Summary
We request that duplicate expenses be marked as anomalous in the Expense Anomaly Detection AI capability, so users can identify potential duplicate expense entries in the FDI front end.
Background
Oracle has clarified that anomaly detection (pattern-based, using historical claims) and duplicate detection (transaction matching) are distinct ML use cases. Fusion Expense offers duplicate detection as a dedicated capability, but only in the form of alerts shown to the user at the point of entry on the source side.
Current Limitations
- Alerts are transient, so finance and audit teams cannot review duplicates after the fact.
- Because duplicates are flagged only through a dismissible alert at entry, users can mistakenly submit them, and duplicate expenses can go unnoticed.
- Duplicates cannot be marked as anomalous in FDI, so they do not appear in anomaly reporting or expense risk analysis.
Requested Enhancement
Enhance the Expense Anomaly Detection model to identify duplicate expense patterns, based on the use case below or any other type of duplicate entry, and mark them as anomalous for reporting in FDI.
Use Case
An employee submits the same receipt twice: once through a corporate card feed and once manually, either because the entry-time alert was dismissed or because it did not trigger. With this enhancement, a business user can see the duplicate pair flagged as anomalous in FDI and then proceed with an audit.
We ask the product team to review and prioritize this request based on broader customer demand. We are happy to join design discussions.
