Power Platform Community Forum Thread Details

When running a Custom Model (AML v2) prediction from Customer Insights (CI), the workflow fails during activation, and we don’t see any corresponding AML job in Azure ML Studio.

Intelligence workflow Id 'xxxx5fe3-xxxx-48a9-xxxxx-5120xxxx0224' failed for instance Id 'xxxx5ea9-xxxx-xxxx-9d92-xxxx0d4084e8'.

Reason: An exception was thrown while activating Microsoft.Customer360.Intelligence.Implementation.Processor.IntelligenceAmlProcessor.

We’ve validated output path/datastore/schema and understand that CI should only integrate with published AML batch pipeline endpoints (not Real-time endpoints). However, the workflow still fails before any AML job appears.

What We’ve Already Checked

- Output path matches AML job output exactly (as parameterized in the pipeline endpoint)

- CSV schema aligns with CI output mapping.

- No incremental refresh sources are used by this CI custom model.

- We’re using a published AML pipeline endpoint (not an Online endpoint); inputs are marked Set as pipeline input.

Has anyone else run into this activation error referencing `Microsoft.Customer360.Intelligence.Implementation.Processor.IntelligenceAmlProcessor` when triggering a Custom model (AML v2) from Customer Insights? If you’ve resolved it, I’d appreciate knowing what fixed it (identity/RBAC, region alignment, networking, or pipeline endpoint setup).

Thank you in advance!

Categories:

Building Power Apps

Connector development