# AI Connectivity: Kong AI Gateway 2.0 Delivers the Infrastructure to Run and Govern Enterprise Agents

This keynote explains Kong’s vision for AI connectivity: the infrastructure enterprises need to make large numbers of AI agents useful, governed, and cost-effective at scale. Using a biking analogy, the speaker argues that agents, like bikes, only become efficient when supported by the right lanes, controls, and coordination, which in practice means platform engineering, APIs, events, models, MCPs, and an AI gateway that delivers context and access to LLMs. The presentation highlights Kong’s move from API management to AI connectivity, including AI Gateway 2.2, Context Mesh, a webhook engine for real-time agent actions, and private beta features such as an agent and MCP registry, agent observability, token vault for identity and credential mapping, and AI cost management with pricing, attribution, anomaly detection, savings analysis, and enforceable budgets. Customer examples from Siemens Energy and FanDuel underscore the real-world challenge of governing thousands of agents without slowing developers down.