Decision-Making Software Replacement In A Full Service Bank

About our client

Our client—further referred to as the Bank—is a large full-service European bank. It provides a wide range of products and services to retail and corporate clients.

Project background

Our client discovered that their existing decision-making module (DMM) had a number of drawbacks. They made modifications and the support of loan decision-making strategies complicated. (For further details see “Troubleshooting the Bank’s Issues”).

It was necessary to replace the decision-making system because its drawbacks significantly decreased the Bank’s competitiveness.

The Bank selected solutions from three vendors—SAS, Experian, and FICO—leaders in the decision-making software market. Due to multiple successful project implementations, ABLE Platform was recommended by FICO as a participant in a close bid to represent its products.

The Bank’s experts carefully scrutinized tender participants’ history to choose the best one for the project. In addition, all participants had to provide relevant references from banks with similar implementation cases.

ABLE Platform’s team used a seeing is believing approach to impress the bank with our competitiveness. The Bank’s representatives were invited to our previous client with the same BRMS implemented by our team.

Seeing the positive results of FICO® Blaze Advisor deployment in action, the Banks’ management decided to implement an analogous module of the decision-making system.

Troubleshooting the Bank’s issues

A thorough examination of the Bank’s existing decision-making module discovered numerous flaws. They included:

1. Impossibility of changing strategies without the engagement of the IT department.
2. The long process of deployment of loan strategies and connection of data sources.
3. Inconvenient interface for configuring strategies.
4. Absence of a fully functional logging process.
5. Lack of technical opportunity to make requests to external services and internal databases via the decision-making module.
6. Technical problems with machine learning module implementation.
7. Performance limitations.
8. Absence of modern software version support.

Solution

Together with the main decision-making module—FICO® Blaze Advisor—we additionally offered FICO® Application Studio (FASt) for building an integration layer and application routing. ABLE Platform’s team also offered the component of our software solution ProcessMix—Decision Module Companion (DMC) as a tool for testing strategies.

FICO® Blaze Advisor didn’t need to be advertised to the Bank. It was known to the Bank’s risk management team as a powerful tool for the development and execution of decision-making strategies. Its deployment allowed ABLE Platform to fulfill the following requirements of the Bank:

• Creation of stable and effective decision-making strategies.
• Reduction of time for changing existing loan strategies and deployment of new ones.
• Optimization of the releasing process of new versions of strategies.

FICO® Application Studio (FASt) was deployed into the Bank’s IT infrastructure to perform the following functions:

• Integration with Bank’s existing internal systems and databases.
• Integration with external services (more than a dozen).
• API building for connection with a decision-making module.
• Effective logging and administering processes.

DMC was not on the to-do list. But our experts showed the module in action as an efficient and simple strategy testing environment. This impressed the Bank’s management greatly. They decided to add DMC as the desired option as well.

Decision-making renovation stages

There were 4 main stages to the Bank’s system renovation:

1. Preparation:
   • license key purchasing
   • deployment of the development environment
   • deployment of the testing environment
   • collection and formalization of the requirements
2. Development of strategies:
   • development of scorecards
   • creation of risk rules
   • calculation of income and loan offers for clients
   • calculation of values to comply with regulatory requirements
3. Integration:
   • API building
   • integration with internal and external services (more than a dozen)
4. Development of effective logging and administering the process of the solution.

Result

Decision-making module replacement allowed ABLE Platform to fulfill all the project’s goals. The load test of the system displayed the following performance figures:

1. The renewed decision-making module allows risk managers to change loan strategies by themselves in a short period of time.
2. Up to 20,000 requests are processed hourly.
3. Up to 5,000 loan applications are processed in less than an hour.
4. Response time:
   • Less than a minute with access to external sources under heavy duty.
   • Less than 20 seconds without accessing external services under heavy duty.

The Bank’s team was satisfied both with the implemented product and cooperation with ABLE Platform’s team.