Practical guide
An evidence-first QA checklist for supervisors
Review accuracy, tone, explanation and resolution with the original conversation, rather than relying on an AI score alone.
Before scoring
Confirm the workspace, service round and published standard version. Check which CS owned the round and who intervened or received a handover. Verify that the captured conversation contains enough context to assess the rules fairly.
Check the answer
Was the information correct according to the business rules that applied at that time? Did the agent explain the next action clearly? Did they address the customer’s actual question? Quote the relevant message when recording a finding.
Check the service approach
Look for professional language and an appropriate greeting. Check whether the agent acknowledged the concern and avoided unnecessary repetition. Use concrete evidence; do not infer a person’s intent or personality from one brief message.
Check the ending
Did the conversation require resolution confirmation, and was it actually present? If the customer stopped replying or the capture ended early, record that uncertainty. A missing excerpt is not automatically proof that the agent failed a requirement.
Approve with a clear audit trail
Treat an AI result as provisional. Confirm supported findings, adjust incorrect scores and leave a reason where needed. Mark uncertain items for review or not applicable according to the standard. This checklist is a starting point, not a replacement for your company’s published QA rules.