The Trade Desk rebuilds Kokai as Zuma with Ask Koa agents for audience building, pacing and troubleshooting
The Trade Desk's new architecture pairs Kokai's programmatic depth with conversational Ask Koa agents for audience building, pacing and troubleshooting.
Mariano De Vitto · September 2026
The Trade Desk has taken its Kokai platform and rebuilt it from the ground up into a new architecture called Zuma, introducing a suite of conversational AI agents under the name Ask Koa. Rather than forcing media buyers to click through dashboards and reports, Zuma lets teams direct campaign work through natural language, marking what the company describes as its most significant platform overhaul since Kokai first launched.
The Trade Desk has spent years positioning Kokai as the backbone of its programmatic offering, giving agencies and in house teams granular control over bidding, audience targeting and reporting across the open web. That granularity has always come with a tradeoff of complexity. Trading desks often need specialists who know exactly which screen, filter or report holds the answer to a pacing question or a delivery problem. Zuma is The Trade Desk's attempt to keep that depth of control while removing the friction of navigating it manually.
From Clicks to Conversations
According to details published by The Trade Desk on its Zuma Q3 2026 release page, Ask Koa is built to live inside the Zuma environment and respond directly to plain English requests, adjusting bids, explaining why a campaign is underdelivering, or assembling a new audience segment on the spot. These are tasks that previously required moving across multiple screens and often relied on a trading desk specialist's memory of where specific controls lived. By compressing that operational layer into a chat interface, The Trade Desk is betting that agents and in house media teams will interact with the platform more like a collaborator than a tool.
This move fits a pattern that has been building across ad tech for the past couple of years. Rather than asking buyers to learn entirely new interfaces, vendors are layering conversational agents on top of existing infrastructure, hoping to lower the learning curve for junior staff while giving experienced traders a faster way to execute routine actions. The Trade Desk joining that pattern with Zuma signals that natural language interaction is becoming an expected feature of programmatic platforms rather than a novelty.
What Buyers Should Watch
For marketers and agencies running programmatic budgets through The Trade Desk, Zuma signals a shift in where day to day trading effort gets spent. Repetitive pacing checks and troubleshooting, the kind of work that has traditionally been assigned to junior analysts, are becoming candidates for automation through Ask Koa. That does not eliminate the need for people who understand programmatic strategy, but it does change what those people spend their time doing, moving them away from manual monitoring and toward interpreting results and making judgment calls.
Why It Matters
Media teams should start evaluating which trading tasks are routine enough to hand to conversational agents like Ask Koa, and where human oversight remains essential. The shift changes both the required skill set for trading desk staff and how time gets allocated across programmatic operations, and teams that treat this as a workflow question now will be better positioned than those that wait for the tooling to force the change on them.
The Signal Brief · Mariano De Vitto — Head of Marketing, Barcelona