Data in
Core banking adapter
Accounts and transactions, nightly
← Solutions Lab
Case study · Bfsi
Mule detection for an urban cooperative bank
Client
An urban cooperative bank in South India
Industry
BFSI
Solution families
Cybersecurity and Digital Trust · Enterprise Integration and Automation
Time to proof
5 weeks
Executive summary
Fraud proceeds were moving through a small number of newly opened accounts, and the bank had no data team to spot the pattern. QAO MuleShield Connect read the bank's core banking data every night, scored accounts for mule risk and gave investigators one clear picture of linked accounts and money flows.
Key outcomes
38
Mule accounts flagged in the first nightly run
85%
Reduction in time to assemble an investigation case
100%
Account actions approved by an officer
Solution and architecture
What we composed.
A core banking adapter extracts accounts and transactions nightly. The feature pipeline prepares mule-risk signals, a scoring connector returns a risk score for every account, and investigators see linked accounts and fund flows in one workbench. Every hold or freeze is approved by an officer and recorded for reporting.
Intelligence
Feature pipeline and scoring
Mule-risk score per account
Workflow
Investigation workbench
Linked accounts and money flows
Systems of record
Reporting pack
Actions and reports for authorities
Solutions Intelligence, applied.
Level 1
Sense
Mapped how fraud surfaced and how long cases took to build.
Level 2
Reason
Agreed which signals justify review and who approves action.
Level 3
Compose
Designed the adapter, pipeline and workbench around the existing core.
Level 4
Prove
Ran on historical data with a planted mule ring against agreed targets.
Level 5
Evolve
Added nightly runs, new signals and reporting formats.
Technology
PythonGraph analyticsPostgresScoring API connectorReactRole-based access