Agents and tools over enterprise systems
On This Lesson 4 topics
This lesson is planned. The outline below lists the topics it will cover. The full lesson has not been published yet.
In this lesson
design tool-using agents with limited authority.
Planned topics
- A tool is a described function the model may request; your code runs it, not the model.
- Give read tools freely; put write tools (refund, address change) behind confirmation or human approval.
- Limit steps, spend, and time per task; log every tool call.
- Each tool returns only needed fields; extra data is extra leakage risk.
Draft outline details
Demonstration
a tool table for the assistant: getOrderStatus (read), getReturnPolicy (read), createReturnRequest (write, needs customer confirm), refund (not exposed; human agent only).
Check your understanding
Who executes a tool call?
Show answer
Your platform code; the model only proposes it.
Practice
Apply this lesson with the module's lab, exercises and quiz.
Marking a lesson complete is optional and you can undo it any time. Your progress stays in this browser.