Session 2 is where things get hands-on. We're going beyond concepts and getting into the actual mechanics of how agents connect to your enterprise systems — tools, integrations, guardrails, and all. Grab a coffee, this one's packed. ☕

Inside this session

1. Recap and outcomes Review Session 1 and align on today's focus — giving agents controlled, safe access to enterprise systems

2. Tool calling fundamentals Explain tools, function schemas, inputs and outputs, and why structured tool design is the backbone of reliable agentic AI

3. MCP and integration patterns Introduce MCP as a practical integration pattern and compare it with APIs, orchestrators, and integration services already familiar to RPA teams

4. Reusing enterprise assets How existing RPA bots, queues, APIs, databases, and reusable components can be exposed as agent tools — rather than rebuilt from scratch

5. Control and guardrails Validation rules, approval checkpoints, exception handling, retries, and blast-radius reduction for safe agentic execution in enterprise environments

6. Demo: agent in action Walk through an agent that reads context, calls a knowledge source, updates a system, and escalates uncertain cases to a person — live

7. Design discussion Participants list the first 3 to 5 tools an enterprise agent in their environment would need — and what safeguards each one requires

8. Wrap-up Summarize patterns and set up the transition into multi-agent orchestration in Session 3

And before you go — the design discussion at the end is where it gets really personal.

Come thinking about your own enterprise environment, because you'll be mapping out real tools and real safeguards for your own setup.

See you there, super sooonnnn! 🚀