AI Agents with Java

Chapter 3. Tool Integration

A Note for Early Release Readers

With Early Release ebooks, you get books in their earliest form—the author’s raw and unedited content as they write—so you can take advantage of these technologies long before the official release of these titles.

This will be the 9th chapter of the final book. Please note that the GitHub repo will be made active later on.

If you’d like to be actively involved in reviewing and commenting on this draft, please reach out to the editor at [e-mail].

The previous chapter introduced the RAG workflow, a technique to add context to an LLM to produce better answers to user questions.

RAG works in cases where we need to provide information to LLMs from documents, and when most of the queries use the RAG context, but it hits the wall in the following cases:

Real-time data

Current weather, stock prices, or exchange rates.

Dynamic Actions

Booking a flight, sending an email, or querying a live database.

On-demand Context

Information that shouldn’t be in the prompt unless specifically requested.

In this chapter, we’ll learn how to fix these problems by using function calling/tools to let LLMs invoke methods defined on the application side.

Function Calling and Tool Binding

As noted in the previous chapter, models don’t always have the requested information, either because you didn’t train them with the data or because the information required needs to be available in real time.

To avoid these problems, most models ...