248: How Moody's is Rethinking AI in Finance

How do you bring innovation to life inside an organization whose job is to help other people see risk before it shows up on a balance sheet?

In this episode of the Innovation Storytellers Show, I sit down with Jason Lee, Chief Intelligence Officer at Moody's Analytics, for a conversation that lives at the crossroads of national security tradecraft, financial crime investigation, and modern data-driven decision making. Jason has spent decades inside large, complex systems, from federal intelligence work to investment banking to building a security consulting firm, and he shares what he has learned about creating new programs inside environments where bureaucracy, budgets, and skepticism can slow even the best ideas down.

We start with Jason's origin story because he makes a compelling point: innovation rarely comes from a formal job description. In his career, it often showed up as a "collateral duty," a leader asking him to solve a pain point, build a new unit, or design a process when the rules had not yet been written. From creating early fraud detection frameworks in banking to uncovering unconventional data sources in government work, Jason frames innovation as a mix of creativity, relationship-building, and a willingness to learn from other industries without copying them.

From there, we get into how Moody's is thinking about AI right now, especially the shift from large language models toward large reasoning models. Jason explains why reasoning matters more than hype when the stakes include fraud, terrorism financing, and organized crime. He walks through what it means to use models for scenario analysis, how "tipping and cueing" can help analysts focus on what matters, and why he believes humans have to stay in the loop, especially when errors can have real-world consequences.

One of my favorite parts of the conversation is when Jason brings storytelling back into the center of analytics. He explains how workshops with prospects help uncover what clients actually need, even when they cannot fully articulate it yet, and why "data experience" matters when the information is complex and intangible. We also talk candidly about where innovation programs can stall, whether it is budget politics, unrealistic KPIs, mismatched expectations across business verticals, or leaders who want short-term wins when the real value takes years to compound.

If you are building inside a big organization, selling complex ideas to busy decision-makers, or trying to make AI useful without losing trust, this episode will give you a lot to think about, so what part of Jason's approach resonates most with how you see innovation playing out right now, and where do you think teams are still getting stuck?