Organizations report 60-70% reductions in false positives using machine learning models in transaction monitoring. AI-assisted investigation workflows automate data aggregation and narrative drafting. However, concerns include hallucination risks in SAR filings and the need for governance frameworks. Analysts must form independent hypotheses before reviewing AI output to maintain investigative judgment and prevent cognitive surrender.

Platforms operating both licensed financial products and unlicensed payment-adjacent products face unique challenges. No regulatory obligation exists to collect identity information from unlicensed users, creating friction and competitive disadvantage. Regulators increasingly expect robust controls regardless of licensing status, with DOJ frameworks extending to unlicensed businesses. Reputational exposure justifies strong controls even without explicit legal obligations.

Organizations explore behavioral and signal-based identity verification beyond documentary methods. Triangulating identity confidence from IP address, device fingerprint, transaction history, and metadata builds persistent identity pictures. Mule account detection relies on behavioral signals and network analysis. Stablecoins and unhosted wallets present frontier challenges where existing vendor solutions and regulatory frameworks have not yet caught up.