The Limits of Conventional Analytics
Over the last 50 years, Freddie Mac has helped people realize their dream of owning a home more than 80 million times. The company has funded $11.6 trillion in mortgages and financed $6 million in rental units.
In 1970, Congress chartered Freddie Mac to support the U.S. housing finance system. Rather than lending directly to borrowers, Freddie Mac buys loans from approved lenders.
As market and economic conditions change, Freddie Mac must remain flexible and continuously deliver on its commitment to affordable, adequate housing. In a sea of unstructured and semi-structured data, it’s challenging to achieve meaningful predictions and key insights to inform business decisions. Working with hundreds of thousands of customers, and mining nearly four terabytes of data, they found business intelligence and manual practices didn’t scale.
Making Sense of Data – More Quickly and Accurately
“With DataRobot, we can analyze these large, complex datasets and gain valuable insights more quickly,” said Lakshmi Purushothaman, Vice President, Innovation in Data Science, Engineering, and Analytics.
The analytics team creates models that span across the organization, bringing value to internal teams, lenders, and their end customers.
As Freddie Mac collects front-end information from lenders and their customers and analyzes housing markets and properties, AI helps the business make sense of the data. The platform extracts data elements from various text documents and images much more quickly and accurately than with the previous manual approach.
Increasing Analytics Team Productivity by 2.7X
The agency modernized its AI and ML infrastructure, shrinking the MLDev and deployment cycle to deliver meaningful value to the business rapidly. The DataRobot platform helps Freddie Mac rapidly home in on the winning models.
Ultimately, the Freddie Mac analytics team attributes significant efficiency to the platform:
- 2 to 10 times faster proof of concept
- 1700+ hours saved in model development time per analytics project
- A 2.7X productivity gain for a corresponding jump in time to market
This efficiency means that the data science team can focus on more use cases and scale more readily.
“Our ability to leverage data science to help us identify disparities, remove barriers, and enable informed decisions from our data, which has been exploding in terms of variety, volume, and velocity over the years, has been made much easier with DataRobot,” Purushothaman said.
“We’ve automated the stuff that data scientists didn’t really like doing so they can focus on what really drives change,” added Aravind Jagannathan, Chief Data Officer. “AI/ML has been critical in terms of the efficiency we’ve achieved by allowing us to scale massively.”
Managing Governance and Explainability with an AI Center of Excellence
The DataRobot AI platform offers essential interpretability and explainability for stakeholders and compliance teams. The team saves time and work because the DataRobot platform collects the needed documentation—available in one-click reports. Explainability tools help clarify AI models into business-speak, detailing what’s behind the models.
To promote model governance and manage risk, the Freddie Mac team also created an AI Center of Excellence (CoE). Among its roles, the CoE ensures that the various people involved in analytics projects understand the governance required.
From the outset of the relationship to the current stage, data scientists work closely with the DataRobot team via office hours and deep-dive workshops to explore use cases and apply best practices throughout the process.
“When I think about how DataRobot has enabled us and supported us with our current use cases or ideas, it’s really tied to helping us with our objectives,” Jagannathan said. “From a customer perspective, I’ve found that’s rare in a partnership. It’s fantastic to bear the fruit of that relationship.”
Speed to Market at Lower Cost
As Freddie Mac looks ahead, the organization is optimistic about the power and potential of AI to drive its business goals.
“The value is ultimately making sure we’re oriented to the customer always,” said Tatyana Krol, Senior Director of Business Intelligence and Analytics. “We’re augmenting the decision-making with AI and letting people do what they’re best at. It’s definitely going to have a transformative value over the next several decades.”