Case study: leading an enterprise AI rollout

In the last session, Andre put the pieces together. The strategy course had produced the roadmap: which initiatives, in what order, with what budget and expected value. This course had produced everything the roadmap needed to work in a real organization of 420 people: an assessment of readiness, an operating model, a skills plan, a change plan, a responsible AI program, a risk register, rules for suppliers, governance with decision rights, measures and a plan for scaling. Graham wanted it in one document that the board could read alongside the roadmap at its December meeting. This lesson walks through that document, the transformation plan for 2027, and the reasoning behind it. Everything in it is a plan; none of the 2027 events have happened.

The starting point

The plan opened with the readiness scorecard from lesson 2. Cascade Ridge scored 2.85 out of 5, at stage 2 of the maturity ladder, experimenting. Leadership alignment and technology were strong at 4. Culture scored 3. Data, governance and skills each scored 2, and those three gaps would block the roadmap if they were not closed. The target for the end of 2027 was stage 3, operationalizing: three AI initiatives running in production with owners, monitoring and support, governed by an approved policy, with every member of staff trained in AI literacy.

Graham asked for the transformation's purpose in one sentence. Andre's draft read: "By the end of 2027, Cascade Ridge will use AI safely and visibly to fill more appointments, reduce paperwork and support overnight care, and will have the people, data and governance to adopt further AI faster and with confidence." The board would be asked to hold the program to both halves of that sentence: the business results and the capability.

The phases for 2027

The plan sequenced the year using the four phases from lesson 1, with different initiatives in different phases at once.

Operating model and governance

The operating model was the hub and spoke from lesson 3: a small, mostly part-time center of excellence led by Andre, with Noor, Kwame, Hannah and the planned data analyst, supported by Dr. Kaur and Lorraine; business owners for each initiative; and a champion at every one of the twelve sites. Governance followed lesson 9: the board for policy, budget and risk appetite; the steering group, chaired by Andre with Graham as sponsor, for pilots, gates and tier 3 uses; and the center of excellence for intake and tier 1 approvals. Decision rights were written down, and any owner or steering group member could switch off a system causing harm at once.

The responsible AI program from lesson 6 and the risk tiers from lesson 7 sat inside that structure. Every initiative had a completed responsible AI checklist and a tier before its pilot. The call-note summaries and the client-support assistant were tier 3; the no-show reminders and lab-report extraction were tier 2; the review-comment themes were tier 1.

People: skills and change

The skills plan from lesson 4 gave every member of staff the AI literacy and responsible AI modules, and gave each role a path matched to its work. The change plan from lesson 5 rested on Graham's three commitments as sponsor, the written statement on roles, the champion network and a staged plan for each pilot. Both carried a cost in staff time that the cash budget did not show.

The 50,000-dollar cash line itself was split four ways: about 22,000 dollars for the share of Andre's time spent leading the program; about 18,000 for designing and delivering training; about 6,000 for governance, including legal advice on the policy and the register of AI systems; and about 4,000 for communications such as clinic briefings and printed guides. Lorraine checked that the four parts added to 50,000, and they did.

Risks to the transformation

The roadmap already carried risks for individual initiatives. The transformation plan added risks to the program as a whole, with owners and responses.

Measures for the board

The board would receive a one-page report each quarter, combining the roadmap's value measures with the transformation's capability measures from lesson 10.

Graham liked the speed measure most. "If the governance works, that number should go down every quarter," he said. "If it goes up, we've built a bureaucracy."

Decisions for the board

The plan ended with the decisions the board would be asked to make in December, alongside the roadmap's budget request.

• Approve the AI policy and principles.
• Approve the governance structure and decision rights, with the steering group's membership.
• Approve the transformation plan for 2027, including the people-side time commitments from each site.
• Approve the readiness targets and the quarterly board report.
• Release the first budget tranche for data foundations, the no-show pilot, the partner build of the assistant and the program line.

Andre also asked the board for something that cost nothing: permission to report bad news as plainly as good news. The first stopped initiative, the first incident and the first missed target would each be reported with what had been learned. Without that permission, he argued, the reports would slowly become optimistic, and the board would lose the ability to steer.

What the case shows

Read as a whole, the plan shows the main ideas of the course working together. Readiness came first, so investment went to the gaps that mattered. The operating model fitted the organization's maturity. Skills and change were planned role by role, with honest communication about jobs. Responsible AI and risk management were built in from the start and scaled to risk. Vendors were governed through contracts and shared responsibility. Governance had clear decision rights and a fast path. Adoption, outcomes and capability were all measured, with feedback loops that led to action. And scaling was planned in waves with production requirements, to avoid pilot purgatory.

Noor's last comment in the session was about time. "None of this is finished in 2027," she said. "If it works, at the end of the year you'll be at stage 3 with three tools running well and a team that knows how to do the next one. That's a transformation. It just doesn't look dramatic from the inside."

What happens if the board says no

Andre also prepared for the board approving less than he asked. If the board approved the roadmap but cut the program line, the plan named what would go first: the regional AI labs and some of the communications, not the literacy modules, the champions or governance, which protected safety and adoption. If the board delayed the client-support assistant, the partner contract would not be signed and the people-side plan would shrink by the hours for client services staff. If the board wanted to go faster, Andre's answer was that the constraint was not money but the time of a few people, and that adding initiatives before the data analyst joined would increase key-person risk. Thinking through these cases in advance meant the meeting could be about choices rather than surprises.

This kind of contingency planning is itself a mark of readiness. An organization that can explain what it will drop, protect and change under different decisions understands its own priorities. Exam scenarios that describe a budget cut or a change of direction usually reward answers that protect the foundations of safe adoption, such as governance, training and ownership, and trim the optional extras first.

Lessons for any organization

Cascade Ridge is a mid-sized veterinary group, but the shape of its plan applies widely. A hospital, a regional bank, a manufacturer or a government department would use different initiatives and different laws, and probably more committees. It would still need to know its readiness, choose an operating model suited to its maturity, train people by role, lead change honestly, build responsible AI and risk management into each use, govern its suppliers, give decisions clear owners, measure adoption and capability as well as value, and scale in steps with gates. The exam is written for leaders in many industries, and its scenarios test those same habits whatever the setting.

Preparing for the exam

Case questions on the exam combine several of these ideas in one scenario: an organization with a gap in readiness, a pilot that is stalling, a governance bottleneck, a workforce concern and a regulatory question, and an answer that must deal with the most important of them first. The habits from this course help: find the gap that blocks the goal, prefer answers that assign clear ownership and accountability, scale controls to risk, communicate honestly, measure adoption and outcomes, and move in phases with gates. The exam readiness course practices these scenarios in full.