AI-powered checks

Before your request enters the standard approval process, it can go through two AI-powered validation checks:

  • Customer Duplicate Detection (CDD) identifies potential duplicate customer records.
  • Customer Compliance Check (CCC) identifies potential matches against exclusion or sanctions lists.

If either check finds a match that exceeds the configured threshold, the request doesn't proceed directly to the standard approval process.

You or an administrator can adjust the threshold settings. For example, if your organization finds that the duplicate detection and compliance thresholds are too strict.

See the M3 Core User and Administration Library (Cloud) and select Value+ Content > Administrator > Enterprise Automations > Customer Account Creation.

Duplicate detection

When you submit a customer creation request, the system uses Infor Artificial Intelligence (AI) to perform Customer Duplicate Detection (CDD) to identify existing customers with similar information.

If no match is found, the workflow proceeds to the compliance check.

If the check finds one or more potential duplicates, the request is routed back to you, the requester, for review. You receive a task that includes:

  • The customer information that you submitted, such as the customer name, phone number, and document link
  • The duplicate detection results, including the highest similarity score and the four most similar customers that were found
  • Any additional customer information that you provided
  • A link to CRS610, where you can compare the request with existing customer records

After reviewing the information, you can:

  • Approve the request if you determine that the customer is not a duplicate. The request then moves to the compliance check.
  • Reject the request if you determine that the customer is a duplicate.
  • Provide a reason in the Duplicate Check Override Reason from Requester field. You can explain the decision to reject the customer creation request or to approve it despite the duplicate check results.

If you reject the request, you'll receive a notification that includes:

  • The customer name
  • The duplicate similarity score
  • The similarity threshold that was applied
  • The four most similar customers that were found

Example notification when a request is rejected

Duplicate Similarity Score: 0.87

Similar Customers:
• ACME Corporation (Score: 0.87)
• Acme Corp Ltd (Score: 0.72)
• ACM Industries (Score: 0.45)
• A.C.M.E. Inc (Score: 0.38)

Scores interpretation

The duplicate similarity score indicates how closely the proposed customer matches existing customer records. A higher score means there's a greater chance that the customer already exists.

In this example, the score is 0.87, which suggests a moderate likelihood that the proposed customer is a duplicate. Before you approve the request, compare the proposed customer with the listed customers and verify whether a duplicate record already exists.

Use these guidelines to interpret the score:

  • 0.90-1.00: Very high likelihood of a duplicate. Review the request carefully before proceeding.
  • 0.70-0.89: Moderate likelihood of a duplicate. Verify the customer details manually.
  • 0.50-0.69: Low likelihood of a duplicate. The customer is likely different, but you should still review the details.
  • Below 0.50: Very low likelihood of a duplicate. The customer is unlikely to be a duplicate.

Compliance check

After a request passes duplicate detection, the system performs the Customer Compliance Check (CCC). This check compares the customer against exclusion and sanctions lists to help prevent the creation of restricted customers.

If no match is identified, the request proceeds through the standard approval workflow.

If the compliance check finds a match that exceeds the configured similarity threshold, the request is automatically rejected without requiring manual review or approval. This safeguard helps prevent the creation of customers that may match restricted or sanctioned entities.

To adjust the sensitivity of the compliance screening process, you can modify the Compliance_Similarity_Threshold setting based on your business and compliance requirements.

A lower threshold increases sensitivity, causing the system to reject requests based on partial or coincidental text matches. For example, a shared word or name fragment between the customer and an unrelated entity on an exclusion list may trigger a match. If legitimate customers are being rejected, consider increasing the threshold to reduce false positives while maintaining an appropriate level of compliance screening.

If the request is rejected, you'll receive a notification that includes:

  • The customer name
  • The highest compliance similarity score
  • The compliance similarity threshold that was applied
  • The four closest matches from the exclusion or sanctions lists

Example notification when a request is rejected

The request for Acme Corp was rejected due to a failed compliance check 
(score: 0.92; limit: 0.01).

Compliance Top 4 Customer list:
  ACME Corp Sanctions Entry (Score: 0.92)
  Acme Corporation [Restricted] (Score: 0.78)

Summary of CDD and CCC behavior

This table compares the behavior and outcomes when a potential match is identified:

Aspect Customer Duplicate Detection Customer Compliance Check
Purpose Identifies potential duplicate customers by comparing the request with existing customer records. Screens customer information against compliance, sanctions, and exclusion lists.
When a match is found Routes the request to a human approval step for review. Automatically rejects the request.
Can the decision be overridden? Yes. Requesters can proceed with customer creation if the duplicate is determined to be legitimate. No. The process does not include a human review or override step.
Rejection notification Shows the match score, configured threshold, and up to four similar customer records. Shows the match score, configured threshold, and up to four matching compliance records.
Rationale Similar customer records may represent valid business entities, such as subsidiaries, regional offices, or branches. Compliance and sanctions matches represent legal and regulatory risks that require immediate rejection.