AI use case

AI agent for source of wealth due diligence in private banking

An AI agent that reads a prospective private client's documents, extracts and corroborates how their wealth was built, checks plausibility against benchmarks and external sources, and drafts the source of wealth and enhanced due diligence narrative for the relationship manager and compliance analyst, who decide on the risk rating and the relationship.

By Len Debets · Last verified 26 September 2026 · 3 public deployments

1 hours
Cycle time
Bank of Singapore (organization claim).
USD 640,000 to USD 3.8 million
Indicative value per year
A private bank onboarding 2,000 new clients a year. Worked example, see how it is calculated.

What problem does it solve?

Private banks must understand and document how a client acquired their total wealth, not only the funds that arrive in the first account. For entrepreneurs, heirs and executives this means reading hundreds of pages of financial statements, tax notices, corporate filings, property valuations and payslips, reconciling them into a coherent story and writing it up. Bank of Singapore described this as work that took its relationship managers about ten days per report.

The work is subjective and varies with the experience of the person writing it, so reports are inconsistent, and gaps that surface in compliance review send the file back and delay the account opening. Regulators have pushed in both directions: after major money laundering cases they expect more rigorous source of wealth work, and in Singapore the regulator has also asked private banks to shorten account opening times.

How does it work?

  1. Collect documents. The relationship manager uploads the client's documents and the fact find; the agent classifies them and lists what is missing for the client's profile.
  2. Extract and reconcile. Income, business sales, inheritances, investment gains and assets are extracted with dates and amounts and reconciled into a wealth timeline.
  3. Corroborate. The agent checks plausibility against benchmarks (for example typical salary for a role, company revenue) and approved external sources such as company registries and news, alongside the separate PEP, sanctions and adverse media screening results.
  4. Draft the narrative. A structured source of wealth report is drafted in the bank's template, citing the document and page behind each statement and flagging gaps and inconsistencies.
  5. Human decision. The relationship manager verifies and refines the draft; the compliance analyst challenges it, requests more evidence if needed and decides the risk rating and whether to proceed.
Audience
Employee facing
Autonomy
Copilot
Adoption
Early adopters
Channels
Internal tools

What is it worth?

Benchmarks are computed from the public deployments below: one data point per organization per KPI, with who made each claim.

Value benchmarks for AI agent for source of wealth due diligence in private banking
KPIMedianReported rangeData pointsClaimed by
Cycle timeNot pooled
1 hours
11 organization
Interactions handledNot pooled
about 50
11 organization

Value drivers: Speed and cycle time, Compliance quality, Employee productivity, Customer experience.

Indicative value

A private bank onboarding 2,000 new clients a year

USD 640,000 to USD 3.8 million

Value of staff time released from source of wealth reports per year

How this is calculated

Formula: newClients * hoursPerReport * shareSaved * hourlyCost. The low scenario uses every low input, the high scenario every high input.

InputLowHighBasis
New private clients needing a source of wealth report per year newClients, clients per year2,0002,000The reference bank.
Staff hours per source of wealth report today hoursPerReport, hours per report816Editorial assumption for effort, not elapsed time. Bank of Singapore reports elapsed writing time fell from 10 days to one hour; replace with your own effort data.
Share of effort saved with an AI drafted report shareSaved, fraction of effort0.50.8Conservative against the benchmark on this page, because verification and compliance review remain.
Blended hourly cost of relationship managers and analysts hourlyCost, USD per hour80150Editorial assumption, replace with your own fully loaded cost.

What it leaves out: Leaves out the larger commercial effect of faster account opening (assets that arrive sooner, fewer abandoned onboardings) and the cost of tooling and external data.

Who already uses it?

3 public deployments, strongest evidence first. Grades: A regulator or audit, B the organization itself, C vendor case study, D anonymous or estimate.

Bank of Singapore

Singapore · Wealth and asset management · 2025

ProductionGrade B

Bank of Singapore, the private bank of OCBC, rolled out an agentic AI tool that drafts source of wealth reports for know your customer due diligence. Relationship managers upload the client's documents (financial statements, tax notices, property valuations, corporate filings, payslips) and SOWA reviews them and generates a standardized report, checking plausibility against benchmarks such as salary and company revenue from Bank of Singapore and OCBC data. The relationship manager verifies and refines the draft before it goes to compliance. The bank says report writing time fell from 10 days to one hour, with fewer inconsistencies and omissions. The tool runs on the bank's private cloud.

