AML case management is the software in which an investigator works through each alert raised by AML transaction monitoring, decides whether it is suspicious, and records the decision. Each alert becomes a case. AML stands for anti-money laundering.
A bank may raise thousands of alerts a day, and most of them turn out to be harmless. For each one, the investigator must find out who the customer is, what is normal for them, and where the money came from and went. The investigator also checks the KYC record, the customer's earlier transactions and any linked accounts. Then the investigator writes down the reasons for the decision, so that an auditor can check it later. Doing all this by hand can take hours for one case.
Now an AI model does the first part. It gathers the history, finds the linked accounts and writes the first draft of the case. The investigator reads the draft, checks the evidence and makes the decision. The model drafts; the person approves. This is called human in the loop. Security teams use the same idea in an AI SOC analyst, which sorts computer security alerts. Sorting many alerts into the few that matter is called alert triage.
If the investigator decides that the transfer is suspicious, the case software prepares a suspicious transaction report (STR). It is sent to the Financial Intelligence Unit, India (FIU-IND), through regulatory reporting, on FIU-IND's online portal, FINGate. The bank must file it within 7 working days after its principal officer, the senior employee responsible for these reports, confirms the suspicion. If the alert is harmless, the investigator closes it with a reason. In both cases, the record is kept for years.