Prompt Chaining
Passing the output of one model call as the input to a subsequent call to complete complex tasks.
Full Definition
Prompt chaining decomposes a complex task into sequential steps, where each step is handled by a separate model call and the output is passed forward as input to the next. This breaks tasks too large or complex for a single prompt into manageable pieces, enables specialisation (each step can use a different system prompt or model), and makes the pipeline easier to debug — you can inspect and correct outputs between steps. Chaining is foundational to agentic workflows and is used in document summarisation pipelines, multi-stage data extraction, and content generation systems.
Examples
Step 1: extract key facts from a long document → Step 2: draft a report from those facts → Step 3: edit the draft for tone.
Step 1: classify a customer email by intent → Step 2: route to the correct response template → Step 3: personalise the template using the customer's name and history.
Apply this in your prompts
PromptIt automatically uses techniques like Prompt Chaining to build better prompts for you.
Related guides