  • Cycle time: 1 hours, per source of wealth report
    "Time taken to write the report has been shortened from 10 days to one hour, with greater accuracy and consistency."
    Claimed by: organization

Bank of Singapore

Singapore · Wealth and asset management · 2026

PilotGrade C

In 2026 Bank of Singapore began using HELIOS, an agentic AI platform that streamlines due diligence and credit risk profiles when onboarding high net worth and ultra high net worth clients. More than 100 relationship managers (about a quarter) in Singapore, Hong Kong and Dubai had begun using it. Its chief executive said roughly 50 clients had been fully onboarded through it. The bank aims to cut account opening from more than 30 business days to 15, which is a target rather than a result. The move follows a request from the Monetary Authority of Singapore to shorten account opening times for private bank clients.

  • Interactions handled: about 50, clients fully onboarded at time of reporting
    "Jason Moo, Bank of Singapore's CEO, said that roughly 50 clients have been fully onboarded through the platform, taking the HNW and UHNW segments together."
    Claimed by: organization

Deutsche Bank

Germany · Banking · 2026

ProductionGrade C

Deutsche Bank Private Bank put an agentic AI solution live in its Singapore and Hong Kong booking centres that researches, documents and prepares Source of Wealth assessments, which the bank calls one of the most resource intensive parts of know your customer checks. It reads client documents alongside approved external data, flags gaps and inconsistencies, and hands the assessment to bank staff for review; relationship managers in Dubai use it for accounts booked in Singapore. The bank stresses that accountability stays with its people. No outcome figures are disclosed; the growth figures in the coverage are forecasts.

No outcome disclosed.

How do you implement it?

A model agnostic playbook: what to prepare, the order to build in, and what goes wrong.

Data you need

  • Source of wealth policy and report template per client type and jurisdiction
  • Client documents in digital form, with a secure upload path
  • Benchmark data for plausibility checks (salaries, company financials)
  • Access to company registries, adverse media and screening results

Systems to integrate

  • Client onboarding or KYC case management system
  • Document management and secure upload
  • Screening providers for PEP, sanctions and adverse media
  • Company registry and financial data providers

Complexity: High

It touches a regulated financial crime control. It needs reliable document extraction across many formats and languages, approved external data sources, a clear template agreed with compliance, strict data security and a model risk review.

  1. 1

    Agree the report standard with compliance

    Write down what a complete source of wealth report contains for each client type (entrepreneur, heir, executive, investor) and which evidence each statement needs.

  2. 2

    Start with extraction and drafting

    Let the agent extract, reconcile and draft with citations to document pages, while all judgments stay with the relationship manager and analyst.

  3. 3

    Add corroboration sources carefully

    Connect approved external sources one at a time and record which source supported which statement; never let the model rely on its own general knowledge as evidence.

  4. 4

    Measure quality, not only speed

    Track compliance send backs, missing evidence and analyst edits alongside turnaround time, and compare with manually written reports.

  5. 5

    Keep data inside a controlled environment

    Client wealth documents are highly sensitive; process them in a private or dedicated environment with strict access control, as Bank of Singapore does on its private cloud.

Guardrails

  • Every statement in the report cites a document page or an approved external source
  • The model never decides the risk rating or the onboarding outcome
  • Gaps and inconsistencies are flagged, not smoothed over
  • Client documents processed in a controlled environment with access limited to the case team
  • No use of the model's general knowledge as evidence of wealth

KPIs to instrument

  • Elapsed time from complete document set to submitted report
  • Compliance send back rate and reasons
  • Share of report statements with a valid citation
  • Analyst edit rate per report section
  • Time from first contact to account opening

Human in the loop

The relationship manager verifies and refines every draft before submission; the compliance analyst challenges it and decides the risk rating; senior management approval applies for PEPs and other high risk clients as the bank's policy requires.

Common failure modes

Plausible but unsupported narrative
The draft reads well but a key wealth event has no evidence. Require a citation per statement and flag uncited text.
Extraction errors in complex documents
Amounts or dates are misread from scanned statements or foreign language filings. Show source snippets next to extracted values for verification.
Over reliance by reviewers
Analysts approve well written drafts with less challenge. Sample reports for independent review and track challenge rates.
Data leakage
Sensitive documents reach systems or models outside the controlled environment. Enforce data residency, masking and access control.

What are the risks and rules?

EU AI Act

Depends on design

Anti money laundering due diligence is not listed in Annex III, so an assistant that drafts source of wealth reports for a human decision is not high risk by default. It becomes high risk if it adds remote biometric identification of the client (Annex III point 1(a); verification that only confirms a claimed identity is excluded) or feeds an assessment of a natural person's creditworthiness, for example for lending to the client (Annex III point 5(b)). GDPR Article 22 on solely automated decisions applies if it ever refused a client on its own.

Guidance

Controls to put in place

  • Inventory entry and model risk review for the drafting and extraction components
  • Report template and evidence standard approved by financial crime compliance
  • Citation coverage check before a report can be submitted
  • Independent sampling of AI drafted reports by a second line reviewer
  • Data residency, encryption and access controls on client documents

Frequently asked questions

How much faster can source of wealth reports be?
Bank of Singapore says the time to write a report fell from 10 days to one hour with its Source of Wealth Assistant, with relationship managers verifying and refining each draft. Deutsche Bank has also put an agentic source of wealth solution live in Singapore and Hong Kong.
Does the AI decide whether a client is accepted?
No. It drafts and flags; the relationship manager, the compliance analyst and, for high risk clients, senior management decide. Bank of Singapore has relationship managers verify each draft before internal review, and Deutsche Bank says that while tasks can be automated, accountability remains with its people.
Can the AI verify wealth on its own?
It can check plausibility against benchmarks and approved sources and point out gaps, but every statement must be backed by a document or an approved external source, never by the model's general knowledge.

How to cite this page

Blits.ai AI Use Case Library, "AI agent for source of wealth due diligence in private banking", last verified 26 September 2026, https://www.blits.ai/ai-use-cases/source-of-wealth-diligence. Licensed under CC BY 4.0. Method: how we verify use cases.

Changelog
  • 27 September 2026: First published

Related use cases

BankingPayments and cards

AI for PEP and adverse media screening

AI that continuously scans news, court records, registries and other open sources in many languages for negative information and political exposure linked to customers, counterparties and beneficial owners, discards look alikes, and summarises credible risk for the analyst with the sources attached.

Deployments
7 public, best grade B
Reported handling time reduction
at least 60%
Save the Children, vendor claim
BankingPayments and cards

Dynamic AML customer risk rating with machine learning

Explainable machine learning that produces the money laundering risk rating itself: it computes and continuously updates each customer's rating from due diligence data, products, geography, behaviour and screening results, and shows which factors drive the rating and when enhanced due diligence is warranted.

Deployments
1 public, best grade B
Autonomy
Supervised agent
BankingPayments and cards

AI for perpetual KYC and event driven customer due diligence

AI that keeps each customer's due diligence file current by replacing calendar driven KYC reviews with continuous, event driven refreshes: it watches for trigger events such as a change of ownership, address, behaviour or a new adverse finding, refreshes the file automatically where it can, and involves an analyst only when something material has changed. The risk rating itself and the first file for a new business client are separate use cases.

Deployments
5 public, best grade B
Reported cost reduction
40%
JPMorgan Chase, organization claim
BankingPayments and cards

AI assistant for digital account onboarding and KYC

A customer facing AI assistant that guides a new applicant, a person or a small merchant, through a digital account, card or relationship application: it collects and checks identity and supporting documents, orchestrates the know your customer and anti money laundering checks, prefills what it can and sends only the unclear cases to a human reviewer with a summary. The ownership research for complex corporate clients is a separate back office job.

Deployments
3 public, best grade C
Reported productivity gain
30x
M-DAQ Global, vendor claim
BankingPayments and cards

AI for business onboarding (KYB) and beneficial ownership discovery

An AI agent that builds the know your business (KYB) due diligence file for a new or reviewed corporate client, before any account is opened: it collects registry, incorporation and ownership documents, resolves the entity across sources, maps the ownership chain through holding companies, nominees and trusts to the ultimate beneficial owners, screens the entity and its owners, and presents a risk scored case for a compliance analyst to decide.

Deployments
3 public, best grade C
Reported automation rate
25%
BNY, organization claim