Executive Prompt Playbook
~14 min read
Estimated time: ~14 min read — for the in-app brief plus opening the primary source.
What this is
Copy-ready prompt patterns for meeting prep, email, decision memos, vendor comparisons, and coaching notes — plus constraints that improve reliability.
Prompting is briefing — Goal, Context, Constraints, Format.
- Vague one-liners create fluff and fabrications.
- Constraints ban invention; format makes review easy.
- Save three team prompts for recurring work.
- Use only allowed data; enterprise controls differ from consumer apps.
Next action: Write one strong prompt for a real task today and pin it for your team.
Deep dive
Vague prompts force the model to invent your brief. Strong prompts are complete briefs.
Core pattern every time: Goal → Context → Constraints → Format. Optional: Role (chief of staff). Always: list assumptions separately; if insufficient, say so — do not invent.
Goal: what good looks like and for whom. Context: only facts you can stand behind; say 'use only provided materials' when attaching sources.
Constraints: ban invented numbers, names, citations, discounts, legal conclusions; set tone and length. This is the safety rail most people skip on the first try.
Format: bullets, tables, subject lines, redlines — make review fast. Prefer short scannable output over long prose.
Meeting prep example: Goal prepare me for a 30-minute supplier review. Context [paste]. Constraints flag unknowns; no invented metrics. Format 5 bullets + 5 questions + risks + assumptions.
Vendor comparison example: force a criteria table you define; forbid inventing features not in the source text; end with open questions for the vendor.
Customer email example: Goal reassure + next step. Context order facts + approved remedy from policy. Constraints no invented offers. Format subject + 3 short paragraphs.
Policy FAQ example: answer only from attached approved policy; cite section; if missing say so; escalate path.
Always paste only data allowed by policy. Prefer enterprise workspaces over consumer chat for Amber/Red data.
Improve loop: if output invents something, tighten constraints and re-run — then save the winning prompt in a team library.