AI models and agent frameworks are advancing rapidly, but intelligence alone does not make an organisation agentic. To create real value, agents must safely interact with enterprise systems, data, people and business processes.
This keynote explores the architectural challenge beyond the model: building a control and execution layer for enterprise AI. MCP is an important step toward interoperable agent-to-tool interaction, but production systems also require identity, least-privilege access, policy enforcement, human approvals, orchestration, observability and auditability.
The session will offer a practical framework for thinking about how open protocols and agent systems can move from impressive demonstrations to trusted, scalable enterprise action. The central question is not whether an agent can use a tool, but whether an enterprise can safely allow it to do meaningful work.
___________________________
Presentation Language: English
Captioning will be available for attendees in 50+ languages through Wordly. See instructions in each room to utilize captioning.
