United KingdomFinance and insurance2019 – ongoingSecure Trust Bank PLC
Explainable AI models for credit application scoring
Secure Trust Bank put explainable AI credit scoring live on two portfolios, a first for a UK lender
Problem
Lenders want the predictive gains of machine learning in credit scoring, but regulators and risk teams need to see why a model reaches a decision. Secure Trust Bank wanted better application scoring than its traditional linear models gave.
Approach
The bank built application scoring models with Archetype, explainable AI software from Jaywing. The models went live on two of its portfolios, and the report called Secure Trust the first UK lender to do so. Jaywing signed a three-year engagement to monitor and recalibrate the models.
Results
No published figures
Takeaway
The bank reported a better-performing model than its linear baseline without publishing the size of the gain. The bad-debt reduction was a projection at the time of the report.
Categories
- Industry
- Finance and insurance
- Business function
- Finance, risk and legal
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Vendor solution
- Goal
- Quality and risk
- Evidence
- Vendor case study
Sources
Source summaries
The article reports that Secure Trust Bank deployed credit application models built with Jaywing's Archetype on two portfolios. It says the models gave a significant uplift in predictive power over a linear model, with no size given. Bad-debt reductions were described as projected only, and Jaywing will monitor the models for three years.
Full details are in the original sources.