TerminologytermStep 3: How you talk to modelsAll

Term: Prompt

~5 min read

Estimated time: ~5 min read — for the in-app brief plus opening the primary source.

What this is

A prompt is the instruction and context you give a generative model. Clear goals, constraints, and format dramatically improve usefulness.

Everyday example

Instead of typing only 'Rewrite this,' a strong prompt says: 'You are a chief of staff. Summarize this board pack in 8 bullets for a CFO. Flag risks. Do not invent numbers. List assumptions.'

A prompt is the instruction plus context you give the model — role, task, constraints, and input.

  • Vague prompts produce vague, confident drafts.
  • Strong prompts name the audience, the format, and what “good” looks like.
  • The prompt is not a spell; it steers a pattern engine.
  • System prompts (hidden) plus user prompts (what you type) work together.

Next action: Rewrite one recurring request this week with role, output shape, and a quality check.

What changes in how you lead

How decision rights, process, and ownership should change.

  • Shared prompt patterns belong in the team, not in one person’s chat history.
  • High-stakes prompts should include a verification step.

Compare related ideas

Prompt vs System prompt

The prompt is what you type this time. The system prompt is the standing brief — role, rules, and limits — that sits underneath every turn.

Open System prompt

Deep dive

Pattern: Role/Goal → Context → Constraints → Output format → list assumptions to verify.

Sensitive data in prompts is still data leaving or entering a system — classify first.

Reusable prompt libraries beat one-off cleverness for Legal, HR, Marketing, and ops teams.

Related terms

Related weekly lessons

terminologypromptproductivity