Deutsche Bank’s private bank is using AI to assemble evidence about how clients accumulated their wealth—not to approve the resulting assessments. In a new WealthBriefing interview, chief operating officer Yiping Li detailed the controls behind that division of work, including source references and confidence indicators designed to help employees challenge the output.
From scattered evidence to a review memo
The system supports source-of-wealth assessments, a part of Know Your Customer, or KYC, checks. These require banks to establish and document how prospective clients built their wealth. Deutsche Bank’s September announcement said the tool draws on material already in a client’s case and approved external sources, automating research, documentation and preparation of the information needed for an assessment.
Inside the bank’s digital KYC platform, the solution extracts information, sorts it into categories and structures it for use. It then cross-checks sources and reconciles the material before producing a narrative and memo for a person to assess. The bank says the system also helps identify gaps or inconsistencies in the underlying information.
That workflow has been operating since early September 2026 in the bank’s Singapore and Hong Kong booking centres. The interview describes its current scope as selected income-related parts of source-of-wealth assessments. The September announcement also said advisors in Dubai had access for accounts booked in Singapore.
Li said the aim is to improve the quality and consistency of KYC files while reducing handling time for KYC officers and relationship managers. The bank’s stated business goal is to reduce administrative work so relationship managers can spend more time advising clients.
A draft employees must be able to challenge
The output includes references to websites or documents and confidence levels for sources. Those indicators are intended to point reviewers toward material needing closer examination. They sit alongside a clear allocation of responsibility: specialists validate the content, resolve flags and retain final approval and case closure, the bank told WealthBriefing.
Outputs need to be grounded in identifiable sources and presented in a way that allows the reviewer to understand and challenge the content and conclusion
Yiping Li, chief operating officer of Deutsche Bank Private Bank, speaking to WealthBriefing
Other safeguards begin before a reviewer reads the memo. The bank specifies preferred external sources, corroborates information across sources and uses deterministic checks where appropriate—checks based on fixed rules. It describes these controls as working together to address outputs that sound plausible but are incorrect, rather than relying on a single check at the end.
Europe brings a scaling problem, not just more users
Work is underway to widen the tool’s scope and scale deployment across European regions, the bank told WealthBriefing. That follows the broader rollout planned in its September announcement. The expansion is still described as work in progress, while the live use detailed in the interview remains limited to selected assessment areas in Singapore and Hong Kong.
Li identified obstacles beyond the AI itself: fragmented systems and documents, regulatory and banking secrecy requirements, engineering capacity, processes and culture. Employees also need training in using the tools and challenging their outputs. His emphasis was on reusable AI capabilities and governance, rather than building a separate technical solution for every application.
Potential future KYC uses include gathering documents, identification and verification, and understanding complex legal ownership structures. Those are opportunities Li outlined, not additional capabilities confirmed as live. For the deployed wealth-review tool, the operating boundary remains explicit: AI prepares the material; employees exercise judgment and remain accountable for the completed case.
Reader comments
Newest comments first. Replies stay oldest first.