AI baseline measurement

The point is the second measurement

A baseline measurement on its own tells you very little. Its whole purpose is comparison. If you never measure again, you have paid for a snapshot that will be out of date within a quarter. Organisations that get value from it decide up front when they will repeat it, usually after twelve months.

What it typically covers

• Current use. Which AI tools are already in the building, including the ones nobody approved
• Data quality. Not whether it is perfect, but whether the same term means the same thing in two departments
• Skills. Who can judge an AI output rather than only produce one
• Decision-making. Whether anyone has the mandate to say yes or no

Why the shadow use matters most

The finding that surprises directors most is the first one. Staff almost always use AI tools before any policy exists. A baseline that ignores this measures an organisation that does not exist. It also makes the case for an AI policy concrete rather than theoretical.

Baseline measurement and readiness scan

The two overlap but answer different questions. A baseline asks where are we now, so we can measure movement. A readiness scan asks what should we do next, and ranks the options. In practice a good AI Readiness Scan delivers a baseline as a by-product.

At Radical the free version of this is called the AI Photo.