Investigation
Trace entities, events & connections
Investigation is Nexma turned toward people, organizations, money, and events — and where they touch the physical world. You trace entities and the links between them across geography and time, resolve who is who, score risk continuously, and assemble a sourced profile in minutes instead of days. Fraud, sanctions exposure, anti-money-laundering, due diligence, and intelligence all run on the same surface.
It is the same operating system — Ontology, DataStore, Jax, and the engines — pointed at relationships instead of infrastructure.
What you can do
- Resolve entities across sources. People, organizations, and assets are matched and deduplicated across disparate feeds, so one real-world actor is one node — not five near-duplicates.
- Traverse the network. Follow ownership chains, shell structures, transaction patterns, and co-location to surface relationships no single record reveals.
- Score risk continuously. Proximity, relationships, and behavioral patterns roll into a live spatial risk score that updates as new data lands.
- Correlate spatially. Cross-reference entity locations against sanctions lists, risk zones, and regulatory boundaries on one map.
- Automate due diligence. Jax assembles a comprehensive entity profile from many integrations — a background check that would take an analyst days, returned in minutes with every source attached.
Core concepts
An investigation in Nexma is a typed graph, not a folder of documents. Entities and the links between them live in the DataStore, governed by the active Ontology, so a query can span people, companies, transactions, and places in a single statement — and every edge carries its source.
- Entity resolution is the foundation. Before any analysis is trustworthy, the same actor across sources must collapse into one node. Resolution runs against the Ontology's identity rules, not ad hoc string matching.
- The graph is queryable. "Who ultimately owns this account, and where are they?" is a traversal plus a spatial join — composed by Jax, returned with provenance.
- Geography is a first-class axis. Risk is rarely about an entity alone; it is about what the entity is near, who it transacts with, and which jurisdiction it sits in. Nexma keeps the spatial axis live.
A finding is only as defensible as its trail. Every node, edge, and score in an investigation carries the source it came from and the time it was observed, so a conclusion can be audited and reproduced.
How it works
You ask the question in plain language. Jax composes the traversal, the spatial join, and the risk roll-up against the world model, and returns an answer you can act on.
1Question: "Who controls this shell company, and is anyone in the chain sanctioned
2 or operating from a high-risk jurisdiction?"
3 1. Resolve collapse name and account variants to canonical entities
4 2. Traverse walk the beneficial-ownership graph to ultimate owners
5 3. Join cross-reference each owner against sanctions + jurisdiction layers
6 4. Score roll proximity, relationships, and flags into a risk score
7 5. Render a network view + a map, every node sourced and timestampedEvery layer keeps its sources and method, so the result is auditable and reproducible — not a black-box flag.
What an investigation surfaces
| Stage | What you get |
|---|---|
| Resolve | Deduplicated entities across all sources; canonical identities; confidence on every match |
| Traverse | Ownership chains, shell structures, transaction patterns, co-location and co-movement |
| Correlate | Sanctions and watchlist hits, risk zones, regulatory boundaries mapped to each entity |
| Score | Continuous spatial risk score driven by proximity, relationships, and behavior |
| Report | A sourced entity profile assembled by Jax, ready for review and handoff |
Example
A compliance team needs to clear a new counterparty before onboarding.
- Drop the counterparty's name and registration into the project; Jax resolves it against existing entities and external feeds.
- Ask: "Map beneficial ownership to the ultimate owners and flag any sanctions or high-risk-jurisdiction exposure."
- Jax walks the ownership graph, finds an intermediate holding company sharing an address with two sanctioned entities, and surfaces it on the map.
- The risk score updates; the finding is rendered as a network view with every edge sourced and a one-page profile generated for the file.
- The team forks the investigation into a branch to model a hypothetical ownership change and see how the risk score moves.
Swap the Ontology and the same traversal-plus-spatial-join pattern serves intelligence analysis, asset tracing, or licensing oversight. The graph mechanics are generic; the meaning comes from the world model.
Where to go next
- Ontology — where entity types, link types, and identity rules are defined.
- Spatial Analysis — the spatial joins and correlations investigation relies on.
- Object Detection — turn imagery and video into entities an investigation can trace.
- DataBase — reach into history to see how a network changed over time.
- Slides and Spreadsheets — turn a finding into a briefing or a report.