Navigating AML flags for new crypto payment rails

Curious if anyone has experience with the evolving landscape of AML flags when onboarding new clients, particularly those utilizing novel crypto payment rails or cross-border solutions. We're seeing a wider range of 'unusual' transaction patterns flagged by our systems that, on closer inspection, often turn out to be legitimate but just don't fit traditional fiat-centric profiles. How are others adjusting their risk scoring and SAR thresholds without casting too wide a net or, conversely, missing genuine red flags? It feels like the industry is still catching up to the pace of innovation here, especially with the differing regulatory stances globally.