Passive Workflow Intelligence Engine
Passive telemetry intelligence that discovers recurring workflows, quantifies automation potential, and maps them to Claresia skills — all without surveys, interviews, or screen recording.
Shadow Observer is a lightweight, privacy-first telemetry scanner that reads local files already on every employee's machine — SQLite databases, JSONL session logs, git history, and OS indexes. It groups these events into activity sessions, detects recurring patterns using Jaccard similarity clustering, and classifies them into named business workflows. The entire inference pipeline is deterministic: no LLM calls, no hallucination risk, no token costs. Results include automation potential scores, estimated time savings, and direct mappings to Claresia's 56 deployment-ready skills.
| Dimension | Traditional Approach | Shadow Observer |
|---|---|---|
| Data Collection | Surveys, interviews, workshops | Passive telemetry (automatic) |
| Employee Disruption | High — pulls people from work | Zero — runs silently in background |
| Time to Insight | 4-6 weeks | 2 days |
| Accuracy | ~30% (self-reported bias) | 85%+ (behavioral data) |
| Cost Per Assessment | $50K-$150K consulting | $15K-$50K (automated) |
| Coverage | Sample of roles interviewed | Every employee with a device |
| Repeatability | One-time snapshot | Continuous monitoring available |
Install lightweight Node.js scanner on employee devices. Reads existing telemetry files — no agents, no hooks, no permissions dialogs.
Scanner reads 9 local data sources: SQLite DBs, JSONL logs, git history, OS indexes, API metadata. All processing stays on-device.
Deterministic 4-stage pipeline: sessionization, pattern detection, classification, scoring. Zero LLM calls. 82ms average latency.
Named workflows with automation scores, time savings estimates, and direct mappings to Claresia skills. Ready-to-deploy recommendations.