“Our survey shows Singaporean firms lead the world in AI adoption [but many still] lost clients due to inefficient onboarding. This underlines the importance of embedding AI into every stage of the client lifecycle, from onboarding to surveillance, to achieve both speed and stronger risk management. By linking these processes end-to-end, institutions have proven to reduce abandonment while meeting the Monetary Authority of Singapore's heightened supervisory standards,” says Cengiz Kiamil, managing director for Asia Pacific at Fenergo.
The tension between speed and compliance is not limited to Singapore. Globally, 70% of financial institutions reported client attrition in 2025 linked to slow onboarding, the highest level recorded to date. Client abandonment rates now average around 10%.
AI ambition vs operational reality
While AI adoption is accelerating, Flint Global’s 2025 report, commissioned by AWS, finds that the path from pilot to production remains steep. Across Asia Pacific (especially in Singapore, which is often viewed as a regional benchmark), data fragmentation and legacy system constraints remain the primary drags preventing AI models from scaling end-to-end, particularly in onboarding workflows.
See also: AI’s power problem is being overthought, says Microsoft’s VP for energy
In many firms, AI-powered risk scoring and document verification tools perform well in isolation but fail to integrate into holistic client-lifecycle engines. This disconnect contributes to customer drop-off even in jurisdictions with high AI penetration.
Fenergo’s data echo the same trend. AI adoption in KYC/AML has leapt from 42% in 2024 to 82% in 2025, but automation of periodic KYC reviews averages only about a third of total workloads. This suggests that many financial institutions still rely heavily on manual intervention despite large-scale technology investments.
