AI and agents are changing the shape of work: not just how we write code, but how we capture, reuse, and operationalize knowledge inside organizations. The dilemma for data products and data mesh is that most firms still treat the mesh as an operating model—ownership, domains, and principles—without the practical machinery to make it fast, coherent, and trustworthy at scale; as a result, data products remain slow and hard to build, definitions drift, tribal knowledge stays trapped in people’s heads, and “self-serve” quietly becomes “ask the same three experts again.” This talk will show how an Agentic Mesh is the next stage of evolution for data mesh: a layer of always-on, action-oriented agents that capture both explicit and tacit knowledge, enforce contracts and policy, monitor quality and usage, and wrap core data products with intelligence—so the data products you ship (and the mesh that connects them) become genuinely smart, not just well-intentioned.