Our Technology

Our AI models offer a better way to measure credit risk

  • Refined over 9 years
  • Trained on more than 121 million data points
  • Drawing on more than 30,000 predictive variables
Bar graph showing our predictive power increasing as our models and data become better over time, from years 2015 to 2021. Starting from an Industry Benchmark (footnote 1) in 2015, our predictive power has increased by 2.5 times the industry benchmark as of 2021.
Man smiling at his phone, surrounded by illustrations of the loan process: a prequalification letter, a computer, analytics, approval, and disbursal of funds.

Our unique approach sets us apart from other FinTech platforms

  • Uses next-generation underwriting, a fully automated approach with data- and optimization-driven characteristics
  • Leverages knowledge of consumer behavior across the full credit spectrum, enabling us to effectively price, manage, and monitor risk at a higher level
  • Incorporates risk, prescreen, conversion, fraud, and lifetime profitability models

Our platform is highly scalable

  • AI-driven underwriting engines enable highly automated decisions
  • Artificial intelligence performance increases with scale
  • Our 100% digital lending platform is engineered to be deployed into new applications and asset classes
  • 91% of the loans are fully automated
A combination bar and line graph showing cumulative loan originations and number of loans between June 2015 and June 2022. The bar graph shows growth from near zero to 1.2 billion in cumulative loan originations. The line graph shows growth from 17 to 304,016 loans.
An illustration of the cycle of Lean Six Sigma methodology: Define, Measure, Analyze, Control, and Improve.

Our culture of continuous innovation transcends the business

  • We operate under a Lean Six Sigma methodology
  • We leverage data science, artificial intelligence and machine learning not just to assess credit and fraud risk...
  • ...but also for marketing, prepayment, conversion, compliance and customer service

Get in touch

1 AI models’ predictive performance measured using the Kolmogorov-Smirnov (KS) statistic on the ADF portfolio.