article · Prompt engineering · Workflow
Prompting for designers: from request to spec
A good prompt reads like a good brief: context, role, constraints and acceptance criteria. How to write prompts that give usable results.
Preview: all content, comments and numbers here are placeholders until the Lab is connected to the server.
01
A prompt is a brief
What you need to brief a junior designer, you need to brief a model: who is the user, what is the problem, what must not change?
02
Constraints make quality
“No more than 10 suggestions” and “rank by impact” turn a long generic answer into a work list.
03
Ask for analysis before solutions
Separate analysis from proposals. When the model jumps straight to UI it repeats common patterns, not what fits your problem.
04
Keep what works
Every prompt that gave you a great result deserves to become a template with variables. That is how the prompts in this Lab began.
Next in the LabAn ERP is not a database with a UI