|
DL: Carat, NBCUniversal, FreeWheel and Newton have built an AI agent tool that helps plan and optimize upfront TV commitments, with an unnamed luxury retailer as the first client. The article argues the big question is no longer whether AI can recommend media but who gets to approve a decision once it puts an advertiser’s money on the line. Other agencies are taking different approaches. Tinuiti has opened its measurement data to outside agents, and Stagwell is positioning its Machine OS as an operating layer for its relaunched media group. The author sees the competition turning on data access and permissions as much as on the AI itself. Executives from Dentsu, FreeWheel and WPP Media all described human oversight as a built-in feature. Routine tasks get lighter supervision, while anything that commits budget or affects activation needs a person’s sign-off. Marketers should evaluate agency AI. They should know whether the tool only recommends or can actually place buys, set spend thresholds that require a named approver, understand what data drives the recommendations, and insist on records of what the agent suggested, what was approved and how it performed.
Read More: Content Grip
|