British multi-national investment bank and financial services holding company, serving more than 40 million customers through global businesses, including Wealth and Personal Banking, Commercial Banking, and Global Banking & Markets.
As a large Financial Services institution providing services in over 60 countries and territories around the world, the client needed to architect an online credit origination system to process requests in real-time. The existing system was cumbersome, sometimes taking up to 6 months to update and deploy new decisioning models. In addition, the customer experience had to support an on-line expectation – providing a decision in seconds.
Complicating the situation:
- Historical customer data needed to be processed on a daily basis. With data sizes growing over time, the processing was taking too long for the daily refresh.
- As part of the decision process, data from external Credit Bureaus have to be fetched and merged with the customer data.
- The volume of requests could vary from day-to-day, up to thousands of requests per second. This made planning and investing in hardware a difficult and costly challenge.
The client needed a way to streamline the TCO of their credit origination system, while simultaneously providing the seamless online experience that their customers expected.
With SAS® Intelligent Decisioning, Core Compete enabled the client to provide digital, seamless, frictionless customer journeys that will radically improve the time it takes to approve customer credit applications for credit cards, loans, and mortgages. The underlying SAS Viya-enabled “Decisioning Fabric” provides a heterogeneous set of tools to support the delivery of these customer journeys.
Core Compete complemented SAS’ out-of-the-box solution for an end-to-end robust, flexible and elastic decisioning ecosystem:
- Built data pipelines to integrate querying and scoring of mainframe data with the decision flows.
- Supported the ability for Data scientists to leverage their favourite algorithms such as Gradient Boosting or Random Forest to achieve model uplift, write their models in their favoured language (R, Python, SAS) then deploy into production seamlessly
- Integrated the ability to capture log decision outcomes for auditing and governance
- Built simulator to replay decision responses and test potential changes to the decision flows before deploying changes into Production
- Architected the SAS decision execution to leverage load-balancing across a highly available architecture.
The client has started rolling out the decisioning framework across their global network, starting in UK and Europe. Leveraging in-memory analytics in conjunction with SAS Micro Analytic Service (MAS) – both part of SAS Intelligent Decisioning – allows the client to provide an optimal customer experience for both existing customers and new customer requests.
- Core Compete enabled the client to reduce their previous 6-month time-to-deployment to minutes for analytical models.
- Based on actual results, the client expects to see at least a
10x return on investment
The client is in process of extending the infrastructure to Google Cloud Platform (GCP) to leverage auto-scaling to support future growth. In addition, they plan additional digital transformation initiatives around Open Source and Machine Learning for the credit lending business.