CASE STUDY
Prompt Genius
ROLE
Product Design
DURATION
4 Months
TOOLS
Figma, Figjam
Overview
PromptGenius teaches people to prompt AI well — not with a library of examples to copy, but with real-time coaching while they write.
Challenge
Millions of people use ChatGPT and walk away thinking it's not that smart. It's not the model — it's the prompt. Every existing resource taught prompting after the fact: tutorials, threads, finished examples. Nobody was in the room while the user was actually stuck, mid-sentence, writing something vague.
Design Solutions
I designed a live scoring system that reads a prompt as it's written — scoring clarity, specificity, context, and structure in real time, with inline feedback that explains why something's weak, not just that it is. Then I tested it: 20-25 real users, writing real prompts, watching whether feedback in the moment actually changed how they wrote.
Outcome
It did. Users measurably improved prompt quality through the coaching loop alone — proof that teaching during the task beats teaching before it.
CASE STUDY
Prompt Genius
ROLE
Product Design
DURATION
4 Months
TOOLS
Figma, Figjam
Overview
PromptGenius teaches people to prompt AI well — not with a library of examples to copy, but with real-time coaching while they write.
Challenge
Millions of people use ChatGPT and walk away thinking it's not that smart. It's not the model — it's the prompt. Every existing resource taught prompting after the fact: tutorials, threads, finished examples. Nobody was in the room while the user was actually stuck, mid-sentence, writing something vague.
Design Solutions
I designed a live scoring system that reads a prompt as it's written — scoring clarity, specificity, context, and structure in real time, with inline feedback that explains why something's weak, not just that it is. Then I tested it: 20-25 real users, writing real prompts, watching whether feedback in the moment actually changed how they wrote.
Outcome
It did. Users measurably improved prompt quality through the coaching loop alone — proof that teaching during the task beats teaching before it.