How AI can help in User Research

Disclaimer: This is not an ad for NotebookLM. It's just one of the AI tools that I really like to use. It's a proprietory RAG software. There could be other open-sources tools that enable the same functionality.

Introduction

For my thesis, I conducted several studies including ethnographic research, product analysis, and user interviews.

To design an experience that helps student have a healthier relationship with technology, it was first important understand how students use technology today, how technology is designed and where it can go next to serve user needs.

User Interviews

A diverse group of students aged 17-25 across different geographies and academic journeys were interviewed. The interview protocol included questions about their daily routine, study habits, use of smartphone and computers, social media usage, and their organization of thoughts and ideas.

A lot of early research also focused on asking people about tasks that they tried to accomplish every day and if they had fixed time to accomplish that; and the tasks they struggled with every day.

Once I had interviews conducted and I had a few hours of conversations to sift from, a challenge I felt was the feeling of having forgotten some key insights. While I had notes from the early insights, I didn’t have anyway of remembering everything that the students mentioned.

This is where AI comes in, organizing large amounts of data, and sifting through it was extremely useful.

Using AI to do User Research with you, not for you

While I didn’t rely on AI to do the user research for me. I was very much impressed how using AI to aid user research helped me gather, synthesize, and expand on my insights.

I used NotebookLM throughout my thesis, not because I was interested in it generating content for me (I think AI generated text and images are sloppy and classless); but because it’s a genuinely great research organization tool that provides utility of drawing connections between discreet topics and helping me understand my own journey better.

Using NotebookLM to organize interviews

Drawing key insights from the user interviews: Part of the NotebookLM UI is like a super-personalised chat bot that takes in information from the provided sources, and gives insights. The good part is all the sources can be referenced back, which makes error handling easier.

Key Insights from the Interviews

Based on the user interviews, my notes, and additional summarisations that I visualised with the help of NotebookLM. I summarized the insights into 3 key points.

1. Smartphones are a constant source of distraction for students, even when they take measures like putting them face down to avoid using them. Students are aware that they’re a source of distraction.
2. Students preferred analog study tools such as paper notebooks for their tactility. They sometimes used hybrid tools like ePaper note taking tablets for the convenience of search.
3. There’s a key difference between students being distracted by social media when studying and students consciously using the platforms as a way to unwind. The latter is actually a healthy practice of taking breaks from work, the former is more concerning and reduces their focus.

Using frameworks to visualize the interviews into meaningful information

NotebookLM was not only great for compiling user insights. It was also great at applying the frameworks I wanted to apply for my user research and giving me information based on the framework.

For example I wanted to use the Jobs to be Done framework to understand the tasks I had identified to be really useful.

Bonus: NotebookLM also generates podcasts, while most of the times they are a little silly and not useful; sometimes I found them to be good at summarizing key concepts.

Like this audio podcast here summarizes my user interviews. It's not perfect, but as a prototype it's an interesting exploration.

Take a look at the artifacts mentioned in this post

Interview Protocol: Download Here

NotebookLM generated content:
Interview Summaries [PDF]

Using the Jobs to be done framework
https://archive.org/details/jobstobedoneroad0000wunk/

Applying Jobs to be done on "Capturing and organizing ideas" [PDF]
Applying Jobs to be done on "Free Time" [PDF]