Artificial intelligence (AI) is rapidly changing the risk and compliance sector, according to insights shared during a recent webinar titled "Mind the gap: AI’s big leap in risk and compliance." The discussion brought together industry leaders to examine how financial institutions are adopting AI, the challenges they face, and what strategies are needed moving forward.

Paul Nola, Partner at We Live Context, referenced Moody’s 2025 global study "From reactive to proactive: How AI is transforming risk and compliance," noting that active use of AI among organizations has increased from 9% to 24%, with up to 53% when including pilot programs. However, only about a third of these organizations report significant impact or measure their success with AI tools. “AI is moving out of innovation labs and into frontline risk and compliance functions,” said one panelist. This shift is being driven by growing regulatory demands and an increase in bad actors using AI.

Ted Datta, Senior Director at Moody’s, highlighted the pace of change: “The 'wait and see' era for AI in compliance is officially over. Our latest global study shows adoption has surged from 30% to 53% in just two years. This isn’t just a trend—it’s a shift toward proactive, tech-enabled risk management. The real question is: is your firm part of the 53%, and are you gaining an edge?”

Datta also pointed out concerns about oversight: “Here’s a red flag: over a third of firms using AI aren’t measuring its effectiveness. Is this just a growing pain of rapid adoption—or does it explain why 46% report only moderate impact? Either way, it’s a governance blind spot we need to close.”

Panelists agreed that specificity in choosing AI tools is crucial; general-purpose models can help with tasks like summarizing emails or drafting policies but may not be suitable for more complex decisions such as transaction approvals. A clear strategy combining traditional machine learning with customized models was recommended.

Explainability was another key topic. Panelists stressed that transparency in how AI makes decisions builds trust—a necessity for risk professionals. One noted, “If you can’t explain it, you probably shouldn’t rely on it,” emphasizing the importance of data-driven frameworks.

Data quality emerged as foundational for successful AI projects. The panel warned that poor data often leads to project failures and outlined steps such as mapping, cleaning, validation, and tagging data before implementing AI solutions. Datta stated: “What’s the number one predictor of AI success in compliance? Data quality. Firms using AI are 2.2 times more likely to report high-quality data. Yet only 27% of organizations rate their data as ‘high quality.’ That’s a massive opportunity—mastering your data strategy could be the key to leapfrogging the competition.”

When considering autonomous decision-making by AI systems, most participants favored keeping humans involved—especially for high-risk situations requiring domain expertise and oversight. Datta observed: “With 62% of firms now encouraging the use of LLMs, we’re seeing a new tension emerge. What’s the bigger compliance risk: the rise of unmonitored ‘Shadow AI’ or an over-reliance on AI that erodes human judgment?”

Regulation was seen as beneficial rather than restrictive by many panelists; new rules from entities like the EU and Canada provide necessary guidelines while maintaining corporate responsibility.

Looking ahead, panelists said that roles within compliance departments will evolve alongside technology advances; nearly all surveyed professionals expect their responsibilities will shift toward more strategic work rather than being replaced by automation alone.

Datta summarized this point: “AI isn’t about replacement—it’s about reinvention. Our data shows 96% of compliance professionals expect their roles to evolve, with most seeing a shift toward strategic and advisory functions. Is your firm investing in upskilling to unlock this potential, or are you risking your top talent being left behind?”

For those interested in further details or listening to the full webinar discussion on these findings from Moody's Analytics research team visit http://moodys.com/kyc/ai-study