The Royal Bank of Canada (RBC) is leveraging its proprietary AI foundation model to enhance credit adjudication and personalize its loyalty programs.
Developed by RBC’s AI research division, Borealis, the Atom (Asynchronous Temporal Model) has been trained on extensive financial datasets, encompassing billions of client transactions. This training provides what RBC describes as “deep financial expertise” applicable across a range of operations.
The bank envisions Atom as pivotal to its goal of generating $700 million to $1 billion in enterprise value through AI-driven initiatives by 2027. The model is currently enhancing credit adjudication processes, making them more accurate, consistent, and insightful. It effectively manages vast amounts of complex data, including transaction histories and non-traditional sources, enabling RBC to offer credit to clients, such as newcomers, who may be overlooked by conventional models.
According to Gopala Narayanan, Senior Vice President and Chief Risk Officer of Personal Banking at RBC, “Utilizing AI in credit adjudication has allowed us to refine decision-making, ensuring optimal customer outcomes and fostering higher levels of personalization that traditional processes cannot achieve.”
In addition, Atom is instrumental in the Avion Rewards program, providing tailored recommendations that have resulted in a notable increase in redemption rates, as well as cost efficiencies and enhanced benefit adoption.
“Atom signifies the future of banking at RBC,” states Foteini Agrafioti, Senior Vice President and Chief Science Officer at RBC. “It enables us to customize products and services for individual clients and allows for a deeper understanding of their specific needs.”