Case Investigation
Cases are generated when a transaction flags one or more rules, or when no rule is flagged but the transaction's AI risk score falls within an enabled Auto Case Creation range. The FraudShield case management system is accessed from the Cases page and is designed to adapt to your specific business process.
Investigation Process​
Every organisation using FraudShield will have differing case investigation processes. In general, case investigation will follow this basic sequence:
1. Assign Case​
Cases are created automatically with Case Status: Open and are listed in order of Created Date - use the filters at the top of the Cases page to narrow by Risk Level, Assignee, Case Status, and more.
By default, cases are not assigned to anyone on creation. This can be changed using Automatic Case Assignment.
You may have a "triage" process for the order in which cases are processed, such as starting with the highest risk cases first, prioritising cases that broke certain rules for reporting purposes, or even a first-in/first-out priority.
Cases can be assigned by any Case Manager role. You can have a manager in charge of delegating cases to the rest of the team, or allow case managers to assign cases to themselves.
2. Investigate Case​
Once the case is assigned, the investigation begins. Use the following information to guide the process:
-
Case Information - Use the Transaction information and review the involved Accounts to get a basis for what happened in the transaction itself. Check the Notes section to see if any other case managers have left notes if applicable, or if any evidence has been submitted.
-
Case History - Review the Case History section to understand the history of the case. This notes any changes since its creation, including changes to the Risk Level, Status, or Assignee. You can also leave notes here for processes occurring outside FraudShield, e.g. ID documents being verified, background checks being performed.
-
Processed Case Rules - Review the Processed Case Rules section to see the rules that were processed for the transaction. Of most interest will be the rules that were flagged, which dictate the fraud and risk implications for the transaction, e.g. Transaction above reporting threshold or Structuring under reporting threshold. You can easily view the rule details from this screen to refresh on the more complex rules.
The rules that were assessed but not flagged are also visible, allowing you to potentially skip parts of the investigation process where they aren't relevant.
-
AI Risk Assessment - If the transaction was scored, the AI Risk score and the Top Rules by Impact show which rules had the most effect on the model's assessment. For a case created from the AI risk score, no rule was flagged, so these rules are the natural place to start. See How AI Risk Scoring Works for how to read them.
3. Resolve Case​
After the investigation is finished, case resolution could take a few forms:
- Investigation Complete - You have determined the investigation is complete and no further steps are needed.
- Escalation - The case may need to be escalated to a more senior case manager for further steps to take place.
- Regulatory Submission - The case may need to be reported to a regulatory body for legislative reasons. See the Regulatory Reports guide for more information.
- False Positive - You may find that the case or rule involved is a false positive, and begin a review process to update the rule plan.
Once the investigation is concluded, set the case status to Closed. If Closure Review is enabled and the case is in scope, select Proposed Closure instead - a reviewer then approves or rejects the closure.
View the Rules guide for an in-depth look at using the FraudShield rule builder and rule plans to implement your monitoring scenarios.