Skip to content
Up to 4x faster fraud detectionForensic-Grade Method

Fight internal fraud with

Investigative Analytics
for Sensitive Data.

Members of:
Swedish-Polish Chamber of CommerceAI Chamber

Data lakes are data swamps.

The answers are in there. So is everything else: duplicate records, dead-end documents, signals scattered across systems that don't talk.

BCNN goes in, connects the dots across text and numbers, and comes back with source-linked findings and a graph your team can keep querying.

The Networks Notebook.

The Networks Notebook is BCNN's investigation engine. Four parts working together: a graph database, a BI engine for text and numerical data, GraphRAG, and a swarm of AI agents that build and verify hypotheses. It delivers evidence-backed answers, source-linked, ready to inspect and challenge.

Relationship Schema
Project ID: 8821-X
Swipe
EMAIL_LOGTRANSACTIONS.CSVOFFSHORE_REGARCHIVE_FILESREL: J. SMITHINTEL_REP_01.PDFSUBJ: A. VKOVTX: $4.2M [SWIFT]SHELL_CORP_LTDFIND_KYC_FLAG
Traced path
Flagged finding
Contextual data

The invoice, the payment and the email, analyzed together.

Patterns, anomalies and leads, out of chaotic data.

Confidentiality and compliance, GDPR and EU AI Act aligned.

First answers in 3 weeks.

Send us a representative sample of your dataset. In three weeks, we return evidence-backed analysis you can act on - and a clear path to what comes after: the full investigation and full monitoring.

01
Week 1

Scope & sample

We define the question, agree what "answered" looks like, and ingest a representative sample.

02
Week 2

Connect the dots

We extract entities, relationships, events, and transactions, and build the graph that links them.

03
Delivery

The first package

You receive a structured dataset, the relationship graph, an analysis report with source-linked findings, and access to our AI interface.

04
After the investigation

Scale and leave it watching

We scale to the full dataset, deepen the investigation, and leave the monitoring running: the patterns we mapped, watched inside your own infrastructure. We hold none of your data. You hear when something moves.

Ontology Map
1. Ingestion Sources
scan_004.pdf
legacy_db.sql
incident_report.docx
ERP & CRM
OSINT_report.pdf
behaviour_patterns.csv
financials.xmlPROCESSING
2. Semantic Mapping
Person BPerson ACompany XFRAUD RISK
3. Structured Output
Entity Detected
Fraud Pattern A
Confidence: High
Identity Resolution
Person A = Person B
Confidence: Verified
Action
Export to SQL
> Export to .CSV

Meet our founders.

Michał Domański

Michał Domański

CEO BCNN

Business and Operations

8 years building tech businesses, after scaling business development in pharma and IT. Works with European public institutions on AI adoption. TEDx speaker, featured in Business Insider, and ambassador for the Swedish-Polish Chamber of Commerce.

Paweł Gołąb

Paweł Gołąb

CTO BCNN

Technology and data analysis

Over 10 years with AI and investigative analytics systems, including air-gapped analytical software for a top intelligence agency. Lecturer at the Faculty of Mathematics, Informatics and Mechanics, University of Warsaw, and co-author of AI in the Polish Public Sector.

Case Studies.

BCNN works on datasets where mistakes have consequences: investigations, disputes, audits, compliance reviews, and sensitive decisions. Every finding links to the document, record, or relationship behind it.

Deploy where your data lives.

BCNN adapts to the client's infrastructure. We work within existing cloud environments, sovereign compute, private data centers, on-premise hardware, or fully air-gapped systems.

Commercial Cloud (BYOC)

Already using AWS, Azure, or Google Cloud? BCNN works within your existing environment, security model, and billing.

Azure / AWS / Google

Sovereign Data Centers

For projects requiring data residency or jurisdictional sovereignty, we deploy on dedicated private compute infrastructure.

Data residency support

On-Premise & Air-Gapped

For highly sensitive datasets, BCNN runs in restricted or fully offline environments. No data leaves the premises.

Fully offline