June 21st, 2019

Underwrite.ai Transforms Credit Risk Decision-Making Using AI

RSS icon RSS Category: Customer, Data Science, Financial Services, H2O Driverless AI, Machine Learning Interpretability

Determining credit has been done by traditional techniques for decades. The challenge with traditional credit underwriting is that it doesn’t take into account all of the various aspects or features of an individual’s credit ability.  Underwrite.ai, a new credit startup, saw this as an opportunity to apply machine learning and AI to improve credit lending decisions. The company considers many more factors to construct a credit profile, which will help a lender make a decision that can be much more accurate.

By using H2O Driverless AI, an automatic machine learning platform, they were able to achieve better results than traditional credit scoring techniques. Here’s how – Driverless AI took in a large amount of data, determined relationships and patterns, and constructed a predictive model that considered approximately 763 different features. From Driverless AI, the company was able to automatically build the Adverse Action Report that lenders need in order to make a financial decision and do that within nanoseconds. Not only was this predictive model accurate, it was faster in making a decision too.

Marc Stein, underwrite.ai’s Founder and CEO emphasizes that “Not only can we reach a decision, we can also explain the decision and be fully compliant with the regulations. And it takes less than half a millisecond”. The platform provides the user with many visualizations and graphs to understand and explain the results. Because machine learning develops a non-linear model, it can be confusing to truly grasp the meaning derived from the data. With Driverless AI machine learning interpretability capability, feature interactions are not only understandable but explained.

How much better was the new credit risk model than traditional techniques? Well, the model had an accuracy of .95813, which according to the Marc Stein was ground-breaking: “This is the first time I’ve ever had a model at .95,” Marc Stein adds. After implementing Driverless AI, underwrite.ai has surpassed their target of producing a decision for a client in 20 milliseconds to producing a decision in half a millisecond.

“Using thousands of data points instead of the handful used by traditional credit scoring services, underwrite.ai generates a much more reliable lending risk profile for creditors to use when evaluating a credit or loan application,” explains Marc Stein. “By using Driverless AI to build and deploy our nonlinear models, we can provide lenders a fully automated process that delivers a highly reliable credit decision with a fully compliant explanation for the decision in less than half a millisecond. H2O and Driverless AI play an integral role in our platform, speeding up the development and deployment of models exponentially.”

Big thanks to AI and automatic machine learning for transforming our long-standing credit system.

About the Author

Priya Jain

Priya is a Marketing Intern at H2O.ai. She is pursuing a Bachelor's degree in Cognitive Science and Economics from Rutgers University. In her free time, Priya likes to play table tennis, listen to music, write poems, and play the flute.

Leave a Reply

H2O LLM DataStudio Part II: Convert Documents to QA Pairs for fine tuning of LLMs

Convert unstructured datasets to Question-answer pairs required for LLM fine-tuning and other downstream tasks with

September 22, 2023 - by Genevieve Richards, Tarique Hussain and Shivam Bansal
Building a Fraud Detection Model with H2O AI Cloud

In a previous article[1], we discussed how machine learning could be harnessed to mitigate fraud.

July 28, 2023 - by Asghar Ghorbani
A Look at the UniformRobust Method for Histogram Type

Tree-based algorithms, especially Gradient Boosting Machines (GBM's), are one of the most popular algorithms used.

July 25, 2023 - by Hannah Tillman and Megan Kurka
H2O LLM EvalGPT: A Comprehensive Tool for Evaluating Large Language Models

In an era where Large Language Models (LLMs) are rapidly gaining traction for diverse applications,

July 19, 2023 - by Srinivas Neppalli, Abhay Singhal and Michal Malohlava
Testing Large Language Model (LLM) Vulnerabilities Using Adversarial Attacks

Adversarial analysis seeks to explain a machine learning model by understanding locally what changes need

July 19, 2023 - by Kim Montgomery, Pramit Choudhary and Michal Malohlava
Reducing False Positives in Financial Transactions with AutoML

In an increasingly digital world, combating financial fraud is a high-stakes game. However, the systems

July 14, 2023 - by Asghar Ghorbani

Ready to see the H2O.ai platform in action?

Make data and AI deliver meaningful and significant value to your organization with our state-of-the-art AI platform.