Smart Feedback Loop

Overview:

The goal of this project was to design an AI-powered learning platform that provides real-time, rubric-based feedback to design students, reducing the typical 48–72 hour delay in receiving meaningful critique.The platform focuses on enabling continuous learning by helping students understand, apply, and improve based on structured feedback rather than waiting for manual reviews.

Project Duration:

12 weeks (Capstone Project)

Role:

UI/UX Designer (Team Project)

• Led end-to-end UX process (research → design → validation)

• Contributed to AI-assisted feature design and user flows

• Designed wireframes, prototypes, and high-fidelity interfaces

Tools Used:

• Figma (UI/UX Design & Prototyping)

• Miro (User flows, journey mapping)

• Pen & Paper (Ideation)

Process:

1. Research & Discovery:

• Conducted user interviews:

  • 10 design students

  • 2 design mentors

Key Insights:

• Delayed feedback slows learning and iteration cycles

• Students struggle to interpret and apply qualitative feedback

• Lack of structured evaluation criteria (rubrics) creates ambiguity

• Defined problem statement:
  → “How might we enable immediate, actionable feedback that supports continuous learning?”

2. Wireframing & Prototyping:

• Designed core flows:

  • Design submission

  • AI-assisted feedback generation

  • Feedback interpretation and learning loop

• Developed low to high-fidelity wireframes:

  • Focus on clarity of feedback

  • Easy navigation between submission, critique, and iteration

• Built interactive prototypes to simulate:

  • Real-time feedback experience

  • Iteration cycle based on insights

3. Visual Design:

• Created a clean, structured interface prioritising:

  • Readability of feedback

  • Clear separation of evaluation criteria (rubric-based)

  • Minimal cognitive load

• Designed UI to support:

  • Layered feedback (AI insights + mentor alignment)

  • Guided improvement flow

4. Usability Testing:

• Tested prototypes with students (5–7 users)

Key Observations:

• Users valued immediate feedback over delayed mentor reviews

• Structured rubric feedback improved clarity and actionability

• Some users needed guidance on interpreting AI suggestions

Iterations Made:

• Simplified feedback language and visual grouping

• Improved onboarding to explain AI-assisted feedback

• Enhanced navigation between feedback and iteration

5. Implementation & Launch:

• Delivered high-fidelity designs and interaction flows

• Defined how AI integrates as a feedback layer (not a replacement for creativity)

• Validated concept through prototype testing

(Note: Full development not in scope)

Results:

(Prototype-based validation)

• Reduced perceived feedback wait time from days → immediate (conceptual impact)

• ~75% of users found feedback more actionable compared to traditional methods

• Improved confidence in iteration and self-learning

Key Takeaways:

• The speed of feedback directly impacts learning efficiency

• Structured, rubric-based critique improves clarity and usability

• AI is most effective as an assistive layer, not a replacement for human creativity

• Clear guidance is essential when introducing AI-driven features

Next Steps:

• Integrate mentor feedback alongside AI insights (hybrid model)

• Improve AI explainability for deeper learning

• Test with larger user groups for validation

• Explore further integration with design tools