Lesson 374 of 1570
What Tools Agents Can Use
Modern agents can use tools — like a browser, an email client, a calculator, a calendar.
Lesson map
What this lesson covers
Learning path
The main moves in order
- 1What Tools Agents Can Use
- 2How AI Agents Use 'Tools' Like a Browser or Terminal
- 3The big idea
- 4Letting AI Pick the Right Tool from a List Instead of Hardcoding It
Concept cluster
Terms to connect while reading
Section 1
What Tools Agents Can Use
Modern agents can use tools — like a browser, an email client, a calculator, a calendar. They pick the right tool for the job.
An agent that can only chat is limited. An agent that can use tools is powerful — and dangerous if not supervised.
Three powerful tools agents have
- Browser (search and click anything)
- Email (send messages on your behalf)
- Payment APIs (actually buy things)
Key terms in this lesson
The big idea: Agents are only as powerful as the tools they can use.
Section 2
How AI Agents Use 'Tools' Like a Browser or Terminal
Section 3
The big idea
An AI by itself can only output text. 'Tool use' is the bridge that lets that text become an action. The model writes 'call read_file with path=foo.txt' and a wrapper actually reads the file and gives the contents back. That wrapper plus the model is what we mean by 'an agent with tools.'
Some examples
- Claude in Claude Code uses a Bash tool — it writes 'ls -la' and the wrapper runs it for real.
- ChatGPT's data-analysis mode uses a Python tool — your CSV gets analyzed in a sandbox.
- Gemini uses a Search tool to fetch live web results before answering.
- Manus uses a Browser tool that literally opens a Chromium window and clicks buttons.
Try it!
Open Claude.ai or ChatGPT and ask it to 'analyze' a CSV file you upload. Notice when it switches into tool-using mode (it shows 'analyzing' or runs Python).
Section 4
Letting AI Pick the Right Tool from a List Instead of Hardcoding It
Section 5
The big idea
If you hardcode which tool runs, you're not really using an agent. Give Claude a list of tools with clear names and descriptions and let it pick — that's where the leverage is.
Some examples
- Claude given search, fetch, and summarize tools picks search first when the goal mentions 'find'.
- ChatGPT with sendEmail and sendSlack tools picks Slack when the recipient is a coworker name.
- An agent with calculator and code-runner picks the calculator for simple arithmetic and code-runner for loops.
- Claude given multiple database tools picks the read-only one when the user asks 'how many rows?'
Try it!
Define three tools with clear descriptions in your next agent and watch Claude route between them based on the user message.
Key terms in this lesson
End-of-lesson quiz
Check what stuck
15 questions · Score saves to your progress.
Tutor
Curious about “What Tools Agents Can Use”?
Ask anything about this lesson. I’ll answer using just what you’re reading — short, friendly, grounded.
Progress saved locally in this browser. Sign in to sync across devices.
Related lessons
Keep going
Builders · 40 min
MCP — How Agents Connect to Tools
MCP (Model Context Protocol) is a standard way for agents to safely talk to tools.
Creators · 50 min
Tool Use at the API Level: The Primitive
Underneath every agent framework is the same primitive — the model returns a structured tool call, you execute it, you feed the result back. Master this loop and every framework looks familiar.
Creators · 50 min
OpenAI Tool Use: Functions, Web Search, Files, MCP, Shell, and Computer Use
Models get more useful when they can act through tools. Learn the difference between hosted tools, your own functions, and MCP-connected capabilities.