BCG's new survey of 300 global marketing leaders found the bottleneck isn't ambition (96% say AI is reshaping marketing) but plumbing: 43% of teams now spend $15M+ a year on AI, yet only about a third have moved past discrete-task assistance to agent-led workflows.
Why it matters: Most teams are buying agentic tools faster than they can build the data and measurement layer those tools need to actually run autonomously.
Yahoo's new Agent Network lets advertisers plug 23 partner agents (DoubleVerify, IAS, Kochava, Pacvue) into the DSP over open MCP and APIs across targeting, activation, creative, and measurement, instead of being locked to native automation.
Why it matters: The open MCP layer sitting on top of the ad platform is becoming the norm — the question is which partner agents get real operator pickup.
Agentic BrandStore Enterprise hit general availability, enriching catalogs with conversational metadata so ChatGPT, Gemini, and Claude can discover and recommend individual SKUs — riding a 393% YoY jump in AI-sourced retail traffic that converts 42% better.
Why it matters: Answer-engine optimization is moving to commerce catalogs: structuring product feeds so AI agents can read, trust, and recommend them.