Auto Publishing an Evaluation from Auto Work Queues
Required Feature Flags
The following feature flags and permissions are required to use this feature:
Feature Flag | Technical Name | Description |
Auto Work Queues |
| Enables Auto Work Queue templates and automated assignment |
Auto-Publish Evaluations |
| Allows fully automated evaluations to be published without human review |
SmartScore |
| Required for automated line item scoring |
Required Permissions:
Manage Work Queue Templates (
admin.assignmentbuilders.configure) — to create and edit Auto Work Queue templatesEvaluate (
quality.evaluations.evaluate) — to act as a fallback evaluator if auto-publish fails
What Is Auto Publishing?
Auto publishing automatically publishes evaluation results to reports when using a fully automated scorecard, without a human evaluator needing to hit Publish.
A system user called "Co-pilot" is used to signify evaluations that have been auto-published.
Go to Evaluate > click Auto Work Queues.
Once you have selected and saved the agents to be evaluated, the next step is to add a contact type and assign evaluators. Select a filter where contacts will be randomly sampled from.
Note: The auto-publish options only apply if you have selected a scorecard with only automated line items. If you select a manual scorecard (one with manual line items or manual evaluation outcomes), the process will continue as normal with a human evaluator scoring the contacts.
If you select an automated scorecard, you will see the following options:
Option 1: Yes, publish without review -- Results from the evaluation will automatically go to reports upon completion.
Option 2: No, require human review -- An evaluator will need to open the evaluation and click the publish button before results appear in reports.
Lastly, you will need to set your target number of evaluations per agent and evaluator. The evaluator assignment is a necessary step because the Co-pilot will attempt to auto-publish these evaluations to reports, but if any errors occur during the automated process, the evaluations will be reassigned to one of the evaluators you nominated.
Auto Publishing Logs
To keep track of events relating to Co-pilot and auto publishing, the following events are logged:
Evaluation auto published: This log records when an evaluation has been published without human review.
Edited evaluation: This log records when a human evaluator edits a published evaluation in the case of a system error.
Safety net evaluation: This log records when an evaluation has been reassigned to an evaluator because some automated line items could not be completed, meaning the evaluation will not be auto-published.
Notes
An evaluation will only auto-publish if all automated line items are completed.
A human evaluator can edit a published evaluation and amend the evaluation to themselves. This will be captured within the audit logs.
If an evaluation fails to auto-publish, it will stay in progress on the contacts to evaluate table for the fallback evaluator to pick up.
Users who have the permission to access these features and to build auto work queues will be notified via in-app notifications if an evaluation fails to auto-publish.