| Announced on | Rollout starts | Rollout ends |
| May 13, 2025 | May 13, 2025 | September 10, 2025 |
Zendesk Quality assurance (QA) is introducing a new metric called accuracy to better evaluate AutoQA’s performance. This metric measures how closely AutoQA’s predictions align with human assessments and replaces the previous acceptance rate metric.
This announcement includes the following topics:
What is changing?
A new accuracy metric now measures the consistency and agreement between AutoQA-generated ratings and those provided by human reviewers, replacing the previous acceptance rate metric.
A score above 75% indicates strong agreement, demonstrating the effectiveness of AutoQA.
Additionally, a new Ratings by accuracy level bar chart displays the level of agreement between AutoQA and human reviewers:
- High accuracy: Scores between 75% and 100%
- Medium accuracy: Scores between 25% and 74%
- Low accuracy: Scores between 0% and 24%
When accessing the AutoQA dashboard, new customers will see only the new accuracy metric. Existing customers will see both the previous acceptance rate metric and the new accuracy metric until the rollout end date on August 18.
Why is Zendesk making this change?
While the acceptance rate penalized changes from N/A to numeric scores, the new accuracy score focuses solely on numeric scores. This distinction is crucial for understanding the effectiveness of AutoQA.
What do I need to do?
To prepare for the transition from the previous acceptance rate metric to the new accuracy metric, we encourage you to familiarize yourself with its details. See Understanding the AutoQA dashboard in Zendesk QA.
If you have feedback or questions related to this announcement, visit our community forum where we collect and manage customer product feedback. For general assistance with your Zendesk products, contact Zendesk Customer Support.