Simplifying test analysis into a guided learning experience with AI

Increased student engagement by 41%

About the company

Aakash

India’s biggest edtech platform specializing in preparing students for (NEET) and (JEE) entrance exams

3M+

Students trained

5,000+

Partners

300+

Branches (India)

My role

End-to-end, product designer

Discovery & research

User flows & IA

Wireframing & prototyping

Visual design

Timeline

1 Month (Dec 2025)

The team

1 PM, 5 Developers,

1 Designer

Business problem

Increased faculty time and operational cost spent explaining reports

After each test, students receive detailed reports across multiple sections, while teachers spend significant time explaining the results.

Goal: help students understand their performance without needing a faculty walkthrough

Understanding the problem

To better understand the gaps in the existing experience, I conducted

UX Audit

Student 1-1 interview

Student 1-1
interview

Faculty Interview

UX Audit

Observations from 1-1 interviews

Interviewed 10 students. Here's are the key things I found,

Core insights

Insights which shaped the design

Due to these issues, students were relying on faculties and using them as fallbacks which was operationally heavy

Thus we decided to

Transform static reports into a learning experience which guides students to prepare better

Solution 1

Making data easy to understand

For easier data consumption, we did two things

Converted all important sections into a seemless IA

Easy to consume

Follows user mental modal

Prevented information overload

Progressive Disclosure

Grouped similar information

Solution 2

Built personalised AI engine to guide learning

Introduced an AI summary that simplifies performance and recommends what to focus on next

AI Performance Summary

Summarized view of strengths & improvements

Clear action points to guide what student

Add Personalisation

Personalised recomendations

Explains why you lost marks and how can you improve

Solution 3

Added reasons behind performance

Introduced mistake spotlight and strong/weak chapters to highlight learning gaps and improvement areas

Identified common mistake patterns behind lost marks

Added peer benchmarks to show comparison analysis

Highlighted strong, weak & need improvement chapters

Usability testing

We tested the redesigned report with students to validate whether they could understand their performance, identify gaps, and find what to do next.

9/10

Understood why they lost marks

9/10

Identified their weak areas

7/10

Identified a next action

-

AI Summary Adoption

Impact after pilot

We piloted in 3 branches with cohort size of 7000 students for two months (June & July). Final rolled out in August.

Analysis engagement

38%

2,665 students explored deep level analysis

Self report Understanding

38%

Students could better identify why they lost marks

Faculty reported fewer queries

8/10

Reduced queries around test performance

Design decisions & trade-offs

Every decision has been taken considering the risks involved

Guidance over information overload

Faster understanding of performance

Less raw data upfront, but clearer priorities

Progressive disclosure over instant access

Lower cognitive load

More clicks, but a simpler mental model

AI interpretation over manual analysis

More personalized learning direction

Requires trust and data validation

Learnings

Familiarity is the biggest blind spot

I learned that reducing information isn’t enough. The real value comes from turning data into clear guidance, reasons and actionable steps.

The End

Thank you for your time

To explore final designs. Let’s connect and talk.

JUMP TO