Approved
Public, synthetic, or explicitly approved information used for a suitable low-consequence task.
AIforBC Learning · Curriculum 1.0
AI for Teams
Participant Workbook
Turn one recurring team task into a safer, measured AI workflow.
- Organization
- Participant
- Workshop date
- Facilitator
01 · Boundary card
What information and decisions stay out?
Pause and confirm
Internal or sensitive context that requires an approved tool, purpose, controls, and owner.
Do not enter
Passwords, credentials, prohibited information, or personal/confidential data without an approved basis.
Our team rules
02 · Task brief
Give the system enough context to do bounded work.
03 · Output review
Confident language is not evidence.
- The output answers the assigned task.
- Material claims trace to authoritative sources.
- Facts, assumptions, and recommendations are separated.
- Names, numbers, dates, citations, and calculations are checked.
- The result stays inside the approved information boundary.
- A qualified person completed the required review.
- Uncertainty and exceptions are recorded and escalated.
Corrections or questions
04 · Workflow scorecard
Choose a small, frequent, reviewable first workflow.
| Candidate workflow | Frequency | Current friction | Potential value | Feasibility | Consequence |
|---|---|---|---|---|---|
Selected first workflow:
05 · Pilot brief
Define the test before changing the process.
- Workflow and owner
- Current process and baseline
- Approved inputs
- Expected output
- Required review
- Exceptions and stop conditions
- Success measure
- Pilot dates and decision date
06 · 30-day playbook
Practise, measure, review, decide.
Prepare
Confirm tool, information boundary, examples, owner, baseline, and review method.
Practise
Run supervised examples. Record corrections, full-cycle time, and friction.
Standardize
Refine the task brief and review checklist. Document exceptions and escalation.
Decide
Compare the result with the baseline. Expand, revise, stop, or choose another workflow.