Proof · Fictional ExampleCo
See what “done” looks like before you buy.
Free policy PDFs stop at a blank template. These ungated excerpts show the kit’s Conditional-first operating doctrine across Decide, Operate, and MSP delivery — three of sixteen completed sample artifacts.
Fictional company only. Not legal, compliance, privacy, security, or professional advice. Not an endorsement of any vendor.
ChatGPT Team: Conditional → standing Approved
Buyers need one finished example of when Approved is the right call. This re-review upgrades a clean Conditional pilot — and teaches Approved ≠ unrestricted.
Company: ExampleCo (90-person professional services) · ID: APPROVAL-2026-0813-CGPT
ChatGPT Team ran Conditional May–August. August quarterly recommended Promote to Approved because scope was stable: business workspace only, MFA + admin ownership live, no connectors/plugins/API/uploads, training/opt-out correct. Personal/free ChatGPT stays Not approved for any company work.
| Use Approved when… | Keep Conditional when… |
|---|---|
| Users, use cases, and data types are clear and stable | Path is still a pilot, seat-expand gated, or time-boxed |
| Required controls are live and verified | Extra compensating controls are still being proven |
| Re-review can move to normal quarterly cadence | Reviewer wants a nearer forced check-in or exit criteria |
Excerpt only. Full checklist + register entry ship in the Business Pack.
High-severity incident: customer data in personal ChatGPT
Operate path is not “write a policy and hope.” This completed triage shows stop-exposure → facts → data types → severity → close — for a common real failure mode.
Incident: Customer support thread pasted into a personal free ChatGPT account under deadline pressure.
| Field | Response (excerpt) |
|---|---|
| Discovered | 2026-06-12 14:40 local (honest self-report) |
| Tool / account | ChatGPT personal free — Not approved |
| Approved path | ChatGPT Team workspace only |
| Data involved | Customer names, emails, contract ID EXC-48219, service-issue summary |
| Severity | High — customer identifiers in unapproved personal AI |
Excerpt only. Full triage checklist + non-punitive close notes ship in the download.
Northline IT → ExampleCo QuickStart engagement
Free templates do not include practitioner packaging. This fictional MSP engagement shows intro email, discovery, fixed-fee QuickStart, and handoff into the same ExampleCo Decide/Operate samples.
| Field | Fictional value |
|---|---|
| MSP | Northline IT |
| Client | ExampleCo (90-person professional services) |
| Engagement | AI Usage Policy QuickStart (fixed-fee mini-project) |
| Fee (illustrative) | $2,400 mid-size SMB band — not a benchmark |
| Kickoff → delivery | 2026-05-28 discovery → 2026-06-20 complete |
| First quarterly | 2026-08-12 (booked at handoff) |
Intro email teaches Conditional-first employee language: which tools are Conditionally approved or standing Approved; personal/free AI is not a company path; how to request review; what to do if sensitive information is shared.
Excerpt only. Full MSP pack includes readiness questionnaire, pitch, scope language, QBR outline, and follow-on map.
Thirteen more completed samples in the download.
Including shadow-AI survey synthesis, employee announcement, 30-day calendar, quarterly review, exception request, Fireflies Not-approved, Copilot / Otter / GitHub Copilot / Perplexity / Gemini / Canva Conditional approvals, and the full Conditional-first register.