Prismatic OSINT Business

Overview

Prismatic OSINT Business provides adapter pattern implementations for business and financial intelligence sources, integrating with corporate registries, sanctions screening databases, financial disclosure systems, and business intelligence platforms to build comprehensive entity profiles for due diligence and risk assessment. Each adapter normalizes source-specific data formats into the platform's unified entity resolution schema, enabling cross-source correlation through the knowledge graph.

The application implements multi-registry company verification, cross-referencing entity identifiers (ICO, DUNS, LEI, registration numbers) across Czech, European, and global corporate registries to build verified entity profiles with confidence scoring derived from the platform's NABLA epistemic framework. When multiple registries provide conflicting data about an entity's status, ownership, or address, the system preserves both signals following the contradiction preservation axiom rather than arbitrarily choosing one source over another.

Sanctions screening integrates OFAC, EU consolidated, and UN sanctions lists through a unified screening API that supports fuzzy name matching, alias resolution, and date-of-birth verification for natural persons. Corporate structure mapping traces parent-subsidiary relationships across jurisdictions, enabling risk propagation analysis where a sanctioned parent entity's risk status cascades to its subsidiaries with distance-weighted risk score attenuation.

Architecture

Query Router --> Source Adapter --> Response Parser --> Entity Normalizer --> Profile Store
      |               |                   |                    |                    |
  Multi-Registry   ARES/Justice       JSON/XML/HTML        Unified Schema      PostgreSQL
  Parallel Query   Open Corporates    Normalization        Cross-Reference     Knowledge Graph
  Cost Optimizer   OFAC/EU/UN         Validation           Confidence Score    Meilisearch
      |               |                   |                    |                    |
      +---------------+-------------------+--------------------+--------------------+
                                       |
                                Sanctions Screener --> Fuzzy Name Matcher
                                       |
                                Structure Mapper --> Ownership Graph

All parsing and entity normalization follows pure function principles. Network requests execute in isolated processes under the OTP supervision tree with rate limiting and circuit breaker patterns per source. Source credential management handles API key rotation and OAuth2 token refresh automatically.

Key Modules

Key Features

Company Intelligence

Multi-registry verification ensures that entity data is corroborated across independent sources before being established as fact:

defmodule PrismaticOsintBusiness.QueryRouter do
  @spec verify_entity(keyword()) :: {:ok, VerifiedEntity.t()} | {:error, term()}
  def verify_entity(opts) do
    sources = Keyword.get(opts, :sources, [:ares, :justice, :open_corporates])
    identifier = extract_identifier(opts)

    results = Task.async_stream(sources, fn source ->
      {source, query_source(source, identifier)}
    end, max_concurrency: length(sources), timeout: 10_000)
    |> Enum.reduce(%{}, fn {:ok, {source, result}}, acc ->
      Map.put(acc, source, result)
    end)

    case EntityNormalizer.merge_and_score(results) do
      {:ok, entity} ->
        {:ok, %VerifiedEntity{
          entity: entity,
          sources: Map.keys(results),
          consistency: compute_consistency(results),
          discrepancies: detect_discrepancies(results),
          confidence: compute_confidence(results)
        }}

      {:error, reason} ->
        {:error, {:verification_failed, reason}}
    end
  end
end

• Multi-registry company search and verification across Czech (ARES, Justice.cz), European (Open Corporates), and global sources
• Officer and director identification with cross-registry verification and historical appointment tracking
• Corporate structure mapping tracing parent/subsidiary relationships across jurisdictions with ownership percentage tracking
• Filing and disclosure tracking including annual accounts, ownership changes, and regulatory submissions

Financial Intelligence

• Sanctions screening across OFAC, EU, and UN lists with fuzzy name resolution and alias matching
• PEP (Politically Exposed Person) identification through regulatory and public source cross-referencing
• Adverse media monitoring for negative news, regulatory actions, and enforcement proceedings
• Financial statement analysis with key ratio extraction and temporal trend detection

Sanctions Screening Engine

The sanctions screening engine implements multi-list fuzzy matching with configurable thresholds and evidence-based result classification:

