Some clarifying questions to get started with,
- Is this part of the Zoom app or a stand-alone app
- For the sake of discussion, I’ll consider Indian geography
- The overall objective of designing such an app,
- Resource constraints like Engineering, QA bandwidth, and budget for the same
- Time constraints, assuming I’ve to launch this in 4 months
- I’d also want to understand the users a bit better, are we designing for
- Young school kids (UKG to 5th standard)
- Middle school children (5th to 10th standard)
- Senior school students (10-12th standard)
- Students enrolled in the Universities
- Students with special needs for instance could be students with low to zero vision
- Students enrolled in coaching institutions for competitive exams or tuition
Answers to the above questions by the interviewer
- Part of the Zoom app
- The objective is to launch, drive adoption and engagement
- No constraints with respect to the budget
Moving ahead, would like to talk a bit about trends here.
Since covid lockdown, we saw a lot of Schools, Universities, and Coaching institutions adopt digital infrastructure and Zoom was one of the products that saw a massive surge in terms of usage. The Internet penetration in India is increasing as also the smartphone users in India. India is a mobile-first digital economy and hence I’d like to focus on mobile devices.
Also, with respect to the users I’d want to pick students who have enrolled for coaching exams. As India is still developing and a large part of users actually have to work hard to survive in the competition. A lot of users are not able to travel daily to coaching or join offline classes for their career, but having this tool that can allow them to have access to the quality education and also improve their experience while engaging in a lecture would improve their chances of success. I believe that’s a very niche and underserved market and also this market is quite big in size.
Check with the interviewer on the same, he is aligned moving ahead.
Next, I’d list down some needs for students:
- Lectures could become a boring and isolating experiences sometimes, can this be made interesting
- I want to understand the important concepts from this lecture
- I want to retain the most information
- I want additional help with certain topics in the lecture
- I want an uninterrupted lecture experience
Among the above needs, I’d like to pick a need that is most pressing which has a high intensity, and has a higher frequency of occurrence.
Based on the above criteria, I’d like to go ahead with
- Listening to live lectures can become boring and isolating experience
- I want to understand the important concepts
I’d check with the interviewer before moving ahead.
Next I’d like to brainstorm some solutions around the same
- In frame pop-quiz as soon as a topic is finished, this quiz can be generated by AI on the fly, will help users understand their retention and give recommendation
- Auto-generate paraphrased and translated transcription of lectures that users can consume to understand the topic better
- Cut and create a snippet of the lecture video and use that as a note
- Ask questions about the topic of Generative AI, here Gen AI assists students while the teacher is not interrupted. The teachers can intervene only when Gen AI is not able to clarify.
- AR-based lecture for important topics, students can see things in action as the instructor explains the concept. This will help with retention and also improve engagement.
I’d like to prioritize these solutions based on some meaningful criteria and check in with the interviewer:
Moving ahead, I’d like to prioritize the features based on the impact they could bring and the effort required
FeaturesImpactEffort
| Auto Pop-quiz |
High |
Medium |
Consume transcription
(paraphrase + translate) |
Medium |
Medium |
| Create a snippet of the lecture |
Low |
Medium |
| QnA with GenAI |
High |
Medium |
| AR infused lectures |
High |
High |
I’d like to start with two features:
- QnA with GenAI
- Auto pop-quiz
Metrics:
- No. of users asking questions to GenAI
- No. of users getting a satisfactory response from GenAI
- No. of users still asking doubts to the instructor
- No. of users completing auto-pop quiz
- No. of users skipping the pop-quiz
- Avg. score of users when they see the pop quiz
- No. of installs for this conference app
- Avg. rating on the play store and app store
At the end I’d like to summarize by going over objectives, user needs and then working on feature implementation. However, for GenAI we need to be cautious, as the outputs are biased, there is a hallucination issue with GenAI and the data that it is trained on the publicly available data and hence could be misleading or incomplete. Overall, I believe GenAI can be a fantastic assistance/partner to students and help them achieve more with their preparation.
Some pointers, the app size and performance should not take a hit. The battery should not drain out for the smartphone. A common concern is around the users being in a high network coverage area and using a decent smartphone. The GenAI can function only on the server hence requests might take time to get served, need to keep these constraints as well while designing the solution.