Recommend up-sell/cross-sell opportunities
Use Case (Fraud, AML)
A large bank was in need of a better transaction scoring system that will filter in real-time whether a wire-transfer is permissible or needs further scrutiny. Suspicious transactions require a SAR report to be filed. Existing systems had various business rules that are based on decades old business rules which determine whether a transaction is below a risk threshold or not. activities have new crime patterns that are not yet part of banking systems.
Missing labeled data is a huge challenge in use cases such as this. In order to circumvent thisproblem, we had to use various novel unsupervised technique that will work even when clearly labeled data is missing.
Use Case (Cross-sell/Up-sell)
A regional bank asked for our help to increase their up-sell and cross-sell opportunity. Today they provide many services and products to their millions of customers. That includes home mortgage loans, auto loans, savings and checking accounts, safety deposit boxes, lines of credit, fixed deposits, and commercial loans. Their customer segment is highly fragmented and they are not sure if all their customers are fully utilizing this bank’s services. activities have new crime patterns that are not yet part of banking systems.
This opportunity in this case is quite interesting. By finding utilization patterns of their existing customers and through combining this knowledge to market data and trends in the banking domain, we were able to recommendation products that are highly suitable to some of the customers but are under-utilized. This has not only allowed this bank to increase their cross-sell and up-sell opportunities but also started engaging more with their customer base than they did ever before.