Prompt chains

Prompt chains link multiple LLM calls together, using the output of one step as the input for the next to execute complex, multi-stage reasoning.

Prompt chains transform isolated AI queries into structured workflows. By passing variables through a sequence (like a LangChain LLMChain or a SequentialChain), developers can decompose high-level goals into manageable tasks. For example, a three-step chain might first extract key entities from a raw transcript, then generate a summary based on those entities, and finally translate that summary into a specific target language. This modular approach improves accuracy by reducing the cognitive load on a single prompt and allows for precise debugging at every stage of the execution pipeline.

https://python.langchain.com/docs/modules/chains/