The Data and AI Readiness Framework

DARF exists to answer one question in two parts: can you trust the answer, and can you trust the system to act on it? Most of what gets sold today as “AI readiness” blends those into a single score. They are two different failure modes, and an organization can be excellent at one while being dangerous at the other.

The two pillars are sequential, not competing. An organization can have a flawless semantic layer sitting on zero operational readiness, which is great answers that nothing is allowed to safely act on. Or governed execution wired to a semantic layer nobody trusts, which is fast action on the wrong thing.