prompt chaining
A workflow that splits a task into fixed steps, each a model call that feeds the next, often with a check in code between them.
In more detail
Each step is simpler and easier to test than one big prompt, and a gate between steps can stop bad output early, at the cost of extra latency. Use it when a task divides cleanly into known steps.
Related terms
- workflow (LLM) — A system in which your code fixes the sequence of model calls and tools, as opposed to an agent that chooses its own steps.
- routing (LLM) — A workflow that classifies each input and sends it to a specialised prompt, tool or model.
Taught in
- Workflows Before Agents — AI Engineering: LLM Apps, RAG & Agents