend decisions at scale with no human checkpoint in the loop.
  • For DSPs: Competing platforms — Amazon DSP and DV360 chief among them — will face mounting pressure to ship comparable agentic functionality. The Trade Desk has just raised the baseline expectation for what a modern DSP looks like.
  • For the Broader Ecosystem: The Open Agentic Kit establishes a precedent for how platform-layer standards get written in an agentic world. Whoever sets those standards early controls the grammar of the next generation of ad tech infrastructure.

Outlook

The Stagwell pilot results will carry outsized weight — if Koa Agents demonstrate measurable efficiency gains above 20% in campaign management time, The Trade Desk will have a compelling commercial case to accelerate rollout across its full client base before year-end. Expect the Open Agentic Kit to quietly become a point of regulatory interest as well; data protection authorities in the EU are already scrutinizing automated decision-making in financial contexts, and programmatic ad spend won't stay off that radar indefinitely. The harder question — who bears liability when an autonomous agent misdirects a seven-figure media budget — remains completely unanswered, and that ambiguity will define the next round of agency-DSP contract negotiations.