Ask us how we can help you succeed.
One of the most meaningful changes in 2025 was the significant reduction in ‘false positives’ across firms, in other words, alerting customers about suspected crimes that were in fact legitimate activities. Some organisations saw decreases of up to 60% this is a result of more sophisticated AI models capable of understanding context, refining alerts and drastically helping analysts. Large language models played a key role in improving alert quality and speeding up analysis, enabling compliance teams to focus time on genuinely suspicious behaviour, rather than the mundane task of sifting through low-risk alerts.
Alongside this, agentic AI workflows gained momentum. These systems are now capable of autonomously gathering customer information, building risk profiles and drafting Suspicious Activity Reports for human approval. This evolution was the beginning of a more automated financial crime operations landscape, where technology performs the heavy lifting and human specialists oversee decision-making.
A significant development in 2025 was the operationalisation of Perpetual KYC, enabled by rapid advancements in AI. Instead of retrospective reviews, sophisticated AI-powered systems can now maintain constant oversight of customer activity and behaviour, triggering enhanced reviews the moment a risk signal changes. This shift from static to continuous monitoring has made KYC more dynamic and aligns with the real-time nature of financial crime.
What is on the horizon in 2026?
Looking ahead, 2026 is expected to be the year that AI in financial crime becomes tightly governed. Regulators in the EU, UK and the US are emphasising transparency, explainability and human oversight. The EU AI Act, coming fully into effect by August 2026, and the Colorado AI Act in the US, both force firms to ensure that AI models are ethical, and free from unfair bias. As a result, human validation will become mandatory for many decision points, ensuring that AI enhances risk assessment rather than replacing human judgement.
Real-time KYC and AML will continue to advance throughout 2026, with financial institutions moving towards automated due diligence and continuous monitoring. AI systems capable of assessing identity, transaction patterns and behavioural signals during interaction will enable faster onboarding, and more effective crime prevention.