AI Dictionary › Prompting
Catena di Ruoli nel Prompt
Role-chain prompting is a technique in which the task is divided into successive phases, each entrusted to a different role adopted by the model, so that the output of one role becomes the input of the next, simulating a workflow with multiple specialized figures intervening in sequence.
Each phase of the prompt assigns the model a specific role with a narrow task: a first pass might ask the model to take on the role of an analyst gathering and organizing information, a second pass has it act as an author turning that information into text, a third has it act as a reviewer checking the final result. Each role works only on the previous role's output, keeping the overall task broken down into distinct responsibilities.
This technique is useful for replicating articulated editorial or production processes within an automated flow, where separating responsibilities by role helps keep the quality of each phase high: drafting content that moves from a research phase, to a writing phase, to a review phase, each conducted by the model with a dedicated role instruction.
The term combines two techniques already known in prompting, role prompting and prompt chaining, describing the specific case where the chain of successive steps is organized around an explicit role change at each phase, rather than around a simple sequence of sub-tasks.
Grace offers scenarios where you have to pass a task through several roles in sequence, checking that each step correctly receives the output of the previous one.
From our network
HSE Genius: AI for Safety Data Sheets
Extract SDS data, H phrases and ECHA compliance checks in seconds, powered by AI.
Visit hsegenius.com →From the Agora Intelligence blog
📱 Download the Android app (beta) iOS coming soon
Say what you mean. Get what you need.
Grace Certified, the AI coach that trains and certifies your prompt engineering, by Agora Intelligence.