Interleave reasoning and action steps — the prompt pattern that powers most modern tool-using AI agents.
Five passes over the same idea, each from a different angle. Do them in order, or jump to whichever you need.
ReAct (Reason + Act) prompting alternates between thought steps and action steps. The model reasons about what to do, takes an action (e.g. API call, search), observes the result, reasons again, and repeats until the task is done. This is the foundational pattern behind LangChain agents, AutoGPT, and most tool-using LLM systems.
Where this topic shows up outside its home domain:
