FinCEN has proposed revised customer due diligence requirements that would let banks rely on third-party sources and stop re-verifying customers each time they open an account. That sounds like relief. It isn't a transfer of responsibility. Institutions remain ultimately on the hook for CDD, even though some had hoped the ownership database created under the 2021 Corporate Transparency Act (CTA) would absorb part of that burden.
The practical shift for your program: the hard question moves from "how often do we collect this?" to "can we prove this source was sound?" This checklist covers the controls a BSA Officer should confirm now, before any final rule lands. It reflects a proposal, so details may change. Treat each item as readiness work, not a compliance deadline.
Prerequisites
Before you start, have these in hand:
- Your current CDD and beneficial owner identification procedures, including the account-opening workflow.
- An inventory of every external data source feeding customer records, including vendors, registries, and affiliate data.
- Your vendor management policy and any existing due diligence files on data providers.
- Your most recent independent testing results for the AML/CFT Framework.
- A named owner for the review (typically you or a delegate in compliance, with input from operations and technology).
Checklist
1. Inventory every third-party source used in a CDD decision.
Done means you have a written list, by source, showing which CDD element each one informs (identity, ownership, control, nature of business). Not done means sources are scattered across onboarding teams with no central record.
Good looks like: a single register that maps each source to the decision it supports and the person accountable for it.
2. Define a fitness-for-purpose standard for each source.
Reliance only works if "reliable" has a definition. Set criteria such as provenance of the data, how it is collected, how often it updates, and known error or coverage limits. Document the standard before you apply it.
Good looks like: a short, written test that a second reviewer could apply and reach the same conclusion.
3. Retain evidence that you assessed the source, not just that you used it.
This answers the central question: what proves reliance was appropriate? Keep the assessment itself, the date, the reviewer, the conclusion, and any limitations you accepted. A screenshot of the data is not an assessment.
Good looks like: an examiner can pull one customer file and trace from the data point back to a dated source assessment.
4. Record the reliance decision at the customer level.
For each file, note which elements came from a third party, which you verified directly, and why. Without this, you can't later show where a failure originated.
Good looks like: the Customer Risk Profile shows source attribution for each key data element.
5. Separate "unnecessary re-check" from "information needs revisiting."
Ending re-verification at every account opening removes duplication. It doesn't remove your duty to keep customer information adequate and current. Write down the specific triggers that override the no-recheck default: changes in ownership or control, adverse media, unusual activity, risk rating changes, or conflicting data between sources.
Good looks like: a trigger list in procedure, with each trigger tied to a required action and a documented outcome.
6. Reconcile reliance with your Periodic Review and Ongoing Due Diligence cycles.
A customer you verified once and don't re-verify at each new account still needs scheduled Periodic Review at an interval matched to risk. Confirm the two processes don't leave a gap.
Good looks like: high-risk customers still receive scheduled review regardless of how many accounts they hold.
7. Decide what role CTA ownership information will play, and cap it.
The CTA requires FinCEN to collect a business owner's name, date of birth, address, and an ID number. That is useful input. But a government-held record is one source, not a conclusion. Your procedure should treat it as data to corroborate against your own findings, not as the answer.
Good looks like: your procedure states that CTA-derived data is weighed alongside other evidence, and that discrepancies trigger follow-up rather than being waved through.
8. Build a discrepancy-handling process.
When a third-party record conflicts with what the customer told you, someone must resolve it and document how. Silence on conflicts is where reliance turns into liability.
Good looks like: a defined escalation path, a resolution record, and a decision on whether the conflict is itself a risk signal worth review for a Suspicious Activity Report.
9. Review vendor contracts and controls.
Check what your agreements say about data accuracy, update frequency, change notification, audit rights, and incident reporting. Contract language can't shift regulatory accountability, but weak terms leave you unable to evidence reliability.
Good looks like: contracts give you the information and access you need to support the reliance standard in item 2.
10. Add third-party reliance to independent testing and governance.
Make sure your independent test scope covers source selection, assessment records, and trigger handling. Brief senior management and the board on the accountability point so no one assumes the proposal reduces exposure.
Good looks like: testing workpapers sample reliance decisions, and management minutes reflect that accountability stays with the institution.
Common Mistakes
- Treating access as assurance. Having a data feed doesn't mean you've validated it.
- Documenting the outcome but not the reasoning. Files show the data was used; they don't show why it was trusted.
- Letting "no re-verification" become "no review." Reduced repetition is not reduced oversight.
- Treating a government database as a safe harbor. It is a source, and your obligation to understand and verify the customer remains.
- Updating policy without updating workflow. If onboarding staff keep running the old process, your documented standard won't match practice.
Next Steps
Because this is a proposal, start with the low-regret work: complete the source inventory (item 1) and draft the fitness standard (item 2). Both help you today and position you for any final rule. Then gap-assess items 3 through 10 against current practice, rank the gaps by customer risk, and assign owners with dates.
Consider submitting comments if your institution has operational concerns about how reliance and re-verification would work. Implementation details and safeguards will matter more than the headline, and practitioners are well placed to say what evidence standard is workable.
The takeaway for your team: fewer repeat checks can be a real efficiency gain, but only if you can show the sources behind your decisions were sound. The burden doesn't disappear. It moves from collecting data to defending it.





