The fix has been fully validated and is being prepared for deployment to all impacted instances. This process is expected to take several hours to complete.

Update timeline

1. identified May 11, 2026, 12:05 PM UTC

   Degradation Started: agentforcePerformanceDegradation

2. monitoring May 11, 2026, 12:05 PM UTC

   A subset of customers are experiencing an issue affecting Agentforce custom prompt templates. Customers may receive the following error message “Can't open prompt template. The URL contains a record ID for a prompt template that doesn't exist. Check the URL and try again”. We're in the process of engaging additional engineering teams to further investigate the issue. We’ll provide an update in 30 minutes or sooner if additional information becomes available.

3. monitoring May 11, 2026, 12:19 PM UTC

   We're currently investigating whether this issue is a result of a recent release. If confirmed to be related, we'll attempt to perform a rollback of the recent release to resolve the issue for affected customers. Attempts to use the "Draft with AI" functionality dependent on custom prompt templates are also failing. Standard prompt templates are not impacted. We’ll provide an update in 30 minutes or sooner if additional information becomes available.

4. monitoring May 11, 2026, 12:57 PM UTC

   We are attempting to perform a roll-back on the release on a test instance. If this is successful, it will be applied to all impacted instances. In the event that the rollback is not successful, we'll attempt an emergency code fix. The development and testing of the fix is being done in parallel and is expected to take 60 minutes to complete. We’ll provide an update in 60 minutes or sooner if additional information becomes available.

5. monitoring May 11, 2026, 01:11 PM UTC

   The roll-back of the recent release on a test instance was unsuccessful. We've identified the need to implement a configuration change to the impacted instances. The configuration change will act as a fix for the issue. Once this configuration change is developed, it'll be validated on a test instance and then rolled out to all impacted instances. We’ll provide an update in 60 minutes or sooner if additional information becomes available.

6. monitoring May 11, 2026, 02:10 PM UTC

   Work is still underway to develop the configuration change needed to resolve the issue for affected customers. This is expected to take some more time to complete. We’ll provide an update in 60 minutes or sooner if additional information becomes available.

7. monitoring May 11, 2026, 02:41 PM UTC

   We’ve completed the development of the configuration change and are currently verifying it on a test instance. If this proves successful, it'll be rolled out to all impacted instances. We’ll provide an update in 60 minutes or sooner if additional information becomes available.

8. monitoring May 11, 2026, 04:45 PM UTC

   We’ve validated the fix, and it’s now progressing through our deployment build pipeline. We expect this process to take approximately three to four hours. Once complete, we'll perform a final round of validation before rolling out the fix to all affected customers. Upon further investigation, we have determined that the start time of the impact is earlier than initially understood. We have now revised the start time of the Trust post to more accurately reflect the time customers may have begun to experience impact. We'll provide an update upon completion of the fix or if significant information becomes available.

9. monitoring May 11, 2026, 09:01 PM UTC

   The fix has been fully validated and is being prepared for deployment to all impacted instances. This process is expected to take several hours to complete. We'll provide another update once the fix has been deployed to all impacted instances, or sooner if additional information becomes available.