Definition
ReAct is an agent architecture pattern where the AI alternates between reasoning (thinking about what to do) and acting (executing tools or actions). In each step, the agent generates a thought explaining its reasoning, takes an action, observes the result, and then reasons about the next step.
Example
Thought: I need to find the customer's order status. Action: Search orders for customer ID 123. Observation: Order 456 shipped on April 5. Thought: I have the info. I can respond now.
Why it matters
ReAct creates transparent, interpretable decision-making. You can see exactly why the agent made each choice, which is critical for debugging and trust.