All posts

How AI Agents Help With UBO Identification

Identifying the ultimate beneficial owners (UBOs) of a business client is usually the slowest part of corporate KYC. Ownership runs through holding companies, trusts and nominee arrangements, often across jurisdictions whose registries hold different information. Fenergo's 2023 survey put the average KYC review for a corporate client at 95 days and $2,598 (Corporate Compliance Insights), and ownership analysis is a large part of that work.

Why it is hard

  • Layers. A UBO may hold a stake through two or three companies; the effective share is the product of each layer.
  • Different documents for each entity. Shareholder registers, trust deeds, certificates of incumbency and UBO declarations all describe a piece of the structure.
  • Gaps. A missing document for one holding company leaves the chain broken.

How Adient approaches it

Three agents share the work:

  1. Document processing reads every document, classifies it and extracts the people, companies and ownership relationships it mentions. It also flags documents that should be there and are not.
  2. Company structure builds an ownership graph from those relationships, calculates effective ownership through each layer and marks who reaches the UBO threshold.
  3. Sanctions screening then checks every person and company in that graph against sanctions lists.

The analyst sees the result as a diagram with the missing documents next to it. If the agents got something wrong, for example two people with similar names, the analyst adds a note and the report takes it into account.

The aim is not to remove the analyst from UBO identification. It is to hand them a mapped structure and a list of open questions instead of a stack of PDFs.

See a complete case

We walk you through a case from start to finish: documents in, ownership mapped, registries and sanctions checked, draft report, analyst review and final report.