How Lumen Works

Three phases. Continuous operation. From raw signal to actionable intelligence.

Collect

Multi-source ingestion

●40,000+ global media outlets via GDELT event streams
●Reddit community monitoring with account provenance analysis
●News API integration for breaking coverage
●15-minute collection intervals — continuous, not batch
→

Analyze

Behavioral intelligence

●Keyword-overlap clustering identifies narrative groupings
●5-signal coordination scoring distinguishes organic from orchestrated
●Wire-service deduplication prevents false positives from AP/Reuters syndication
●Source reputation weighting across mainstream, fringe, and state media
→

Report

Actionable intelligence

●Active week reports: threat assessment, attribution, response recommendations
●Quiet week reports: six OSINT frameworks scan for precursor signals
●NATO STANAG 2511 source credibility ratings on all citations
●PDF delivery with analyst review (Analyst tier) or AI-generated (Monitor tier)

Methodology

Intelligence community standards, applied to the private sector

Analysis of Competing Hypotheses

ACH

Developed by Richards Heuer at CIA. Multiple hypotheses are evaluated against available evidence. Rather than seeking confirmation, ACH systematically identifies which hypotheses are most consistent with — and least refuted by — the evidence. Lumen generates structured ACH matrices for every attribution assessment.

ICA Probability Standards

ICA

Intelligence Community Assessment standards define precise probability ranges: Almost Certainly (95%+), Very Likely (85–95%), Likely (70–85%), Roughly Even (45–55%), Unlikely (15–30%), Very Unlikely (5–15%), Remote (<5%). Lumen never uses ambiguous language — every confidence judgment maps to a defined probability range.

NATO STANAG 2511

STANAG

NATO Standardization Agreement 2511 provides a two-dimensional rating for intelligence sources. Credibility (A–F) rates the source itself; content rating (1–6) rates the specific information. Every source cited in a Lumen report carries both ratings, enabling consumers to weight intelligence appropriately.