Publication: OSINT Maturity Model — building the capability in a company (from ad-hoc to a Centre of Excellence)
OSINT Maturity Model (AgentD22 Data solutions v2.1):
- Level 1 — ad-hoc: one-off requests, spreadsheets, manual searching in Google and registries. No processes, tools or budget. Risk: errors, gaps, illegal methods
- Level 2 — structured: standardised checklists, subscriptions to 3-5 commercial databases (Spark, Kontur, World-Check), a dedicated analyst. SLA per request: 2-3 days
- Level 3 — automated: RPA and scripts for routine collection, API integrations, an internal knowledge base (knowledge graph), ML verification of the first 50% of facts. SLA: 4-6 hours
- Level 4 — intelligence-driven: continuous monitoring (sanctions, adverse media, competitors), predictive alerts, LLM-augmented research (RAG), a dedicated OSINT team (5-15 people). SLA: real time
- Level 5 — Centre of Excellence: an internal product or platform, exporting expertise (consulting, training, SaaS), R&D (own models, proprietary datasets), thought leadership. Examples: AgentD22 Data solutions, major banks, Big 4 firms, intelligence services
KPIs by level:
- Coverage: the share of critical decisions supported by OSINT input
- Accuracy: confirmed facts as a share of all facts
- Speed: median time to insight
- Cost per insight: dollars per actionable intelligence unit
- Compliance: zero violations of GDPR, data protection law or ethics
Centre of Excellence structure (Level 5): Head of OSINT → research leads (geopolitical, corporate, cyber, crypto) → senior analysts → junior analysts → ML engineers → platform engineers → legal and compliance officer.
Budget (indicative, 2026): Level 2: 0K-150K per year; Level 3: 00K-500K; Level 4: 00K-1.5M; Level 5: 1M+.
Download the model (PDF plus Excel scorecard): pdf | xlsx
← Back to news