(US and Canada) Aravind Jagannathan, VP Single Family Data Officer, Freddie Mac, speaks with Chris Knerr, Chief Digital Officer, Syniti, about Freddie Mac as a unique organization that has reached a particular position in capital markets and its unusual public-private regulatory structure and mandate.

Knerr shares that Freddie Mac was founded as a government-sponsored entity back in 1970 with the mission, "We make home possible." Single-family deals with the residential side of the marketplace. So Freddie Mac currently have over 11 million active loans, properties, and a US$2.9 trillion value balance sheet. So it is very complex. In terms of scale, it is vast; they also have a multifamily division, which deals more with the commercial side. So in terms of the three lines of business for Freddie Mac, it’s all geared towards supporting customers and end customers in terms of making homes possible, whether residential or commercial.

Jagannathan discusses that when he thinks about AI in terms of challenges, he doesn’t see any particular challenges in this type of frame when he discusses it with his peers and other companies. For example, at Freddie Mac, they leverage AI and ML in really two broad scope areas: single-family risk management and operational use cases. Their AI approach is enabled within data strategy and ties to the overall vision, mission, and business objectives.

The team is focused on operation use cases. So it means that it could be sampling or predictive nature of all of our loans or unique loan characteristics or discovery or leveraging logs to predict system outages. So in the operational side of the processes, when they think about working with business partners, they need to understand what is possible and then get into what is declared business value. "This is where AI/ML is," says Knerr. "It’s not just going through the theory but making sure the organization will do something to have that clear business value in a time-relevant manner."

Based on that management review and prioritization of use cases, they execute and see the results. So it helps in terms of the oversight because they try to execute a transparent approval process. Ultimately, when you think about a single-family perspective, from a risk, origination, servicing, operational perspective, AI and ML from risk management and operational use cases are embedded to improve our process.

Knerr further shares that if as an organization they don’t know where to go, don’t have that strong governance, and then when you think about information risks whether it be from privacy or meeting standards and controls, whether it be from, how is the data protected across the ecosystem. "It is critical. You wanted them to build out and leverage AI and ML for the end-user," says Knerr. "So it is a must that people need to trust and make sure that what you’re building upon is strong. So that’s why data governance is the foundational building block to empower businesses."