Producing AI output is easy. Being accountable for it is the skill.
The AI Operator Program is a three-stage professional certificate for people who are answerable for AI-enabled work — operations leaders, programme and project managers, functional managers, consultants, founders and governance professionals. You bring one real workflow. You leave able to bound it, supply it, supervise it, test it, measure it, and decide whether it should continue.
The questions that have no owner
Ask anyone running an AI-enabled process these seven questions. The pattern in the answers is more interesting than the answers.
- What may the AI access or change? Not what it was built to do — what it can actually reach today.
- What information and instructions does it need? And who maintains that supply when the source changes.
- Who checks the work, and against what standard? A reviewer without a standard is a second opinion, not a control.
- When does an approval or record stop being trustworthy? Most approvals are written as if permanent. Almost none are.
- How do we know the controls work outside the configuration screen? Configured is not enforced.
- Is it producing value after review, rework and risk? Gross output is not the measure.
- Can someone else operate it if its creator leaves? If not, you have a person, not a process.
Most organizations can answer two or three. The ones who can answer all seven did not get there by being careful — they got there by building something.
What you leave with
The certificate is not the product of this program. The product is a chain of operating artifacts built around one AI workflow you are responsible for — and you start accumulating it in Module 1. Learners do not describe governance. They build the mechanisms through which governance happens.
| Stage | Artifacts you produce |
|---|---|
| Core | Scope statement · context package · risk-tiered approval matrix · self-invalidating record · enforcement test |
| Strategic | Business case · measurement plan · tool decision record · stakeholder alignment plan |
| Operations | Operating concept · readiness analysis · environment and authority maps · bounded deployment · executable brief · handoff protocol · control plan · value-stream report · sponsor evidence pack · capstone |
The three stages
One program, ten modules, three stages. Each stage ends in its own credential, so you can stop at any stage and leave with something real.
| Stage | The question it answers | Modules | Time | Fee | Prerequisite |
|---|---|---|---|---|---|
| Core | Can you govern a single AI workflow? | 1 Scope · 2 Context · 3 Approval · 4 Warrant | 16–22 h | $100 | None |
| Strategic Track | Did it produce organizational value? | 5 Strategic Alignment | 26–37 h incl. Core | $125 incl. Core | None |
| Operations | Can you run it as an accountable operation? | 6 Design · 7 Environment · 8 Execute · 9 Evaluate · 10 Deliver | 33–50 h | $165 | Strategic Track completion |
The whole program is roughly 60–85 hours, or about 14 to 24 weeks at three to five hours a week. The work carries forward: what you build in Core becomes the input Strategic measures and Operations runs.
How each module works
A teaching book, three graded applications with visible rubrics, an Operator Journal entry, a discussion, and a knowledge check. Graded work is scored against a published four-level rubric — Emerging, Developing, Proficient (the target), Exemplary. Knowledge checks pass at 80%.
Every worked example comes from a real, dated operation — not a hypothetical one.
The capstone
Operations closes with a performance assessment rather than a written test. You run a bounded publishing lab through a browser terminal under your own identity. You declare opening orders before running anything, take bearings at a declared interval, and order at least two corrections sized to the evidence.
At some point the instructor introduces one unannounced failure. You are not graded on avoiding it. You are graded on whether your bearings detect it, whether your correction is proportionate, and whether your record shows who performed, approved, owned and could accept the residual risk.
Who this is not for
Someone who wants better prompts does not need this. Someone who wants to build or fine-tune models does not need this either. There is good instruction available for both, and this is neither.
This is built for the person who already has AI inside work they are accountable for, and has noticed the accountability has nowhere to sit.
Enrol
Enrolment and delivery are on our learning platform, lms.drbill360.net. Each stage is purchased there.
The AI Operator Program is aligned to O*NET-SOC 15-1299.09 — Information Technology Project Managers. Delivered by CIWLEARNING, a registered assumed name of Cafe Iman West, LLC. © 2026 Cafe Iman West, LLC. All rights reserved.