"The promise is enticing: agentic AI-driven media trading could (theoretically) wipe out many of the problems caused by its programmatic predecessor. Namely, ad tech middlemen. Instead of having ads pass through countless vendors who each take a cut, agents would effectively cut the line, so more of the money reaches the publisher, and the results are easier to see. If you strip it back entirely, it's like middlemen built the pipes; agents let publishers keep the water."
"Their preference is to see if holding groups or media agencies are willing to work with them directly on agent pilots, where their own seller agent talks directly to the buy-side agent. What they don't want: the recreation of a multi-step programmatic ad tech stack, whereby the various cuts taken by vendors along the way can run as high as 50 percent."
Publishers have moved agentic systems from demos into production for tasks such as RFP interpretation, planning, pacing, and curation. Agentic AI-driven media trading aims to eliminate ad-tech middlemen, reducing vendor cuts and increasing the money that reaches publishers. The envisioned architecture pairs seller and buyer agents coordinating directly, ideally without a coordination agent in the middle, to avoid recreating a multi-step programmatic stack. Publishers prefer pilots with holding groups or media agencies where seller and buy-side agents connect directly. Industry executives describe programmatic as due for an upgrade and are testing agentic approaches to improve transparency and efficiency.
Read at Digiday
Unable to calculate read time
Collection
[
|
...
]