James, formerly known as CrowdProcess, is developing the artificial intelligence (AI) platform for credit risk.
The system was designed by and for credit experts with the goal of making sure that risk teams have continuous access to state-of-the-art data science techniques and up-to-date validation and monitoring requirements, without having to setup complex software solutions.
James has been adopted by financial institutions from two different continents and has been put up against incumbent solutions by more than 20 lenders, ranging from Tier 1 banks to alternative lenders, with positive results.
Credit risk management platform for financial applications
James is an integrated platform for credit risk management that allows financial organizations to easily create, validate, deploy, and monitor regulation-ready, high-performing predictive models.
When it comes to modeling, James allows clients to leverage state-of-the-art machine learning algorithms to create high-performing predictive models and scorecards.
Once these models are created, clients can use the validation module to get pre-emptive alerts regarding their models’ metrics and access all the information required for internal validation and regulatory compliance.
The validation module was built based on Basel recommendations and validation reports provided by banks following IRB. Validated models can then be seamlessly deployed. James provides clients with all the necessary parameters for on-premises or Cloud deployments. Finally, live models can be monitored at both business and compliance levels.
James’ monitoring dashboards allow clients to keep track of a portfolio’s performance, as well as provide an easy interface to check a model’s performance, stability, and calibration.
Artifical intelligence systems for calculating credit risk
James is an ever-growing AI for credit risk based on the continuous research of four teams.
These teams consists of a regulatory advisory board comprised by members of the European Banking Authority (EBA) and specialized risk consultants who provide insights into regulatory compliance, a credit advisory board that gathers credit risk management best practices.
The board also include members of a data science research team focused on finding the most efficient ways to fulfill those best practices or even improve them; and an AI research team whose goal is to make sure that all this knowledge is easily accessible in a single solution.
State-of-the-art risk management processes
James has so far been used for a wide range of financial applications.
The platform has been implemented to help financial institutions launch credit products, update their risk assessment from standard scorecards to statistical models, and adapt from traditional statistical approaches to machine learning techniques.
Financial lenders that used James saw their models’ performance increase by up to 28%, which has allowed them to get results such as 30% default rate decreases and 10% increases in acceptance rate.