Fix Your Output
Fix Generic AI Output
When AI output feels generic, it’s usually missing details about your situation. Add what’s missing.
The Four Layers of Context
-
Background
Give the AI the situation.
What’s happening? Who’s involved? What’s been decided?
Examples
- “We’re rolling out a new onboarding process next month.”
- “Our team has four people and we’ve been short-staffed since January.”
- “The client renewed their contract but asked us to cut the timeline in half.”
-
Audience
Tell the AI who it’s for.
Who will read this? What do they already know? What level of detail do they need?
Examples
- “The checklist is for new hires in their first week.”
- “Write a report for senior leadership who will use it to approve the budget.”
- “This email goes to the whole department. Not everyone knows the project details.”
-
Purpose
Tell the AI the goal.
What should this do? Inform? Persuade? Reassure? What should the reader do or feel?
Examples
- “Reassure the client that the project is back on track after the delay.”
- “Convince leadership to approve the new schedule.”
- “Help new hires feel prepared before their first day.”
-
Reference Material
Give the AI an example to follow.
Do you have an example? A template? Anything that shows the AI what good looks like?
Examples
- “Here’s the slide deck from last time. Follow the same layout.”
- “I’ve attached an email with the tone I want you to match.”
- “Write the proposal using this template as a guide.”