Automate Credit Scoring with Machine Learning
Looking to serve your customers more efficiently? Want to make quick and well-informed decisions on lending, based on multiple data sources? Now you can, with our real-time bidirectional integration capabilities with market-leading decisioning engines.
The example below shows how our integration with the CredoLab AI credit scoring platform uses opt-in smartphone metadata to incorporate character profiling as part of a blended score.
This enables even the most underbanked of customers to potentially get a “yes” when they may not have been eligible using traditional methods.
Potential and existing customers can apply for a loan via USSD; mobile app; web; field agent; or in the branch. As part of this process they will receive an SMS inviting them to download the CredoLab app.
Once they have downloaded the app, they are required to give their permission to share information on their phone with CredoLab. This data is captured only at the time they approve, not on an ongoing basis, so they customer only needs to have Internet access for a brief time. All data is depersonalised and not used for any other purpose.
The relevant data points are sent to the CredoLab decisioning engine, where the AI algorithm determines a “character score”. This is then sent to the Fern system and combined with other data – including customer demographics, loan repayment history and savings behaviour as appropriate – to determine a final blended score and make a decision.
This process all takes place in real-time and so the customer can expect a decision within a minute!