Ownership graph
Find out who really owns the company, and whether that person is sanctioned.
AdverseMe walks the ownership chain up to 6 levels, screens every owner and officer it finds, and computes the OFAC 50 percent rule across the whole structure.
The ownership graph traces who owns and controls a company, up to 6 levels of ownership above the subject and 2 levels of connected entities around it. Every owner and officer is screened against sanctions lists, effective ownership is multiplied along each chain, and the OFAC 50 percent rule is computed rather than eyeballed.
- Ownership depth
- 6 levels up
- Network depth
- 2 levels around
- OFAC 50 percent
- Single and aggregate
- Edge honesty
- Registry solid, reported dashed
How it works
Registry edges in, an exposure verdict out
- 1
Officers and owners are pulled from registries
For a company subject the engine reads registry data such as UK Companies House officers and persons with significant control, plus the openly available registers it can reach for the jurisdiction. Each edge is stored with its provenance: registry, filing, vendor or analyst extraction.
- 2
Every associated party is screened
Directors, officers and beneficial owners are batched against the OpenSanctions consolidated dataset and against the 13 official list feeds AdverseMe parses itself. A sanctioned director or owner used to pass silently. Now it is a finding on the subject.
- 3
Chains are walked and percentages multiplied
A recursive traversal walks up the ownership edges to depth 6, cycle-safe, multiplying the minimum of each ownership band along the way. When a registry only gives a band such as 50 to 75 percent, the floor is used, so an exposure figure is a provable lower bound, never an estimate.
- 4
The OFAC 50 percent rule is decided
If a sanctioned owner reaches 50 percent or more, alone or in aggregate with other sanctioned owners, the subject is flagged as ownership-exposed. Chains that land between 25 and 50 percent on band minima are flagged as possible exposure with the maximum shown, so an analyst knows whether 50 percent is even attainable. An unknown percentage on any hop is reported as unknown, not guessed.
- 5
Risk propagates through the graph
Each node gets a badge: red for sanctioned or controlling exposure, amber for PEP or elevated propagated risk, yellow for adverse media, grey for nothing found. Propagated risk flows along edges so a clean shell sitting under a sanctioned parent does not look clean.
What you see in the app
The Structure tab: an interactive graph with honest edges
The Structure tab on a company result renders the graph with React Flow. Solid grey edges are registry or filing records. Dashed amber edges are "reported": the analyst model extracted them from text, and the tooltip says so explicitly, with the source and the quote where one was captured.
Edge labels show the percentage and its basis: exact, a band, or unknown. A lock ring on a node marks OFAC control. Node badges follow the red, amber, yellow, grey scale so the eye goes straight to the problem.
Above the graph, the exposure panel states the verdict in words: blocked, possible or clear, with the aggregate minimum percentage and the owners that produced it. Under a person result, the same data appears as a radial relationship view of officers, owners and prior screenings.
Why it matters to a regulator
A percentage with a provenance
Ownership findings are where screening tools most often overstate. A registry band of 50 to 75 percent is not 62.5 percent, and an analyst's reading of a news article is not a shareholder register. AdverseMe keeps the two apart in the data model and in the pixels: band floors are what get multiplied, and anything reported by a model is dashed, amber and labelled.
The exposure verdict is a computation you can show: the chain, the effective minimum, whether 50 percent was reached on its own or in aggregate, and the lists the owner appears on. That is the working an examiner asks for when a blocked-property determination is challenged.
Officers and owners are screened as a deterministic channel written by the engine, not copied from the analyst's narrative, so the associated-party hit in the report is the same one the council scored. See Engine Trace for the full record.
Frequently asked questions
What is the OFAC 50 percent rule and does AdverseMe apply it automatically?
Under OFAC guidance, an entity owned 50 percent or more, directly or indirectly, by one or more blocked persons is itself blocked. AdverseMe multiplies ownership percentages along each chain and adds sanctioned owners together, then flags definite exposure at 50 percent and possible exposure between 25 and 50 percent on band minima.
Where does the ownership data come from?
From the registries the engine can reach for the jurisdiction, such as UK Companies House officers and persons with significant control, plus filings and vendor data where configured, plus what the analyst model extracts from documents. Every edge carries its source, and analyst-extracted edges are drawn dashed and labelled as reported, not as a registry record.
Why 6 levels?
Six levels covers the layered holding structures seen in practice while keeping the traversal bounded and cycle-safe. Connected entities around the subject, such as shared directors, are explored 2 levels out so the graph stays readable.
What if the ownership lane fails?
It says so. Each lane records whether it completed, and the result page shows an amber notice with the reason when ownership tracing did not finish. An empty graph and a broken graph never look the same.
Trace the next company to its owners.
Plans start at $49 a month for 50 screenings. Every plan includes the AI council, the ownership graph, Engine Trace and PDF reports.