Programmatic media buying · PUBLIC RESEARCH BRIEF
The Trade DeskWhich control makes AI-guided media buying easier to trust?
The Trade Desk’s August 27, 2026 announcement describes Kokai Zuma as a platform release for planning, buying and measuring advertising across the open internet, with an updated interface, agentic AI and a simplified measurement framework. Its Kokai measurement materials distinguish reach from inventory quality and business outcomes. These are vendor-described capabilities, not evidence that automation improves a particular campaign.
Updated 2026-10-17 · Simulation results not yet generatedCHANGE ONE THING. LEARN WHAT MATTERS.
Three questions for the GTM team.
Which campaign changes should require explicit human approval rather than automated execution: audience expansion, bid shifts, budget movement, inventory changes or measurement configuration?
Set up this study →Would a measurement story centered on relevant reach, inventory quality or verified business outcomes best support the next budget discussion? Keep media delivery constant and identify what evidence finance would accept.
Set up this study →PROPOSED AUDIENCE
Who should weigh in?
North American brand and agency media leaders, programmatic traders, measurement teams, finance partners and privacy reviewers evaluating The Trade Desk. Include hands-on operators and budget owners from the same buying committee. Proposed audience only; no private campaign, audience or customer data is included.
TWO TIME HORIZONS
Trial today. A habit tomorrow?
Near term · 0–90 days
Over 0–90 days, test synthetic campaign plans with matched recommendation explanations and approval controls. Measure decision accuracy, unsupported confidence, review time and escalation needs; execute no live media changes.
Longer term · 3–12 months
Over 3–12 months, follow consenting teams through repeated planning and optimization cycles to study whether control preferences change with experience, campaign stakes and model updates. Separate workflow efficiency from incremental campaign performance.
What would make the result actionable?
Use versioned synthetic campaigns, known constraints, hidden ground-truth errors, role-specific tasks, approval logs and a predefined measurement hierarchy. Any business-impact model requires controlled media tests, platform delivery data, independent outcome validation and complete fees.
A Gather simulation returns hypothetical customer reactions. Quantifying revenue, traffic or retention needs actual business inputs and validation against observed behavior.