GCP Generative AI Leader Practice Question 88: React Prompting Technique
Google Cloud Generative AI Leader · Test 9 · question 8 of 10
A developer wants a prompting technique in which the model writes out an explicit reasoning step, takes an action such as calling a tool, observes the result of that action, and then repeats the cycle until the multi-step task is solved, reducing the chance the model invents an answer instead of checking one. Which prompting technique does this describe?
Show answer and explanation
Answer:
ReAct (Reason and Act) prompting structures the model's output as a loop of thought, action, and observation: the model reasons about what it needs, takes an action such as a tool call, reads back what happened, and reasons again before its next step. That loop is what lets it check results instead of guessing, which matches the scenario exactly. Option C names the technique.
- A is wrong because zero-shot prompting gives the task alone with no reasoning or tool-use structure at all, the opposite of an explicit reason-act-observe loop. - B is wrong because chain-of-thought prompting only asks the model to reason in text before answering; it does not call tools or observe real results partway through. - D is wrong because role prompting changes the voice or expertise the model answers in; it says nothing about reasoning, tool calls, or observing outcomes.
Reference: Core concepts of AI agents