Summary  
This chapter explains how to expose external functions as “tools” that a language model can invoke on demand and how to build an “agent” that autonomously selects and sequences these tools to achieve a specified goal.  

General domain of usage  
research automation

Two words you'll keep hearing if you spend any time around AI: **tools** and **agents**. They sound vague – they're not.



## Why a Model Alone Hits a Wall

A language model only knows what it read during training – frozen in time, locked inside the model. Ask Claude **"what's the weather in Berlin right now?"** and it can't answer. It has no way to reach out.



## A Tool Is Something Claude Can Call

A **tool** is a function wired up so Claude can use it on demand:

| Tool | What It Does |
|---|---|
| `web_search` | Fetch fresh results from the internet |
| `read_spreadsheet` | Pull rows from a CSV or Google Sheet |
| `send_email` | Draft and dispatch a message |
| `read_notion_page` | Load a doc by URL |

Claude doesn't run the tool itself – it writes a request, the tool runs outside the model, the result comes back as text. That round-trip is the **tool call**.



## An Agent Is Claude Using Tools to Reach a Goal

A single tool call is just a lookup. An **agent** is Claude given a goal and a set of tools, free to choose which to use and in what order.

A research agent asked to **"draft a competitor brief on Notion's pricing"** might run two searches, load your brief template, then synthesize a draft – all without you specifying those steps.

**A tool is a capability. An agent is Claude with a goal, a toolbox, and the freedom to decide.**

A hands-on, beginner-friendly course designed for non-technical business users to understand and leverage AI tools and concepts in their daily work. The course covers foundational AI concepts, effective communication with AI, practical use of AI-powered tools, and building simple AI assistants.

Introduction to artificial intelligence, large language models, and essential AI vocabulary for business users.

Learn how to communicate effectively with AI systems through prompt engineering and tone selection.

Explore how AI can be integrated with business tools for research, note-taking, and data analysis.

Learn the basics of agent personas and how to build and tune simple AI assistants for work.

Tool	What It Does
`web_search`	Fetch fresh results from the internet
`read_spreadsheet`	Pull rows from a CSV or Google Sheet
`send_email`	Draft and dispatch a message
`read_notion_page`	Load a doc by URL

What Is a Tool and What Is an Agent?

Why a Model Alone Hits a Wall

A Tool Is Something Claude Can Call

An Agent Is Claude Using Tools to Reach a Goal