Download PDF
DataRobot > Case Studies > Optimizing Loan Predictions with DataRobot AI Apps
DataRobot Logo

Optimizing Loan Predictions with DataRobot AI Apps

Technology Category
  • Analytics & Modeling - Predictive Analytics
Applicable Industries
  • Finance & Insurance
Applicable Functions
  • Business Operation
  • Sales & Marketing
Use Cases
  • Predictive Replenishment
  • Fraud Detection
Services
  • Data Science Services
  • Software Design & Engineering Services
The Challenge
The fintech company provides consumer financing to merchants and consumers at point-of-sale through more adaptable alternatives to traditional lending programs. They built models to support the company’s projects in various departments including underwriting, accounting, and collections. However, they faced a challenge in the collections department. With tens of thousands of delinquent loans at any given time, there are a lot of calls for the Collections team to make. The more successful calls they have — measured by an industry metric called Right Party Contact (RPC) — the more likely they are to be able to successfully collect on these delinquent loans, and thus bring in revenue for the company. However, with such a great volume of target calls to make and generally low connection rates in terms of reaching the right person or party, any type of optimization or efficiency can make a big difference.
About The Customer
The customer is an innovative fintech company based in the USA. They offer alternatives to traditional consumer lending products. The company primarily provides small consumer loans — under $5,000 — to personal borrowers, as well as financing options for merchant stores. Collecting on these loans is a huge part of the company’s incoming revenue every month, putting the Collections team at the forefront of the company’s business. They have a small team of data scientists and analysts who were already impressed with the DataRobot platform and were finding tremendous productivity and data science gains from using it.
The Solution
The team saw an opportunity to leverage DataRobot’s models and working with DataRobot’s new AI Applications team, built a beta version of a DataRobot Optimizer application. The app, powered by DataRobot’s models, was able to make predictions on the best time of day to call and connect with the large number of delinquent accounts, and make those predictions in under 20 minutes. Those predictions are then pushed to an autodialer system used by the Collections team to work more efficiently through an optimized list that tells them whom to dial at which time. The DataRobot team implemented a new specialized optimization algorithm, with 220x improved throughput, meeting the challenge to complete the job within 15-18 minutes. This new algorithm has been integrated into the Optimizer application and is available for any customer with lead scoring, churn, or upsell use cases.
Operational Impact
  • The DataRobot predictions through the Optimizer App have led to improved connection rates of as much as 0.1-0.2%.
  • The time freed up for data scientists and analysts on the small team.
  • The partnership with DataRobot represented a unique opportunity to lend their frontline insights to a vendor and collaborate on a solution together.
Quantitative Benefit
  • 220x improved throughput with the new specialized optimization algorithm.
  • The job is completed within 15-18 minutes.
  • Overall connect rates of around 3-4%.

Related Case Studies.

Contact us

Let's talk!

* Required
* Required
* Required
* Invalid email address
By submitting this form, you agree that IoT ONE may contact you with insights and marketing messaging.
No thanks, I don't want to receive any marketing emails from IoT ONE.
Submit

Thank you for your message!
We will contact you soon.