Reporting on sn_piwb_itsm_conte_dbv database views

[qprinsloo]

I have created 3 models to predict category, configuration item, and Assignment group for incoming Self-Service incidents and are then leveraging flows to take action on incoming incidents. This is all working as intended and focus has now shifted to the analytics part. I'm trying to leverage the ITSM Predictive Intelligence Workbench dashboard. However, none of the sn_piwb_itsm_conte_dbv database views are being populated with data. Do ML solutions need to be integrated before these database views are populated with data and/or is there any way around this? What actually controls how data is populated in these database views? Currently the most accurate data I can get is on the ml_predictor_results_list but this doesn't allow evaluating conditions on the incident records such as only considering certain contact types. Right now short of creating a new database view I'm not able to get the reporting needed.

https://docs.servicenow.com/bundle/vancouver-it-service-management/page/use/dashboards/application-c... 

[ASA5]

There are some data collection jobs to activate : 

https://docs.servicenow.com/bundle/vancouver-now-intelligence/page/use/performance-analytics/task/sc...

[qprinsloo]

Thanks for the reply ASA5. I checked and the scheduled data collection job "[Predictive Intelligence for Incidents] Daily data collection" is active. This however is dependent on data being populated in these (sn_piwb_itsm_conte_dbv) database views. Taking sn_piwb_itsm_conte_dbv_ml_predictor_results as an example, both my ml_predictor_results and ml_solution tables have data. However, the piwb_model does not. So even if the collection job is setup this doesn't create data in the piwb_model table does it? Its active regardless though. 

[Brian Bakker · Accepted solution]

@qprinsloo 

If your table [piwb_model] is empty, then you have not trained the solution via the PIWB (Predictive Intelligence Workbench). Here is the record it created in my table [piwb_model], after I used PIWB to create a new "Use Case" to predict the Incident Assignment Group, then trained the solution and integrated it, so that it starts making predictions using the Business Rule "PIWB Prediction - Incident" -

Here is the Use Case that I created, trained and then integrated -

Once it starts making predictions on this solution after integration, it will populate your ITSM Predictive Intelligence Workbench dashboard with the expected details over time.

Hope this helps get your dashboard to see how well your predictions are working.

Regards,

Brian

[qprinsloo]

Thanks for the reply Brian! So the piwb_model is only leveraged if integrating the model via the Predictive Intelligence Workbench? We don't want to integrate the model since it should only be applied across a subset of our Self-Service incidents which we are controlling with custom flows. Thanks, this helps! Guess we'll need to setup own database views then.