Prompt chaining
Also called chained prompts, prompt chain, fixed pipeline
Splitting a task into a fixed series of model calls, each with its own prompt, where one step's output feeds the next.
Some jobs are better as several small prompts than one big one. A chain runs them in a set order, and each step gets the previous step's output plus its own instructions. Because every step is its own call, you can see where things went wrong and fix that step alone.
Current models plan many multi-step tasks well without a chain. Chaining still earns its place when the structure must never change or when you need to check each step's output before the next one runs.
For example
One call drafts a release note, a second checks the draft against a checklist, and a third revises it. You can read what each step produced.
Put it to use
Where you'll meet prompt chaining in ChatGPT, Gemini and Claude, how to use it there, and the mix-up most people make.
Daily Bytes Plus is in the Android app.