Identify a bad range reference pattern in a row-level AI function workflow.
The selected range is doing too much. Mei is creating an AI function column for renewal-risk labels. Which part of the setup is the real issue? The selected range includes the AI output and review notes Correct. The AI function should read source evidence, not its own generated column or later reviewer notes. This protects independence and auditability. The source facts are included Those columns are the right evidence. The principle is not to avoid ranges; it is to keep ranges limited to the facts the prompt should use. There is a separate human review area A review area is…
Sign up free: one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in