Governing AI without slowing it down

 AI Governance Â
The Problem
Staff at a global accrediting organization were building with AI quickly, through features in existing platforms and through tools they created themselves. Leadership wanted that work to stay safe, consistent, and sustainable without stifling it. Their data foundations were solid but not yet systematic, and AI ambitions were outpacing them.
My Role
I led this engagement, from risk tier creation through framework design and onsite validation, working across IT, data, AI enablement, HR, and business teams.
What I did
Ran a full-day working session where cross-functional teams walked live projects through the workflow. Together we gave lower-risk experiments more room, set up consultative intake with the technology team, and made proven competency the bar for staff building their own tools.
4.
Tested it against real work
Created an intake process and project registry so staff could see existing work and avoid duplicating it.
3.
Made the approved path easy
Grounded in ISO/IEC 42001 and aligned with existing compliance obligations: a policy defining permitted, restricted, and prohibited AI use; risk tiers based on data sensitivity, audience, and system exposure; and development, testing, and release standards.
2.
Designed the governance framework
Reviewed the data warehouse, pipelines, and member-facing data products, flagged near-term risk from a CRM migration, and delivered a phased roadmap covering stabilization, AI-ready metadata, and data classification.
1.
Assessed the foundation
The Result
A governance model that matches oversight to risk across four controls: cost, data, audience, and sustainability. It channels staff innovation through a clear path instead of shadow IT, and it's built to evolve as AI capabilities change.​