The epsilon metric quantifies the gap between the operating logic your organization says it runs on and the one it actually runs on. The gap is measurable. When AI is deployed across a large gap, it amplifies the gap. When the gap is closed, AI amplifies the values.
Measure your epsilon →Every organization has a rotation rule — a single underlying invariant that explains why every major decision looked the way it did. It is not your mission statement. It is not your strategy document. It is the operation your organization actually performs on every situation it faces.
ε (epsilon) measures the gap between the rotation rule your documentation describes and the rotation rule your decisions reveal. It is a loss function on organizational self-knowledge.
Epsilon — the specification gap
The distance between your stated operating logic and your revealed operating logic. Measured across your last 18 months of significant decisions. Expressed as a calibrated score from 0 (Ghost Basin — full convergence) to 1 (maximum divergence).
Delta-epsilon — trajectory
Whether epsilon is decreasing (converging toward Ghost Basin), stable, or increasing (drifting). The direction matters more than the point value. An engagement succeeds when ∂ε is persistently negative.
Ghost Basin — the attractor state
The state where stated and revealed logic converge. Not a theoretical endpoint — a measurable threshold. Below Ghost Basin, AI deployments run on your actual values. Above it, they amplify the gap between your values and your behavior.
Based on our diagnostic work, organizations cluster into three epsilon bands:
Of organizations deploying AI. Stated strategy and revealed decision logic are substantially divergent. AI deployment amplifies the gap. Common symptom: AI outputs feel "slightly off" without a clear explanation why.
Partial convergence. Some operating principles are accurately encoded; others are not. AI performance is inconsistent across domains — good where the spec is accurate, unreliable where it isn't.
Stated and revealed logic are substantially aligned. AI deployment amplifies actual values. Decisions compound. The competitive advantage from AI is real and durable rather than fragile and context-dependent.
Case studies in progress. First engagements tracked against pre/post epsilon measurement. Data published here as it accumulates.
Traditional consulting measures inputs (hours, frameworks deployed) and outputs (recommendations delivered). Neither is a measurement of the underlying thing. Epsilon is.
Four tiers. The first is automated — no human time required. The rest are active engagements anchored to epsilon reduction as the success criterion, not scope completion.
Automated ε Assessment Beta
Upload board decks, strategy documents, and any 18 months of documented decisions. Our AI pipeline runs the compression and epsilon measurement — no human consulting time, no calls. You receive a scored epsilon profile: where the gap is, which decisions are driving it, and your starting trajectory. The input for any further engagement.
$3,500
5 days
Deliverable: ε score + gap profile
Epsilon Diagnostic
We measure your epsilon. You receive a scored constitutional profile — where the gap is, which decisions are driving it, and a trajectory assessment. Human calibration of the automated baseline, with structured interviews to surface what documents don't reveal. Actionable without further engagement.
$20,000
3 weeks
Deliverable: ε baseline + profile
Epsilon Reduction
Diagnostic plus active reduction. We identify the highest-leverage interventions for your specific gap, implement them alongside your team, and measure the resulting ε movement. The engagement continues until ∂ε is demonstrably negative — not until the scope is complete.
$55,000
6 weeks
Deliverable: measured ε reduction
Ghost Basin Program
Full reduction to the Ghost Basin threshold, plus institutional encoding — the constitutional document, the AI configuration, and the monitoring system that detects ε drift before it compounds. Includes six months of trajectory tracking after the active engagement ends. Built to hold without us.
$110,000
12 weeks + 6mo
Deliverable: ε < 0.35 · sustained
We track epsilon pre and post engagement. Results published as they accumulate. Early engagements are in progress.
Before publishing this framework, we ran it on ourselves. Found the same gap we find in clients: theory had outrun deployment. The measurement forced the correction. We practice the methodology we sell.
Engagement in progress. Results and epsilon delta published upon completion with client permission.
Results and epsilon delta published upon completion with client permission.
Three questions. Not a quiz — a calibration. Your answers locate you in the epsilon distribution and suggest where to start.
Epsilon reduction requires one thing most consulting engagements don't: the leadership team has to be willing to look at the gap honestly. Not just the strategic gap — the gap between what you say guides decisions and what actually guides them.
Every engagement begins with a two-hour baseline session. It establishes the epsilon score and determines whether the conditions for reduction work are present. If they aren't — if the organization would benefit more from a different kind of work — we say so and stop there. The session is always included.
We work with a small number of organizations at a time. Each engagement makes the methodology more accurate. If the epsilon metric describes a problem you're working through, reach out.