Bank of Singapore has deployed an agentic AI tool to automate the drafting of source-of-wealth (SoW) reports, a core Know-Your-Customer (KYC) step for verifying clients’ wealth and transactions. The bank says the agentic AI tool, called Source of Wealth Assistant (SOWA), reduces turnaround from about 10 days to one hour while aligning with regulatory standards.

The agentic AI tool is designed to help relationship managers (RMs) produce more consistent reports and improve account-opening efficiency—an area the industry has been examining.  SOWA ingests documents such as financial statements, tax notices, property valuations, corporate filings and payslips, then generates a structured, standardised draft for review.

How the agentic AI tool works

SOWA aims to minimise human errors linked to manual compilation by surfacing gaps and inconsistencies, and by validating client information against benchmarks such as salary ranges and company revenue using data from Bank of Singapore and parent OCBC.  RMs retain oversight: they review and refine the draft before it moves to internal review teams as part of anti-money laundering and counter-terrorism financing controls.

For security, SOWA runs in the bank’s private cloud, a computing environment dedicated to Bank of Singapore’s use.  OCBC co-developed the agentic AI tool with the private bank, building on prior investments in generative AI and traditional AI.

Why agentic AI matters for KYC in Asia

Agentic AI goes beyond prompt-response systems by taking initiative, coordinating tools and adapting in real time, learning from prior interactions to improve performance.  OCBC says AI now powers more than six million decisions daily across the group. Its in-house AI tools follow the FEAT principles (Fairness, Ethics, Accountability and Transparency) and undergo regular reviews to prevent bias.

“In an increasingly complex risk landscape, AI can automate repetitive tasks like report generation and data validation,” said Kam Chin Wong, Global Head of Financial Crime Compliance at Bank of Singapore, adding that agentic AI enhances efficiency, accuracy and consistency while allowing RMs to focus on client engagement and risk assessment.

Ruth Yeo, a relationship manager at the bank, said SOWA has made the process more intuitive, with RMs now verifying uploaded details and refining the AI-generated draft rather than sifting through emails and documents. .

The launch comes as private banks in Asia seek to accelerate compliant onboarding while meeting stricter KYC/AML requirements. Bank of Singapore positions SOWA alongside other OCBC AI deployments, including tools for investment ideas, internal knowledge search, call summarisation and staff productivity.

Bank of Singapore did not disclose adoption figures or timelines for additional modules. The bank says SOWA will remain on its private cloud and continue to be refined under OCBC’s AI governance framework.