Topograph

What “authoritative” means for company data

Hugo Dellinger
Author

A company's status, a director's name or an ownership percentage is not evidence on its own. A compliance analyst signing off on a company checks its legal status, registered address, directors and owners. For each of these datapoints, the question is not only "what is the data?" but "which official register says so, and when was the data retrieved?"

That is the current bar for data that holds up under a regulator's scrutiny: AML authorities expect customer due diligence to be based on a reliable, independent source.

That kind of data has a name. Vendors, compliance teams and regulators are starting to call it "authoritative". So it's worth being precise about what the term means.

What makes data authoritative

Authoritative data has three properties:

  • It is fetched directly from the official register.
  • It is fetched at the time the verification takes place.
  • It is traced to its source: name of the register and retrieval date attached.

"Authoritative" isn't a judgment on how good the data is. It isn't a scale either: data can't be more or less authoritative than other data. Data is authoritative, or it isn't. It's about source traceability: being able to show where each fact came from, and when. A copy collected earlier can't do that: it shows what the register held on the day it was collected, not what the register holds today.

Why it matters now

Compliance now asks two questions: "do you know who you're dealing with?" and "how do you know?" Regulators increasingly ask for the evidence behind a decision, not just the decision. On October 1st 2026, the EU Anti-Money Laundering Authority (AMLA) published its final draft standards on customer due diligence. They would require firms to take risk-sensitive measures to ensure that the sources of information they use for customer due diligence are reliable and independent. A screenshot of a stored record doesn't answer that question. Data fetched from the register at the time of the request does, because its source and timestamp show where it came from and when.

It also matters because not everything built from register data is authoritative. Beneficial owners reconstructed by tracing a company's shareholder chain are a calculation, not a record the register holds. Shareholders extracted from filed documents with AI are an interpretation of those documents, not data the register publishes. Both are useful, but neither is “authoritative” under the meaning presented here. A KYB file that’d present those datapoints as if they were authoritative blurs the line a reviewer needs to see.

Authoritative data keeps each fact tied to where it came from. That keeps the line visible and makes the final record auditable.

How to tell if data is authoritative

The label is easier claimed than achieved, so it helps to test it. Here are five questions to put to any source of company data, including your own:

  1. Where did this field come from? Every value should name the register it was fetched from, not just "our database."
  2. When was it retrieved? Each record should carry a retrieval date, and you should be able to query the register again at any time.
  3. Can I prove it came from the register? Data fetched from the register, with its source and timestamp, is proof in itself, with no extra document needed. A register extract can back it up but isn't required. A copy supplied by the company being checked is weaker evidence.
  4. What happens when the register doesn't publish something? An honest record says "not published by this register" rather than leaving a blank that could mean anything.
  5. Are different sources kept apart? Register data and data reconstructed or extracted from it should be shown and labeled separately, not merged.

Registers also differ. What one country publishes in full, another publishes in part or not at all, and privacy rules can redact personal details. Knowing what a register actually returns is part of using its data well.