Open the shared drive and you will find the policy. It names a few banned tools. The people who approve use cases have left it alone since the draft. Staff keep pasting work into whichever window loads fastest. People point at the file when someone asks whether you have a policy. Nobody has used it to decide anything.
NIST's risk framework is the reference a leadership day should translate, and I would keep the PDF out of the recital. It is voluntary. It gives you a way to put trustworthiness into how you design, use and evaluate an AI system, including the generative profile NIST added for the risks that belong to those systems. The 1.0 text is under revision, which is another reason to teach the decisions and skip the line by line reading.
Two different jobs
The AI governance workshop runs as two tiers because the jobs differ. All staff need a half day on everyday use. What can go into a tool, when to distrust an output, and what shadow use looks like. Leaders, risk owners and the people who approve use cases need a day on the operating model. Most large organisations need both. Book only the staff session and you leave the approval question unanswered, which is the question the policy file failed to settle.
What leaders leave able to do
Leaders leave able to classify a use, name an owner and approve or refuse it in language a squad can follow. You map the operating model to NIST, ISO 42001 and the EU AI Act, then you translate it into the rules the organisation already has. A proposed use arrives. It leaves with a named yes, a named no or a named limit. A new binder that restates the framework will sit next to the policy file.
Each class of use has an owner, and that owner carries the downside. Staff get one plain statement of what may go into a tool, written against the tools they already open. A use that can send, pay, file or tell a customer something needs a recorded approval. A screenshot in a chat will fail an inquiry six months later. Every pilot has an end date. On that date the permission stays, shrinks or stops.
Squads still need questions they can answer before an agent acts. That list is in AI governance for non-lawyers. Who may let a system send is in who may let an internal agent send. The leadership day is where those answers become a rhythm someone runs, in place of a note in a ticket that closes when the sprint does.
What to leave out
A statute class is the wrong room for people who will never open a statute. So is a policy reading that never names the tools staff already use, and a rule so heavy that the rollout you funded simply stops. People follow guardrails they can remember on a Thursday. Lawyers still mark the pages that touch privacy, employment and the regulator. They should share the yes with the people who run the work.
If you are choosing the day, ask who will approve the next use case after the session, and where you will write that approval down. If the answer is that everyone will be more aware, you have booked the staff briefing and called it the leadership day. Put the name, the limit and the date in the notes.
The NIST AI Risk Management Framework is the document that leadership day should translate.
