Phase 07 · AI agents Core

Tools and function calling

Tool calling lets a model request structured actions that your code executes.

In plain terms

The model does not really fetch weather or query a database by itself. It emits a structured request, and your program decides whether and how to run it.

Why it matters

Tools connect language reasoning to real systems while keeping execution under your control.

How it works

Define tool names, descriptions, and JSON schemas. Validate the arguments, run the function, then send the result back to the model.

When you use it

Use it when the answer requires fresh data, private data, calculations, files, or app actions.

Common mistakes

  • Trusting tool arguments without validation.
  • Giving vague tool descriptions.
  • Returning giant tool results that flood the context window.

Best practices

  • Keep schemas narrow.
  • Return compact results.
  • Require user confirmation for send, delete, buy, or publish actions.

Try it yourself

Give a model a calculator tool and make it solve word problems by calling the tool for arithmetic.

Related topics

What is an agent, really An agent is an LLM wrapped in a loop that can choose actions and use tools. Structured output (JSON mode) Forcing the model to answer in an exact machine-readable shape - usually JSON matching a schema - so code can consume it. Output parsers and structured responses An Output Parser turns model text into the shape your app expects: plain strings, JSON Output, validated objects or Structured Output. Agent executor, ReAct and toolkits An Agent Executor runs an Agent loop: Reasoning chooses an Action, a Tool runs, an Observation returns, and Decision Making selects the next step. Agentic AI goals, tasks and context Agentic AI is AI that works toward a Goal by breaking it into Tasks, choosing Actions, reading Observations and updating State inside an Environment. Prompt injection and AI security Prompt injection is malicious instruction text hidden inside data the model reads.