IAB releases agentic AI roadmap for digital advertising | MarTech
Briefly

IAB releases agentic AI roadmap for digital advertising | MarTech
"IAB Tech Lab released an agentic roadmap outlining how the digital advertising industry can scale agentic buying and selling without tearing down what already works. The plan focuses on extending existing standards with new agentic and high-performance protocols, rather than creating entirely new ones, to support secure and interoperable agentic execution. To move faster, the organization is also making a dedicated investment in AI development, including adding new engineering resources."
"At the core of the roadmap are well-established standards the industry already relies on, including OpenRTB, AdCOM, OpenDirect, VAST and the Deal API, along with measurement and privacy frameworks like Open Measurement, the Conversion API, GPP and TCF. These are being paired with newer protocols such as Model Context Protocol, Agent2Agent and gRPC to support machine-speed execution and coordination between independent systems, while maintaining governance and trust."
IAB Tech Lab released an agentic roadmap that extends existing standards with agentic and high-performance protocols rather than replacing current systems. The roadmap pairs established standards (OpenRTB, AdCOM, OpenDirect, VAST, Deal API, Open Measurement, Conversion API, GPP, TCF) with newer protocols (Model Context Protocol, Agent2Agent, gRPC) to enable machine-speed execution while maintaining governance and trust. The organization will invest in AI development and add engineering resources to accelerate implementation. Planned 2026 deliverables include open-source buyer and seller agent reference implementations, standardized agent profiles, tooling for production agentic workflows, a January 28, 2026 webinar, and in-person Agentic AI boot camps starting in February. The approach emphasizes open, shared standards to scale agentic execution responsibly and avoid fragmentation.
Read at MarTech
Unable to calculate read time
[
|
]