Quality Assurance & Calibration

Last updated 2026-10-24 · fulliautomatix 1.0.33 · getafix 3.2.65

When Quality Enabled and Record Enabled are turned on for your company, the Desk's Quality module (https://desk.<domain>/quality) automates call sampling, structured scorecard grading, agent coaching feedback, and multi-evaluator calibration.

Setting Up Scorecards & Sampling Rules: Scorecards and Sampling Rules are created by a Company Administrator in the Admin Console under Settings → Scorecards and Settings → Sampling Rules (see Alert Rules, QA Scorecards, Sampling & Shifts). Once configured, a background sweep runs every 30 seconds and automatically populates the supervisor evaluation queue.

1. The Evaluation Queue (/quality → Queue)

• The Quality badge in the left navigation rail displays your count of pending evaluations.
• The Queue tab lists pending evaluations ordered with the soonest due date first, showing the agent's name, queue, call timestamp, and due date.
• Click Evaluate on any row to open the grading workspace.

2. Grading a Call in the Workspace

Inside the evaluation workspace:

1. Audio Player & Timeline Markers:
   • Play back the call recording directly in your browser.
   • Any live QA tags saved during live monitoring (Compliance miss, Coaching needed) and any masked Secure Segments are highlighted directly on the audio timeline.
2. Scoring Sections & Items:
   • Each scorecard is organized into weighted sections (for example, Opening weight 1, Resolution weight 3).
   • Mark each item Yes, No, or N/A (items marked N/A are excluded from the denominator so they do not penalize the agent) and add optional per-item comments.
   • Auto-Fail Items: If an item flagged as Auto-fail (such as Verified the caller) is marked No, an auto-fail warning banner appears and the total evaluation score immediately drops to 0%.
   • Your draft answers are saved automatically as you click, or you can click Save to keep a draft and return later.
3. Submitting & Coaching Delivery:
   • Add an overall Evaluation comment with coaching notes for the agent.
   • Click Submit evaluation:
     • The scorecard snapshot is frozen and the final weighted score (0–100%) is recorded.
     • Your pending Quality rail badge decrements immediately.
     • The evaluation moves to the Done tab with its score chip.
     • The agent receives a coaching email and sees the score and your feedback on their Desk's Coaching card.

3. Running a Multi-Evaluator Calibration (/quality → Calibration)

Calibration sessions ensure that multiple supervisors grade calls consistently against the same scorecard:

1. Open the Calibration tab and click New calibration.
2. Select a recorded call and tick two or more evaluators (supervisors), then click Start calibration.
3. Independent Grading:
   • Each participating supervisor opens the calibration session and scores only their own copy (attempting to edit another evaluator's copy is refused).
4. Side-by-Side Comparison & Agreed Score:
   • Once evaluators submit their copies, the calibration view compares all scores item by item, automatically flagging any items where evaluators disagreed and showing the score spread.
   • After discussing the discrepancies, enter the final Agreed score and a calibration note, then click Record agreed score to close the session.