Objective

Run the real QA evaluator on one Zendesk ticket and inspect the result before relying on automated coverage.

Before you begin

You need Quality: Manage, an enabled category with a saved QA policy, and a connected published scorecard. The Enable QA evaluations switch may stay off for a single-ticket test. Have a numeric Zendesk ticket ID ready.

Resolution

  1. Open the test. Go to Categories, open the intended category, select QA Analysis, and choose Grade single ticket. You can also select Test from QA Analysis · QA Grades.
  2. Choose the category and ticket. If you opened the test from QA Grades, choose the category first. Enter the numeric Zendesk ticket ID and review the assigned scorecard and published version shown in the drawer.
    The Test automated QA drawer accepts a Zendesk ticket ID and shows the selected category’s published scorecard version before the evaluation starts.
  3. Start the evaluation. Select Evaluate ticket. This enqueues the real evaluator against the published version. It creates a retained, unpublished AI draft under the normal retention policy; it is not a disposable simulation. No Zendesk field, comment, tag, or status is changed.
  4. Watch the run. The drawer shows progress and then the result, skip, or failure. If it is still running after the watch period, use Check status or open the run in QA Grades.
  5. Inspect the result. Check the score, any critical failure, criterion grades, and cited evidence. A test does not publish feedback to an agent. Use How To Read An Evaluation And Its Evidence for the review sequence.

A ticket with no category or a different category can still be tested using the category you selected; the drawer warns you and uses that category's scorecard. This does not change automatic eligibility or the ticket's classification. If the test is refused or skipped, see Investigate A Skipped Or Failed Evaluation.