Three years after Kokai, The Trade Desk is betting on refinement with Zuma

With this update, the plan is to do the opposite.

“I would say UX is a key focus for the Zuma release of Kokai,” said Samantha Ross, senior director, product marketing at The Trade Desk during a press briefing. “As we’ve been listening to customer feedback over the last year or so, one theme has come up consistently, which is users appreciate the power of Kokai, but they want it to be easier to use, easier to find what they’re looking for, and ultimately easier to take action in, and that is exactly what the UX improvements are designed to address.”

That lines up with how some agencies describe Kokai’s evolution over the past year.

“I wouldn’t say it’s been ‘leaps and bounds’. It’s been more of a steady evolution, with a lot of smaller improvements and refinements to the platform,” said Haley Feazell, vp, global media at Mindgruve.

That shows up across three fronts, all aimed at cutting the time traders spend hunting for answers rather than acting on them. Landing pages now surface pacing, delivery and performance trends the moment a buyer logs in, rather than making them click into reports to find what needs attention. Bulk edits, like updating geo targeting across dozens of ad groups at once, now show whether that change will expand or constrain delivery, and by how much before it goes live. And troubleshooting, historically, the single biggest time sink for traders, according to The Trade Desk’s own product team, gets dedicated view surfacing blockers, recommendations and active settings instead of forcing buyers back into the raw settings table.

That emphasis on control is also showing up in the way buyers describe the changes. 

“The platform is evolving to give us more control over the buy, in some ways returning to the control we had in Solimar,” said Tom Swierczewski, vp of media investment and partnerships at Goodway Group. “The upcoming updates add more views and a simplified UI, making it easier to skip features that don’t support a given buying strategy, while increasing control over the features we consider important for each client.”

Sitting underneath all of it is the Koa AI assistant, entering closed beta as part of this release. It incorporates six agents on the back-end: help and search, campaign creation, insights, troubleshooting, frequency and audience building. The campaign creation agent can read a media plan in whatever spreadsheet format a buyer hands it, flag missing details and inconsistencies, and ask clarifying questions before building out a structure, though a human still has to approve it before anything goes live. Troubleshooting, for now, only diagnoses and recommends. The Trade Desk said the ability for it to apply fixes itself is coming later, which is really the direction all of this is headed: less time spent asking Koa what’s wrong, more time letting it fix things.

“Today, it is only surfacing recommendations, and it can guide you to where you need to go to take that action,” Ross said. “But over time, we will introduce the ability for it to actually apply those fixes for the user if and how the user chooses to deploy Koa on their behalf.”

That’s also where The Trade Desk still has something to prove.

“The Trade Desk seems to still be playing catch up in terms of their AI features,” said Dominic Johnson, associate director of brand media at Collective Measures, who noted that some AI capabilities TTD has in beta are already available on other DSPs.

The bigger thing, though, is how much control The Trade Desk is willing to hand off, and to whom. Moving forward, it will open up the Kokai AI layer to agents that don’t live inside its own platform. Alongside Koa AI assistant, the in-platform tool now entering closed beta, the ad tech vendor lets clients route the same underlying agents through whatever AI tool they already use to run their desks. During the briefing call, The Trade Desk’s Ross cited Claude as an example. It’s a notable concession for an ad tech vendor that has spent years building out its own interface. Rather than betting buyers will come to Kokai for their AI needs, it’s betting they’ll keep using their own, and building the plumbing so Kokai’s data and agents can meet them.

Then again, with the company now claiming all clients using Kokai, up from 85% default adoption it reported just last quarter, that concession looks less like an olive branch and more like a vendor that no longer needs to win the interface fight. Having spent three years, a few bruising earnings updates and tensions with some of the largest holdcos, The Trade Desk doesn’t need them living inside it for everything. It just needs them to buy media through Kokai. Where they go to ask an AI for help along the way matters less.

That shift is reflected at the agency level too.

“All of our clients are now on Kokai,” said one U.S. agency exec who preferred to remain anonymous. They added that their agency transitioned its full client base at the beginning of this year, in part to ensure teams were working from the same platform and taking advantage of its latest capabilities.

The third pillar of these changes is measurement, and it leans directly on language The Trade Desk’s CEO Jeff Green has used on earnings calls about moving buyers off proxy metrics like click-through rate (CTR) and cost-per-action (CPA) toward proof of actual sales impact. For instance, lift studies, which measured whether an ad campaign actually drove more sales or conversions rather than just clicks, used to take five to seven days to set up.

Now they launch in one click and go live within 48 hours. A new GeoLift study lets buyers tie campaign spend to incremental lift in certain markets around brick-and-mortar locations, and a Nielsen IQ partnership brings automated conversion, sales and brand lift studies into the platform for CPG advertisers. Asked what specifically enables the connections between the data and measurement pieces on a technical level, The Trade Desk’s team said they’d need to follow up in writing.