How a Fast-Growing Lender Minimizes Risk, Maximizes Revenue with DataRobot

We succeeded in increasing our loan acceptance rate, so we sell more while keeping risk at the same level. In addition to other demographics, we’re serving unbanked individuals, giving them access to legal capital and a chance to build their credit history.

Navigating a Digital Transformation

For many Armenians, Global Credit is a lifeline. Without the lender, many individuals and businesses may not be approved for a loan due to low income or poor credit scores. For this market segment and others, the financial services company provides fast and accessible consumer, mortgage, business, and agricultural loans to meet customers’ short- and long-term goals. 

Global Credit recently pivoted to invest in financial technology offerings to bolster its competitive position in the market. The company now delivers all lending services online, including 99% of its loans, with the power of risk identification and strong compliance. In order to get here, Global Credit built an easy-to-navigate user interface, integrated third-party applications, moved to online processing, diversified services, implemented a successful digital strategy, and developed a mobile app, making it one of the few companies in Armenia to offer such a convenience. This app gives clients seamless access to all lending services and enhances customer experience.

As Global Credit navigates its digital transformation, Tamara Harutyunyan, Chief Risk Officer and Chief Data Officer, sees data science as a competitive advantage.

“In Armenia, every financial institution has access to the same government-backed data,” Harutyunyan said. “When everyone has access to the same information, we have to differentiate in other ways to obtain a competitive advantage.” 

Harutyunyan’s dual role as Chief Risk Officer and Chief Data Officer gives her a unique perspective — and distinct criteria — for leveraging AI in financial services. Her responsibility to Global Credit’s customers is a guiding light for a responsible AI program with strong governance and transparency. That philosophy allows Global Credit to embrace cutting-edge AI while maintaining strict compliance and minimal risk. 

Streamlining the Entire AI Lifecycle in FinTech 

As a tightly regulated company, Global Credit wants to keep its data science efforts in-house. However, the data team can only spend a small percentage of its time on modeling.

In previous organizations, Harutyunyan has repeatedly relied on DataRobot for predictive analytics. Again, she turned to the platform to empower her team at Global Credit to do more. 

“I’ve brought in DataRobot every time I’ve joined a new company,” she said. “I’ve found it has all the functionality and support we need.”

DataRobot’s end-to-end AI lifecycle capabilities let them minimize modeling time while simultaneously focusing on managing the many parts of their lending engine and driving Global Credit’s digital transformation. 

The value of the platform also extends to AI models in production by providing critical data governance. On the front end, Harutyunyan finds that DataRobot helps enforce a sound data policy. Then, it provides essential visibility and explainability behind models to help comply with regulatory standards, a critical component for managing sustainable AI programs.

DataRobot Experts Augment Internal Expertise 

According to Global Credit, onboarding was faster and easier with support from the data scientists at DataRobot. They brought technical know-how along with ideas and guidance for industry-specific use cases. Together, Global Credit and DataRobot brainstorm solutions to challenges.

“One of my favorite parts of the week is our conversations with our DataRobot data scientist,” Harutyunyan said. “They ensure we’re always up to date on new capabilities and suggest use cases we may not even be familiar with. They get to know our business and always bring solutions.”

Global Credit uses DataRobot and a Microsoft SQL database. From there, the team rapidly uncovers insights. In particular, a real-time look at data drift pinpoints changes to the population, prompting model updates.

“The data drift capabilities are huge,” Harutyunyan said. “With DataRobot, we see the changes that are happening and recalibrate our models to improve performance.”

With support from DataRobot’s AI Applied Experts, Global Credit developed and deployed several high-impact models in just eight weeks. As a result, they increased the loan acceptance rate — and revenue — while keeping risk at the same level. 

Balancing Risk and Reward with AI in Banking 

In one of its top use cases, Global Credit models the probability of loan default. The financial company aims to reach an underserved market, the unbanked, while mitigating risk. 

DataRobot helps the lender find unbanked individuals with a lower probability of loan default. This ensures that Global Credit can keep its acceptance rate — and revenue — as high as possible while controlling risk.

“We succeeded in increasing our loan acceptance rate, so we sell more while keeping risk at the same level,” Harutyunyan said. “In addition to other demographics, we’re serving unbanked individuals, giving them access to legal capital and a chance to build their credit history.”

Global Credit also taps into DataRobot to understand the propensity of former customers to come back. Armed with accurate predictions, the marketing and sales teams can reduce excess spending and focus on the most impactful consumer segments. This gives the company more control over an expensive, but critical, cost center.  

Going forward, Harutyunyan envisions a debt collection scorecard that enables Global Credit to proactively alert customers as they get close to their due date and limit collection resources for those likely to pay.

“If we can text customers a reminder, we’re not using full-time employees for collections and we’re not frustrating our customers,” Harutyunyan said. 

This program would both improve the customer experience and reduce human hours needed to manage debt collection.

The lender is also developing a model to find customers likely to pay loans early, further streamlining the debt collection process. 

A Collaborative Partnership

With the DataRobot platform, Global Credit ultimately gains greater clarity into its customers, supporting more targeted outreach, greater operating efficiency, and better customer experiences. 

“DataRobot gives us very quick insight into our data,” Harutyunyan said. “It helps us understand our customers better, who they are, and how they’re reacting to changes in our products.”

As Harutyunyan speaks with colleagues in the industry, many of whom must build an internal data science capacity, she recommends DataRobot.

When I’m talking with peers from other financial institutions, I always tell them, ‘We have a product and a support team in DataRobot.’ They help us uncover all the possibilities and use cases to solve our business challenges.