React in practice

You have the ReAct loop running. Now let us see how it handles real problems, understand when ReAct is better than prompt chaining, and practice designing tool-using agents.

Prompt chaining is a fixed pipeline: step 1 always feeds into step 2. ReAct is dynamic: the model decides which tool to use and when to stop based on the results. A chain always runs the same steps; a ReAct agent adapts based on what it observes. Think of chaining as a recipe and ReAct as a chef who improvises.