Score the operating habit, not the hype.
Every score has a useful next step.
Build the foundation.
Choose one responsibility, define approved information, and add a visible human review step.
Make it repeatable.
Turn good individual experiments into role-specific workflows the team can use and improve.
Scale from proof.
Rank deeper integrations and automation by value, feasibility, risk, and adoption.
Keep the people responsible.
Important facts, exceptions, relationships, approvals, and final accountability stay with authorized people.
Ten signs that useful AI can survive Monday morning.
The assessment covers work selection, approved tools, data boundaries, context, human review, shared workflows, practical training, judgment, measurement, and internal ownership. It does not measure enthusiasm, tool count, or how impressive a demonstration looks.
Use the result as a conversation starter. Ask one employee from leadership, operations, sales, and customer service to score the company independently. The gaps between answers are often more useful than the average.