Agentic AI in Action: Architecting Autonomous Systems

AI agents aren’t coming; they’re already here, quietly taking on work humans used to do. As systems begin to plan, reason, and execute tasks independently, the way software is built is fundamentally changing. In this session moderated by Innodata, leaders from OpenAI, Meta, IBM, and Mastercard break down what’s actually happening beneath the surface of agentic AI. From orchestration layers and tool access to memory management and safe execution, the conversation will explore how autonomous systems operate in real-world environments. Panelists will also examine what it takes to trust these systems, how you can govern, monitor and design to fail safely, as agents move from demos into meaningful, production-grade use.