Protecting Korean defense data sovereignty while accelerating decisions 10×.
Palantir Foundry-class intelligence with deep Korean-domain fluency. Full data sovereignty at a fraction of the cost.
On-prem · Sovereign · Korean-first
Four challenges in defense data
Challenges to resolve before Korea
reaches the top four defense exporters.
Data Silos
Procurement, patents, quality, military HQs, DAPA, and ADD span six or more separate systems. Integrated analysis takes days to weeks.
Supply Chain Blind Spots
Collaboration between innovation firms, prime contractors, and government programs is not systematically mapped. Risk identification and alternative supplier discovery remain manual.
Decision Latency
Even simple queries require multiple report lookups and database calls. Multi-hop relationship analysis is practically infeasible.
Dependence on Foreign Platforms
Beyond Palantir, no comparable alternative exists. Data sovereignty concerns mount as domestic firms place sensitive data in foreign platforms.
Our answer
Rebuilding Korean defense intelligence
across six capabilities.
From OSINT automation to Grafana operational dashboards — one continuous flow.
Supply Chain Intelligence
Every relationship across innovation firms, prime contractors, government agencies, and technology domains in one graph. Identify risk nodes instantly via natural language.
Technology Gap Analysis
Assess technology holdings by domain. Automatically match required technologies for specific weapon systems to organizations that hold them.
Public Procurement Analysis
Unify open tender data into a graph. Explore award patterns, bidding trends, and emerging sectors through natural language.
Patent & IP Mapping
Connect public patent databases to the company-technology graph. Analyze competitive advantage, map rivals, and surface licensing opportunities.
OSINT Methodology
Open-source intelligence becomes
your most powerful asset.
web intelligence and open-source collection — when systematically automated — surfaces patterns that human analysts miss. OntoCraft connects OSINT pipelines directly to the ontology layer.
Collect
Automatically crawl disclosures, news, patents, tenders, SNS, and web pages. Ingest both predefined sources and dynamic search results.
Crawler · API · RSS · OCR
Process
Noise reduction, deduplication, language normalization, and entity extraction. Korean NER identifies companies, individuals, institutions, and technologies. SNS, disclosures, and security threat signals are also analyzed.
Korean NER · SNS analysis · Deduplication
Link
Match extracted entities against the existing knowledge graph. Auto-generate new relationship candidates with confidence scoring.
Graph matching · Vector similarity
Verify
Cross-source validation, temporal consistency checks, and confidence weighting. A human verifies the final step (HITL).
Cross-source · HITL · Sandbox
Operate
Real-time visualization, alerting, and anomaly detection via Grafana dashboards. An actionable flow for operators.
Grafana · Prometheus · Alert
Weapon System Lifecycle DB
From acquisition to decommissioning —
managed in one graph.
Unify acquisition timing, maintenance requirements, remaining service life, and consumable part status for each weapon system in a single knowledge graph. Decision-makers instantly see what is needed and when.
Acquire
Acquisition schedule per system · procurement history · contract pricing
Operate
Maintenance cycle · availability rate · unit deployment
Sustain
Remaining service life · consumables · obsolescence risk
Retire
Decommissioning schedule · replacement candidates · asset recovery
TRL Assessment Framework
Objective technology readiness assessment
enables sound decisions.
TRL (Technology Readiness Level, levels 1–9) — assess technology maturity and attach those ratings directly as knowledge graph node attributes. Visualize the gap between self-assessment and external evaluation to support credible decision-making.
Coordination with IDS TRL Assessment Association — under review
TRL
1
TRL
2
TRL
3
TRL
4
TRL
5
TRL
6
TRL
7
Commercially viable
TRL
8
TRL
9
TRL 7 and above are classified as commercially viable stages
Synthetic Data
When real data is scarce,
we generate synthetic scenarios.
In defense environments, real data is often classified or simply unavailable. OntoCraft generates synthetic data for validation and simulation based on domain ontologies, compatible with TTA synthetic data standardization efforts.
Validation
Algorithm validation when real data is insufficient
Simulation
Synthetic scenario data generation
Security-friendly
Replace sensitive information with synthetic equivalents
TTA standard compliant
Follows domestic synthetic data standards
IP · Standards
Consolidating our technical advantage
through standards and patents.
ISO Standardization in Progress
Data Model Based on International Standards
ISO 42001 (AI Management Systems) · ISO/IEC JTC 1 SC 31 (Identification and auto-ID) · ISO 11179 (Metadata) · ISO 27001 (Information security) — domain ontologies aligned with international standards. Also coordinating with TTA detailed-level standardization work.
Patent Filing in Preparation
Korean Foundry Paradigm IP
GraphRAG-based Korean natural-language query routing, automated knowledge-extension merge pipeline, and domain dictionary auto-construction mechanism — patent filings are being prepared for these core algorithms.
Positioning
Entering the space no one else occupies.
Korean Foundry paradigm · Korean-language depth · SME-accessible pricing — no company has delivered all three. Until now.
Palantir Foundry
Strengths · Global standard · proven paradigm · deployed with US military and NATO
Limitations · USD 10M+ · limited Korean-language capability · data sovereignty concerns (ITAR/EAR)
Hanwha Systems C4I · LIG Nex1
Strengths · Track record supplying tactical systems to the Korean military
Limitations · Limited data analytics and AI capabilities · no intelligence platform offering
OntoCraft / ONTOFLOW
★ Our position
Strengths · Korean Foundry paradigm · 100% Korean NL · cost-effective pricing · on-premises support
Limitations · Early-stage company — building reference cases
Data Sovereignty · Security
Critical data never crosses a border.
On-premises deployment
Full stack with a single Docker Compose command · air-gap environment compatible
Local LLM
In-house LLM via Ollama · zero external API calls possible
JWT + RBAC
6 roles · 65 permission grants · MFA · Keycloak SSO
Audit logging
Full API call logging · complete change history tracking