How AI scores every conversation

Every closed case, email, chat, call and AI-agent conversation is scored automatically, with evidence quotes and confidence.

For: Supervisors, QA leads and agents

Steps

When a case is resolved or closed, an email or chat is closed, or a call ends, MetriQA scores the conversation a few minutes later.
Calls are scored from their transcript, so they are scored once the transcript is ready.
Go to MetriQA → Reviews to see every conversation with its score. Filter by below pass mark, auto-fail, risky or not scored.
Open a conversation to see the transcript with the AI's evidence highlighted, and each criterion's score, confidence and reason.
Risky conversations (escalation language, legal threats, churn signals) are flagged with the customer's own words.
Admins can select Score now to score anything waiting; reviewers can Re-score a conversation after a scorecard change.

Agents see the scores of their own conversations; supervisors see their teams'.
Evidence quotes are checked against the conversation, so every highlighted quote really was said.
Scoring uses the workspace's AI credits and its AI data policy (redaction) like every AI feature.

Still need help? Contact support