Predictive intelligence Unable to predict.
Options
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
02-12-2026 11:15 PM
I have trained a predictive intelligence model in "Similarity" in tested that it was working fine but now it is showing "unable to predict"
Can someone tell the possible cause and its resolution.
6 REPLIES 6
Options
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
02-14-2026 02:08 AM
Unable to predict” means the model executed successfully, but ServiceNow could not find a valid similarity match due to input, threshold, or data mismatch — I do NOT think ITS not a platform error.
Retrain the model if data distribution has changed (data drift)
Review and lower the Similarity Score Threshold if needed as If no historical record exceeds the configured similarity threshold, the model returns Unable to predict. Adjusting the threshold improves coverage when data patterns drift over time.
Also refer this KB- Error "unable to predict" due to Business Rule - Support and Troubleshooting
Hope This Helps.
Regards
RP
Options
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
02-15-2026 10:24 PM
I have tried retraining and decreasing the threshold to 10, is there any other way + I also wanted to know that I am triggering this model using script include and it is giving the result only if the user has admin role else not, so is there any specific role for that.
