The Trade Desk Kokai Zuma adds agentic AI to automate campaigns

The shift toward agentic AI represents a move from passive automation to active decision-making within the ad tech stack. By automating frequency settings and audience discovery, The Trade Desk aims to reduce the manual overhead required to manage fragmented streaming and digital inventory. This development pressures competitors to prove similar performance gains as buyers increasingly prioritize platforms that can demonstrate measurable CPA improvements through machine learning. As the open internet competes with walled gardens for budget, the ability to simplify complex measurement via tools like Conversion Lift will be critical. Watch for whether the reported 32 percent CPA lift holds steady as the update rolls out to the full global client base.

The Trade Desk is not alone in pushing agentic AI into programmatic buying. In May 2026, Google announced that its Performance Max campaigns now use agentic AI to autonomously adjust creative, bidding, and audience signals across YouTube and Display, a direct competitive response to demand-side platforms that promise similar automation. Meanwhile, Amazon Ads expanded its AI-powered campaign builder to include automated audience sequencing for streaming TV inventory in June 2026, intensifying pressure on independent DSPs to demonstrate measurable performance advantages over walled-garden alternatives.

On the business side, The Trade Desk has been investing heavily in measurement infrastructure to support Kokai's evolution. In April 2026, the company reported first-quarter revenue of $662 million, a 20 percent year-over-year increase, driven by growth in connected TV and international markets. That financial momentum gives The Trade Desk resources to fund aggressive product development like Kokai Zuma. The company also launched Audience Unlimited in early 2026, a privacy-first identity solution designed to replace third-party cookies across open-web inventory, positioning Kokai Zuma's agentic capabilities as the execution layer on top of that identity foundation. Jeff Green has framed this combination as the open internet's answer to walled-garden lock-in, arguing that automated optimization on transparent data will outperform opaque platform algorithms.

Technical validation of agentic AI in ad buying remains limited but growing. A W Media Research survey published in July 2026 found that 41 percent of programmatic buyers had tested at least one agentic AI feature in the prior six months, with automated audience discovery cited as the most frequently tested capability. The Trade Desk's claimed 32 percent CPA improvement aligns with broader industry benchmarks: a separate analysis by AdExchanger in June 2026 noted that DSPs deploying machine-learning-driven frequency optimization reported median CPA reductions between 18 and 28 percent across display and CTV campaigns. Kokai Zuma's performance claims sit at the upper end of that range, which will draw scrutiny from buyers who expect independent verification as the rollout expands beyond early-access clients.