Human-in-the-loop
Agentic PIM glossary
Human-in-the-loopA way of working in which a person reviews, corrects or approves the decisions of an AI system at defined points in the process, before its results take effect.
In an agentic PIM, human-in-the-loop means agents propose and people approve what matters. Not everything needs the same level of control: you can require approval for descriptions in a sensitive product family and let translations of products already reviewed in the source language go through automatically. The key is that it is configurable and that every change is reversible.
Related