Sanctions screening tools fail when risk is hidden behind layers of shell companies — they check individual nodes but miss the paths between them. Sentries, a risk-intelligence system built on ArangoDB, models ownership, control, and family relationships as a graph and computes transitive risk propagation across up to three hops. Each relationship type carries a weighted edge (1.0 for direct ownership, 0.8 for leadership, 0.5 for family ties), and the traversal multiplies weights along each path to produce directRisk, inferredRisk, and totalExposure scores. Because ArangoDB stores entity records, relationship edges, and ontology rules in a single query context, no external application code is needed to join them. The same engine covers defense vendor vetting, financial KYC/AML with UBO roll-up, and supply-chain risk. Upcoming features include propagation-weighted PageRank for continuous background computation, temporal 'as of' queries for audit compliance, and entity resolution for transliterated aliases.

5m read timeFrom arango.ai
Post cover image
Table of contents
The gap sanctions lists can’t closeWhat Sentries doesWhy the math worksWhere this applies in the enterpriseWhat’s coming next“Tracing Hidden Risk” by Arthur Keen
2.3K Impressions