5 ways MAS's new SOW circular + agentic AI will transform HNW onboarding

Singapore just raised the bar again for wealth‑management AML. MAS's July 2024 circular (AMLD 08/2024) on establishing sources of wealth (SOW) tells private banks exactly how to be tougher on financial crime while minimizing client friction.

At its core, the circular reframes the SOW as a risk‑proportionate, evidence‑driven exercise. MAS anchors expectations to three principles:

• Materiality – focus corroboration on the most significant and higher‑risk components of a client's wealth, not on every minor asset.

• Prudence – use more reliable sources (audited accounts and independent third‑party documents) for those material elements.

• Relevance – seek fit‑for‑purpose evidence, avoid outdated or low‑value documents, and leverage credible public sources where possible.

For HNW and UHNW onboarding, this is a quiet revolution. It gives institutions regulatory cover to move away from "collect everything just in case" toward targeted, explainable SOW work. Done well, that means: shorter onboarding timelines, fewer pointless document chases, clearer escalation when material SOW can't be fully corroborated, and ongoing monitoring calibrated to the client's risk profile and SOW story.

The missing piece is execution at scale. This is where agentic AI becomes a force multiplier.

Instead of static workflows, think of a network of specialized AI agents orchestrating the end‑to‑end journey:

• An intake agent structures client declarations and identifies potentially material SOW components.

• A research agent pulls corroboration from company registries, financials, media, property, and corporate records, then benchmarks plausibility against external data.

• A risk‑analysis agent applies the MAS principles to prioritize evidence, flag residual‑risk gaps, and propose mitigants (e.g., tighter monitoring, limited product scope).

• A reporting agent drafts the SOW narrative, escalation memo, and justification, creating a clean audit trail for senior management and regulators.

Humans remain firmly in control: approving high-risk cases, challenging the AI's assumptions, and handling nuanced client discussions. But the heavy lifting, including data gathering, plausibility checks, and documentation, is handled by agents that never tire, never forget a rule, and can be instantly updated when MAS guidance evolves. The firms that win in this new environment will treat the MAS circular as a blueprint for redesigning HNW onboarding around risk‑based SOW and then use agentic AI to turn that blueprint into a living, continuously improving capability.