Question

A talent acquisition team wants to deploy a Copilot Studio agent that ranks job applicants by summarizing resumes and scoring candidate suitability against role requirements. During the responsible AI review before go-live, the architect must determine which action most directly addresses the Microsoft Responsible AI principle of fairness for this use case. What should the architect prioritize?

Accepted Answer

Conduct a fairness assessment that tests the agent's scoring output across demographic groups to detect and mitigate disparate outcomes before deployment

Answer

Enable data loss prevention (DLP) policies to block the agent from sending candidate data to unapproved connectors

Answer

Add a fallback topic so the agent politely responds when it cannot parse a resume format

Answer

Pin the underlying model version so scoring results remain reproducible across releases