Financial crime investigation means joining transactions, customers, counterparties, documents and external data into a picture that a compliance officer can act on and a regulator can follow. Octostar is an investigative intelligence platform that does this on one knowledge graph: transaction monitoring alerts, KYC files, adverse media and corporate registries resolve into the same entities, and analysts investigate them with search, link analysis and an AI assistant on one screen.
The same platform serves law enforcement financial crime teams, which is where its data model and its evidential standards come from. It deploys on-premise or in a private cloud, so customer data stays within your regulatory perimeter.
The platform overview describes the underlying architecture.
Where compliance programmes lose time
Alerts without context
Transaction monitoring produces alerts, but the context needed to close them lives in KYC files, case history, registries and news. Investigators rebuild that context by hand for every alert.
Entities that do not match
The same customer, beneficial owner or counterparty appears under different names and identifiers across core banking, onboarding and third-party data. Networks stay hidden until entities are resolved.
Narratives that take days
Suspicious activity reports and enhanced due diligence files are written by hand from many sources. Quality varies and turnaround is slow.
Models the regulator cannot follow
AI that cannot explain a score or cite its evidence adds model risk instead of reducing compliance risk.
From alert to evidence-based decision
Platform capabilities as applied to AML, fraud and due diligence work.
Entity resolution and network analysis
Customers, accounts, counterparties, beneficial owners and devices resolve into one graph across internal systems and external registries, so rings, mule networks and hidden ownership become visible.
Search across every dataset
Sub-second search over transactions, KYC documents, correspondence, adverse media and sanctions data at petabyte scale, grouped by entity rather than by source system.
Risk scoring, alerting and anomaly detection
Rules and models score entities and raise alerts on new patterns, with the underlying evidence attached. Value and risk models are configurable by your own analysts.
AI-assisted investigation
The AI assistant summarises a case, extracts entities from documents, answers questions over the graph and drafts narratives for reports. Every statement cites its source records.
Due diligence and KYC workflows
Enhanced due diligence, onboarding review and periodic refresh run as structured workflows with third-party data pulled in on demand through semantic drivers.
Dashboards and decision intelligence
Interactive dashboards, predictive analytics and AI classification give compliance leadership a live view of exposure and throughput.
Collaboration across lines of defence
Shared cases, role-based access and structured hand-offs let first-line, compliance and investigation teams work on one record with a full history.
Audit and model governance
Every query, view, score and AI output is logged with provenance, supporting internal audit, model risk management and regulatory examination.
Deployment inside your perimeter
- Step 1
Proof of concept on real data
An Octobox appliance or a private-cloud instance runs the full platform on a sample of your data within a day, with no data leaving your environment.
- Step 2
Connect core and external systems
Core banking, transaction monitoring, onboarding, case management and third-party data are mapped onto the ontology. Federate or ingest per source.
- Step 3
Operate and extend
Run on Kubernetes on-premise or in your private cloud. Your teams tune models, add data sources and build APPs on the open API.
Why financial institutions choose Octostar
- Built for investigation, not only monitoring: the platform comes from law enforcement and intelligence work, where evidence standards are strict.
- One graph across internal and external data, with entity resolution that surfaces networks rather than isolated alerts.
- AI that cites its sources and is fully logged, aligned with the EU AI Act and with OECD and NIST guidance.
- Sovereign deployment on-premise or in a private cloud, keeping customer data inside your regulatory perimeter.
- Open API and APPs ecosystem, so your own teams and integrators extend the platform.
- Petabyte scale with sub-second search, so years of transaction history are available to every investigation.
Frequently asked questions
- Does Octostar replace our transaction monitoring system?
- No. Octostar takes alerts and data from transaction monitoring, onboarding, case management and external providers and gives investigators one place to resolve entities, explore networks and build the evidence for a decision. Existing monitoring and reporting systems remain in place.
- Which use cases does it cover?
- Anti-money laundering investigation, screening alert review, fraud investigation, enhanced due diligence and KYC refresh, and internal investigations.
- How does the AI support regulatory examination?
- Every AI-generated summary, extraction or score is linked to the source records it was derived from and written to the audit log. Analysts confirm outputs before they are used. The platform's AI governance is aligned with the EU AI Act, OECD, NIST and Council of Europe frameworks.
- Where does the data reside?
- In your own data centre or private cloud. Octostar deploys on Kubernetes on-premise or in a sovereign cloud, and the Octobox appliance offers a fully self-contained option. No component sends data to an external service.
- Can we connect third-party data providers?
- Yes. Corporate registries, adverse media, sanctions lists and other commercial datasets connect through on-demand semantic drivers and are mapped directly onto the entity graph.