A tool is a specific capability you give the agent: look up a shipment, send an email, search the knowledge base, reschedule a pickup. The model can't do anything you haven't given it a tool for, which means your tool inventory is the complete definition of what your agent can do.

This is why tool design is product design, and it's the subject of Chapter 3. For now, three things worth knowing.

Fact	What It Means For You
The agent only knows what your description says	A badly described tool gets used wrongly or never
It picks tools from the names and descriptions alone	Two similar-sounding tools will get confused
Every tool you add makes every choice slightly harder	More tools is not strictly better, it can be worse

The Counterintuitive One

Adding a twentieth tool can make an agent worse at using the other nineteen. The model is choosing between options at every step, and more options means more chances to choose wrong. Teams that keep adding capabilities often see reliability drop, then add more capabilities to compensate. Prune your tool list deliberately.

Agentic AI and MCP for Product Managers

Module 1What 'Agentic' Actually Means

Module 2How an Agent Actually Works

Module 3MCP: What It Is and Why You Should Care

Module 4Designing the Tool Layer

Module 5Permissions and the Human in the Loop

Module 6Evaluating Agents

Module 7Cost, Latency and Unpredictability

Module 8Designing What the User Sees

Module 9Risk and Blast Radius

Module 10When an Agent Is the Wrong Answer

Module 11Final Project: Design One Real Agent End to End

Module 12Appendix: The Prompt Library

Tools Are the Interesting Part