AI Discovery
Last Updated: May 7, 2026
Estimated Reading Time: 5 minutes
Overview
AI Discovery analyzes synced CMDB topology and proposes service blueprints. Use it to accelerate service modeling, discover candidate application services, and prepare reviewed blueprints for CSDM work.
When To Use It
Use AI Discovery when:
- service boundaries are unclear
- application services are not fully modeled
- relationship data suggests natural service groupings
- you want a starting point for CSDM Workbench review
- you need to explore possible service blueprints before formal modeling
Before You Begin
Confirm:
- ServiceNow data has synced successfully.
- The active environment is correct.
- Relationship data is present.
- AI capabilities are configured for your tenant.
AI Discovery works best when CMDB relationships are reasonably complete.
Run Discovery
- Open AI Discovery.
- Confirm the active environment.
- Start a discovery job.
- Wait for generated blueprints.
- Review each blueprint before taking action.
The job may take time depending on CMDB size, relationship density, and configured AI limits.
Review Blueprints
For each blueprint, review:
- proposed service name
- included CIs
- relationship evidence
- confidence or quality indicators
- suggested service boundary
- missing or weak evidence
AI Discovery output is a recommendation, not an approved model. Review the evidence before using a blueprint in CSDM workflows.
Use A Blueprint
After review, you can use a blueprint as input for service modeling.
Typical flow:
- Review the generated blueprint.
- Confirm the proposed CI grouping.
- Load or reference the blueprint in CSDM Workbench when available.
- Adjust the model manually.
- Follow your approval process before write-back.
Improve Results
AI Discovery quality improves when:
- CIs have clear ownership and class data.
- relationships are current and directional.
- irrelevant CI classes are excluded from sync.
- services and applications use consistent naming.
- stale and orphaned CIs are reviewed.