Workplace AI reference
Prompting and Context Reference
Good prompting is not a magic phrase. It is the habit of giving an AI tool the right task, useful context, clear boundaries, and a reviewable output shape.
The prompt ingredients employees should practice
AI Lunchroom labs teach prompt ingredients through small workplace tasks so learners can see what changes the output.
- Role context: what job or department the work supports.
- Task: the one finished output the user wants.
- Source material: notes, draft text, facts, examples, or constraints.
- Output format: bullets, table, checklist, summary, draft, or questions.
- Review rules: what not to invent, what to verify, and what needs approval.
Context should be useful, not excessive
Training should teach people to include enough context for useful output while removing sensitive or unnecessary details. More data is not automatically better.
- Use one-use task details for the current lab.
- Save only stable, reusable work context when the user explicitly chooses that path.
- Keep private prompt context separate from public profile data.
- Avoid pasting confidential, regulated, employee, customer, patient, credential, or financial details into unapproved tools.