The methodology ladder
Every engagement starts the same way — the Operational Intelligence Assessment, a fixed-scope look at one or two workflows. What we find determines which of the other nine methodologies actually applies. Nothing here is sold before it's earned by the Assessment's findings.
STEP 0 · FREE
Self-report, about four minutes — where your workflows likely sit.
M1 · $7,500 · 3 WEEKS
The one paid engagement everyone starts with.
M2–M10 below are scoped individually once the Assessment identifies which applies — none are priced publicly yet.
Fixed-scope evaluation of one or two workflows to determine whether, where, and how AI belongs. Intake → legibility capture → boundary and quadrant analysis → verification/economics screen → 90-day roadmap.
Architecture review focused on the one decision that governs a system's reliability — decomposition, boundary identification, and a re-drawn boundary with rationale.
Organization-level readiness across permissions, evaluation, provenance, legibility, and operating model — before agents scale.
Builds the client's evaluation infrastructure to continuous, with grader calibration against correlated-grader risk — the asset that survives every model migration.
Measures and closes the gap between AI governance as documented and as mechanically enforced.
Replaces governance theater with oversight that provably catches things, sized to the approval fatigue curve.
Makes every consequential automated decision reconstructable — the documentary foundation for regulated deployment.
Converts model and prompt churn into managed change inside the client's existing change-management structures.
Converts tribal and procedural knowledge into workflow contracts — machine-referenceable specs of the work, its exceptions, and acceptance criteria.
Quarterly portfolio discipline: what AI work actually costs per verified outcome, and where the portfolio should rebalance.
Not sure which one applies?
That's what the Assessment is for. Start there.