Manager training · agenda published · pilot cohorts forming

Practical AI Governance: controls a working team can actually use

A role-based path for turning an AI policy into an intake process, data rules, human review, incident handling, and evidence that survives a real operating week.

Review the agenda Ask about a pilot cohort

Who it is for

Bring a live decision.

Business and technology leaders responsible for AI adoption
Security, privacy, legal, and risk partners reviewing AI use
Managers who need rules teams can apply without a policy meeting
Procurement and vendor owners collecting AI assurance evidence

Format

Four 60-minute modules with a cross-functional scenario exercise. Available remotely or as an internal workshop.

Before the session

Bring one current or proposed AI use case and any policy your team already has.

Launch status

The agenda and completion standard are published now. Facilitation notes and exercises are being refined with early cohorts, so we will confirm fit and format before proposing a session.

Course agenda

Four modules. Four working artifacts.

60 minutes

Turn principles into an intake lane

Route low-, medium-, and high-consequence uses without treating every prompt as a committee event.

Use-case inventory
Consequence and reversibility
Fast lanes and escalation
Named business ownership

Workshop: Triage six realistic use cases and defend the route for each one.

Take-home artifact: Risk-tiered intake form

60 minutes

Set data, model, and vendor boundaries

Define what may enter which tool and what evidence is required before enabling it.

Data classification
Training and retention terms
Subprocessors and model routing
Access and logging

Workshop: Map one use case from source data through model, output, storage, and downstream action.

Take-home artifact: AI data-flow and vendor evidence checklist

60 minutes

Design human review that has authority

Match review depth to consequence and give reviewers evidence, time, and rejection power.

Reviewer competence
Claims and source checks
Automation bias
Override and escalation records

Workshop: Write an approval rubric for one customer-, employee-, or production-facing output.

Take-home artifact: Human-review control card

60 minutes

Monitor, respond, and retire

Recognize an AI incident, contain it, learn from it, and stop a system that no longer earns its risk.

Operational monitoring
Complaint and exception signals
Incident ownership
Change review and retirement

Workshop: Run a scenario exercise involving an unsupported claim that reaches a customer.

Take-home artifact: AI incident and change-review playbook

Capstone and completion

One governed-use packet

The capstone is a working control set for one use case: owner, risk tier, permitted data, approved tools, vendor evidence, review step, monitoring signal, incident route, and retirement condition.

Completion standard: Completion requires a usable governed-use packet and participation in the scenario exercise. It is not legal advice or an accredited certification.

Frameworks used in the course

Pilot cohort

Shape the course around work you already own.

Send the team size, use case, and decision in front of you. Gary will reply by email with fit and possible next steps. This does not reserve a date or collect payment.