Czech Business Sources

• ARES integration (business registry) with full entity profile extraction and real-time availability
• Justice.cz (commercial register) access for corporate filings, board composition, and insolvency status
• RZP (trade license register) for professional qualification and activity verification
• CEDR (subsidy tracking) for government grant and support payment monitoring

Cross-Registry Verification

• Entity identity verification through multi-source cross-referencing with weighted confidence scores
• Automated detection of inconsistencies between registry sources with audit trail documentation
• Entity resolution algorithms handling name variations, transliterations, and legal form differences
• Telemetry emission for verification latency, match rates, and source utilization metrics

Risk Propagation Through Ownership Graphs

When a sanctioned entity is identified, the risk score propagates through the corporate ownership graph with distance-weighted attenuation:

defmodule PrismaticOsintBusiness.RiskPropagator do
  @spec propagate(String.t(), float(), keyword()) :: {:ok, list(map())} | {:error, term()}
  def propagate(entity_id, base_risk, opts \\ []) do
    max_depth = Keyword.get(opts, :max_depth, 5)
    attenuation = Keyword.get(opts, :attenuation, 0.7)

    graph = PrismaticGraph.ownership_subgraph(entity_id, depth: max_depth)

    propagated = traverse_and_score(graph, entity_id, base_risk, attenuation, 0)

    {:ok, Enum.map(propagated, fn {id, score, path} ->
      %{entity_id: id, propagated_risk: score, distance: length(path), path: path}
    end)}
  end
end

Integrated Sources

Usage

# Full company profile from multiple registries
{:ok, profile} = PrismaticOsintBusiness.company_profile(ico: "12345678")
# => %{name: "Example s.r.o.", registries: [:ares, :justice], confidence: 0.96}

# Sanctions screening with fuzzy matching
{:ok, result} = PrismaticOsintBusiness.sanctions_check("Entity Name",
  lists: [:ofac, :eu, :un], threshold: 0.85)
# => %{matches: [...], clear: true, screening_id: "scr_..."}

# Cross-registry entity verification
{:ok, verified} = PrismaticOsintBusiness.verify_entity(
  name: "Example s.r.o.",
  ico: "12345678",
  sources: [:ares, :justice, :open_corporates]
)
# => %{verified: true, consistency: 0.94, discrepancies: []}

# Corporate structure discovery
{:ok, structure} = PrismaticOsintBusiness.corporate_structure(ico: "12345678")
# => %{parent: nil, subsidiaries: [...], ownership_depth: 3, total_entities: 12}

# PEP screening with evidence
{:ok, pep_result} = PrismaticOsintBusiness.pep_check("Person Name",
  jurisdiction: :cz, include_family: true)

NABLA Compliance

Testing

Company profile tests verify multi-registry data merging, entity normalization accuracy, and confidence score computation against known entity fixtures. Sanctions screening tests verify fuzzy matching accuracy, false positive rates, and list update handling. Corporate structure tests verify ownership chain traversal, risk propagation attenuation, and cycle detection in ownership graphs.

Integration tests exercise the full pipeline from multi-registry query through entity normalization to profile storage. Property-based tests generate random entity names with variations to verify fuzzy matching stability and consistency.

Integration Points

Performance

Telemetry events: [:prismatic, :osint_business, :profile_built], [:prismatic, :osint_business, :sanctions_screened], [:prismatic, :osint_business, :entity_verified].

Related Resources

• Prismatic OSINT Core -- Core OSINT infrastructure shared across all source adapters
• Business Financial Intelligence Specialist -- Coordinates business intelligence analysis workflows
• Czech Business Intelligence Specialist -- Specialized Czech business data analysis and verification
• Competitor Researcher -- Competitive intelligence leveraging business source data
• Intelligence Synthesis -- Multi-source business evidence fusion for entity profiling
• NABLA Axioms -- Contradiction preservation and confidence scoring for business data
• Quality Gates -- Data quality validation enforcement on business intelligence

Connect & Contribute

Created by Tomáš Korcak (korczis) | Open Source under GHL

• GitHub | GitLab | LinkedIn | Contact
• Developer Portal | Architecture | Meet the Creator