As organizations move from AI experimentation to production-scale AI solutions, architects need a clear framework for deciding when agentic patterns are worth using. This session provides a practical, architecture-focused guide to agentic design: what patterns exist, why they matter, and how to choose the right one for the job.

Rather than treating “agents” as a single concept, we’ll unpack the design choices behind them: how agents reason, use tools, manage state, involve humans, recover from failure, and coordinate across tasks. We’ll also discuss the tradeoffs: latency, cost, security, predictability, governance, and user trust, that determine whether an agentic approach will succeed in the real world.

Attendees will leave with a usable mental model for designing agentic AI solutions that are purposeful, grounded, and production-ready.