
AI Fix-It Challenge
🧠 How It Works
-
We give participants:
-
A bad AI output (text or image)
-
The original poor prompt
-
Their task:
-
Rewrite the prompt to fix the output
-
Explain in 2–3 lines why their prompt works better
📥 Submission Method
-
Improved prompt
-
New AI output (optional but bonus)
-
Short explanation
🏆 Evaluation Criteria
-
Prompt improvement quality (40%)
-
Output relevance (30%)
-
Explanation clarity (20%)
-
Creativity (10%)
🥇 Winner Selection
Best prompt correction + reasoning wins.

❌ Given Poor Prompt
“Create a chatbot for a college website.”
❌ Given AI Output
A chatbot can be added to a college website to answer student questions. It can help with admissions, courses, and general
information.
🎯 Your Complex Task
The prompt above is vague, under-defined, and lacks technical clarity, resulting in a weak AI response.
Your challenge is to re-engineer the prompt so the AI can generate a production-ready chatbot specification for a college website.
You must think like a product designer + AI engineer.
✅ Mandatory Requirements for Your Improved Prompt
Your rewritten prompt must force the AI to generate output that includes ALL of the following:
1️⃣ Role Definition
Clearly define the AI’s role
(e.g., AI Product Architect, UX Designer, Technical Consultant).
2️⃣ Target Users
Specify at least three user types:
-
Students
-
Parents
-
Faculty / Admin
3️⃣ Core Features (Minimum 8)
The output must include features such as:
-
Admissions support
-
Course & fee queries
-
Event updates
-
Timetable help
-
Contact routing
-
FAQ handling
-
Multilingual support
-
Escalation to human staff
4️⃣ Functional Constraints
Your prompt must specify:
-
Response tone (professional + student-friendly)
-
Data privacy consideration
-
Response length control
-
Fallback handling when data is missing
5️⃣ Output Structure
Force the AI to respond using:
-
Clear headings
-
Bullet points
-
Feature grouping
-
Optional flow diagram explanation (text-based)
6️⃣ Technology Context
Your prompt should mention:
-
Website integration
-
CMS compatibility
-
AI model flexibility (tool-agnostic)
📤 What You Must Submit
-
Improved Prompt (Main Evaluation)
-
AI Tool Used
-
Improved AI Output (Optional but highly recommended)
-
Explanation (150–200 words) covering:
-
What was missing in the original prompt
-
How your constraints improved output quality
-
Why your structure is more effective
-
