Riten Dixit, a seasoned expert in market and credit risk, reflects on the evolving challenges faced by risk modelers in a rapidly shifting financial landscape. He notes that while models are built to handle volatility through stress testing and simulations, they often fall short when deeper, structural regime shifts occur—like those seen in mortgage markets post-COVID or in reaction to inflation-driven rate hikes.

He highlights how consumer behavior has transformed due to factors like ultra-low mortgage rates and remote work, creating scenarios that legacy models simply weren’t designed to handle. These shifts underscore the need for more coherent, flexible modeling frameworks capable of detecting when the underlying rules of the game have changed.

Rather than replacing traditional models, Dixit believes AI and machine learning should complement them. Where conventional models offer structure and interpretability, AI offers scale and pattern recognition—spotting emerging risks early, such as bank run signals missed by financial ratios alone. The key, he argues, is not to chase complexity but to leverage context and curiosity to connect signals and sharpen human judgment.

Dixit also calls for a shift toward micro-level modeling using tools like geospatial mapping and behavioral clustering, helping risk managers move from broad narratives to actionable local insights. In a world where past data is increasingly an imperfect guide, history should be treated as a boundary, not a blueprint. Looking ahead, the winners in risk modeling won’t be those with the flashiest tools, but those who blend theory, adaptability, and explainability to make better decisions—especially when the future refuses to resemble the past.