Should every interaction be scored?
That depends on your systems, risk and purpose. Define a sampling approach that gives useful coverage and separately identify targeted investigations. A large review count does not compensate for unclear criteria or conclusions that ignore which interactions were selected.
What is calibration?
Reviewers independently assess the same interaction using the same criteria, then compare their reasoning and resolve differences. It helps expose ambiguous questions and inconsistent interpretations. Amazon Connect’s calibration workflow is one example of a tool supporting this process. Amazon Connect Calibration Sessions
Is a quality score the same as customer satisfaction?
No. A quality score reflects your evaluation criteria, while satisfaction captures the customer’s reported experience. Read both with case context. A correctly handled interaction can still leave a customer unhappy about a policy, and a friendly answer can still be wrong.
Can they coach staff directly?
Yes, where that responsibility is included and the person has suitable coaching skills. Agree how feedback is shared and how disagreements are reviewed. Team leaders should remain involved so coaching aligns with the person’s responsibilities and available support.
How do we know quality work is helping?
Check whether recurring errors reduce and whether customers receive more accurate, complete help. Combine evaluation evidence with service outcomes and feedback rather than relying on average scores alone. GOV.UK’s service guidance similarly recommends feedback across the full experience when assessing service improvement. GOV.UK Measuring User Satisfaction