You are here:
Set Up Configuration Item Criticality Prediction for IT Services
Set up the Configuration Item Criticality Prediction model to forecast each configuration item's business importance by using machine learning, CMDB dependency topology, and incident history. Use the criticality scores to prioritize incidents, improve change risk assessment, and focus resources on the most business-critical assets.
Required Editions
| Available in: Lightning Experience |
| Available in: Enterprise, Performance, and Unlimited Editions with Agentforce IT Service. |
Prerequisites
Complete these prerequisites before training the model. At least 400 active Configuration Item records with business criticality values are required.
- Set up Configuration Management Database (CMDB) and verify these DMOs are available: Configuration Item, Configuration Item Value, Configuration Item Relationship, and Configuration Item Definition.
- Verify these IT Service DMOs are mapped: Incident, Incident Configuration Item, Change Request, and Change Request Configuration Item.
- To create the custom ConfigurationItem_Home DMO, navigate to Data Cloud | Data Streams | ConfigurationItem_Home | Start Mapping.
-
In the Data Model Entities section, click the Edit icon and select the Custom Data Model
tab. Click New Custom Object, review the field list and data types,
then save your changes.
The system automatically generates field mappings between the data stream and the new custom DMO.
Install the Configuration Item Criticality Prediction App
- From Setup, in the Quick Find box, enter templates, and select Templates under Apps | Intelligent Apps.
- In the list of available templates, select the dropdown for Configuration Item Criticality Prediction, and then select Create App.
- In the Data Space Name, select the default data space for the application installation, and click Next.
- In the Training Duration field, enter the number of days of historical data to use for model training (default is 365 days), and click Next.
- In the Select Features page, review and select the textual, numerical, and categorical variables relevant to your organization's data, then click Next.
- In the Enter App Details page, provide an App Name, Description, and Log Level.
-
Click Create to begin the installation process.
The App Install History page appears. The installation includes automated tasks for data transformation, model training, and scoring. Initial criticality scores are saved to the Configuration Item Criticality Scoring DMO.
Create AI Accelerator Use Cases
Create use cases to display CI criticality predictions on Configuration Item, Incident, and Change Request pages. Follow these steps for each record type where you want to display predictions. The configuration is nearly identical across all three record types.
- From Setup, in the Quick Find box, enter AI Accelerator, and select AI Accelerator.
- Click New Use Case and enter the appropriate name for your record type based on the table below.
- Click Edit for the new use case, set Maximum Predictors to the number of top predictors to display (for example, 3), and save your changes.
-
In Use Case Model, select Add Model and configure the fields based on
your record type.
For all record types, set these common fields:
- Prediction Platform: Einstein on Data Cloud
- Predictions: Already generated
- Data Space: default
Configure the remaining fields using this table:
AI Accelerator Configuration by Record Type field Configuration Item Incident Change Request Use Case Name Configuration Item Criticality Score Incident CI Criticality Prediction Change Request CI Criticality Prediction Use Case Model Name Configuration Item Criticality Score Incident CI Criticality Prediction Change Request CI Criticality Prediction Score Field Configuration Item Criticality Scoring > Predicted CI Business Criticality Incident CI Score > Predicted CI Business Criticality Change Request CI Score > Predicted CI Business Criticality Reference Record Field Configuration Item Criticality Scoring > Core CI ID Incident CI Score > Incident ID Change Request CI Score > Change Request ID Add Input Field (optional) Not required Field label: CI_NAME, Source field: Configuration Item Name Field label: CI_NAME, Source field: Configuration Item Name Prediction Label Template The predicted score is %%SCORE%% The predicted score is %%CI_NAME%% The predicted score is %%CI_NAME%% -
Save your changes.
You've enabled the AI Accelerator use case. Repeat these steps for each record type where you want to display criticality predictions.
Add AI Accelerator Cards to Record Pages
Add the AI Accelerator component to Incident and Change Request pages to display real-time criticality predictions. Follow these steps for both page types. The Configuration Item page displays predictions automatically once the use case is configured.
- From the App Launcher, find and select Incidents or Change Request.
- Open a record, click Setup, and select Edit Page.
- In the Lightning App Builder, drag and drop the AI Accelerator Card component onto the page layout.
-
Select the component and configure its properties:
- Title: Enter Incident CI Criticality Score for Incident page or Change Request CI Criticality Score for Change Request page.
- Use Case Definition: Select the use case you created earlier (Incident CI Criticality Prediction or Change Request CI Criticality Prediction).
- Threshold Score: Enter the minimum score required for the prediction to appear (for example, if the minimum score is 5, records with a criticality prediction score below 5 are not displayed).
-
Save your changes and click Activate to publish the page.
The AI Accelerator card now appears on the record page, displaying criticality scores for CIs linked to incidents or change requests. Verify the card appears on an active record that has a linked configuration item.
Did this article solve your issue?
Let us know so we can improve!
