Method

One ledger. Every position scored.

An assessment breaks a company into positions — the activities its revenue actually stands on — and scores each one for AI exposure. The result is a ledger, not an opinion.

The unit of account

Positions, not departments.

Org charts hide exposure. A "legal team" is not exposed; the document review it performs is. We decompose the business into positions — discrete activities with attributable revenue or cost — and assess each on its own terms.

A mid-size services company typically holds 20–40 positions. Each enters the ledger with its share of revenue, its dependence on judgment versus repetition, and the evidence for both.

Position
A discrete activity with attributable revenue or cost. The unit everything else is scored against.
Exposure
The degree to which a position is substitutable by current or near-term AI capability.
Cover
What defends a position: regulation, relationships, physical presence, accountable judgment.
Option
Upside a position holds if AI is adopted deliberately. Priced, not hyped.
The M-scale

Calibrated. Comparable. Defensible.

Every position and every company lands on the same five-point scale. The DNA is reused from due diligence: scores mean the same thing across companies, sectors, and years — and they hold up in front of an investment committee.

M1
Proofed
Service survives the shift as-is; moat holds.
M2
Resilient
Marginal exposure; efficiency upside dominates.
M3
Contested
Material positions exposed; options exist on both sides.
M4
Exposed
Core positions substitutable; window to act is open.
M5
Undefended
The moat is the thing being automated.
Process

Four weeks, ledger to statements.

The assessment runs on documents and structured interviews — no system access required. Findings are stated with their evidence; where evidence is thin, the ledger says so.

Week 1
Decomposition. The position ledger is drafted from financials and operating documents.
Weeks 2–3
Scoring. Each position is assessed against current capability, with interviews to test cover.
Week 4
Rendering. The ledger is written into the three statements, each for its reader.
Thereafter
Reassessment on a fixed cycle. Scores move when the evidence moves.
An assessment produces your first statement.