FRAMEWORKS
The method behind the read.
Every engagement runs on published method, so a read you are handed can be checked against the framework that produced it. Three frameworks carry the practice. Everything under them is working method we use inside those three and publish because it is useful to a CTO or an operating partner on its own.
THE THREE FRAMEWORKS
PRISM™ for the read, CLEAR™ for the hold, and the AI Operating System at fund level.
PRISM™ — Technology Due Diligence Framework
v1.2Five dimensions scored 0–100. Every finding priced as CapEx requirement, EBITDA drag, or exit multiple implication. A financial instrument, not a checklist.
CLEAR™ — Hold Period Operating Framework
v2.0Five phases ending at exit, matching fund economics. PRISM™ diagnoses the asset; CLEAR™ operates it.
AI Operating System
v0.9One GP-level engagement that every current and future portfolio company inherits. Governance, readiness, and a repeatable deployment playbook: portfolio AI as operating capacity, not per-portco pilots.
METHOD LIBRARY
Working method, published in full and used inside an engagement rather than sold as one.
AI Cost Optimization Framework
A six-layer model of AI spend (API, infrastructure, tooling, people, waste, risk) that uses governance as the detection layer. Three of the six layers carry 55 percent of spend and are the least visible; working all six recovers 20 to 34 percent of AI spend in the first year.
AI Governance Program
The engagement that runs between the diagnostic and the target architecture: five phases that take a portfolio company from unlisted AI systems to a board-ready, regulator-ready governance posture.
AI Value Attribution Framework
A five-level measurement hierarchy (Cost, Adoption, Productivity, Business Outcomes, Enterprise Value) for verifying what AI investment produced. Most companies can evidence the first two levels and report as though they had evidenced the last two.
AI Value Creation Framework
Turning AI from a cost and distraction into a measurable margin driver. Use case prioritization, governance, and a board-ready AI narrative: all tied to EBITDA, not pilots.
Enterprise AI Control Plane
A unified governance architecture (5 pillars × 4 lifecycle stages = 20 control domains) that treats AI systems as enterprise operational infrastructure requiring identity, access, observability, and resilience controls.