I built a book club app in 9 hours. Here's how.
28 OCT 2026I built a web app in nine hours. Not nine consecutive hours, obviously. I have a full-time job, I'm a mom, and I have hobbies! I had about a week to get it done, working on it whenever I could.
But let's start at the beginning.
It started with a Google Doc
I'm part of a wonderful piano community, and we have a book club there. We agreed to discuss one chapter per week, so our moderator set up a Google Doc. Simple enough, right?
Well, Chapter 1 went reasonably well. But by Chapter 2, things started getting a little complicated. During our first online meeting, after we discussed both chapters, our moderator asked whether anyone had suggestions for improving the Google Doc, as it was becoming increasingly difficult to manage.
As people shared their frustrations with the file. Notifications weren't working for them, so people weren't always aware when someone had added a new reflection or commented on an existing one. To catch up, they had to scan through the whole document to figure out what was new.
And then there was the comment-resolving problem. The checkbox was so easy to click accidentally that people kept marking comments as resolved by mistake. They felt bad about it, especially when those comments were part of an ongoing discussion. The mobile experience wasn't great either. Reading and commenting on Google Docs from a phone was far from ideal.
In my professional life, I've designed applications where a group of people need to share a document, comment on each other's contributions, and collaborate around the content. I knew the challenges involved in creating these kinds of experiences, and I found it interesting to see the same problems emerging in our little book club. The difference? Instead of big projects and budgets, we were discussing a book.
Same collaboration challenges, but thankfully, fewer legal constraints!
Five hours in: from idea to MVP
I decided to use AI to help me build a solution. The goal was simple: have something ready for Chapter 4. I started working with Claude Design, defining three main sections:
- Chapters: where members could share reflections, comment on specific sentences, and reply to comments. Most importantly, I wanted clear indicators showing what was new. That was one of the main things missing from Google Docs.
Activity: a notification page showing who had commented on your reflection, replied to you, or tagged you.
Profile: where members could manage their preferences, including whether they wanted a daily email summarising what had happened.
I also needed three roles: an admin (me), a moderator, and regular members. Each person would have their own account and password, so we could identify who had written each reflection or comment.
My initial idea wasn't to build an app, but a responsive website that everyone could access from their phone or computer. Since I already had my own website and hosting, I wanted to put it on a subdomain of my existing domain.
I explained what I had in mind to Claude, and it suggested using PHP. So that's what we went with. From there, it was a constant back-and-forth with Claude. It suggested things, I questioned them, we made changes, and so on.
The UX decisions came quite naturally to me. I'm used to understanding complex problems, identifying user needs, and translating them into simple interactions. The UI, however, was a different story. I work with outstanding UI designers who make even the most complex things look clean and beautiful. And they do it so quickly! With this application, I was much slower and more indecisive. I kept exploring different options before settling on a look I liked. In fact, I spent almost two hours figuring out the colours of the UI.
After five hours, I had an MVP ready for the moderator to test. Five hours. For someone used to digital products that take months to design and build, that was surreal.
But I wasn't done yet.
Four more hours: testing, refining, and launching
I started thinking about how our moderator had been working before we had the website. She'd used AI to summarise our reflections and prepare for the online meetings, so I wanted to make that process easier. I added a Download all button for each chapter, allowing her to download the reflections and comments as a text file to use with AI.
Then I realised I could install the website on my phone and make it work more like an app. So I asked Claude to make that possible and add push notifications. I also added notification settings so members could choose which updates they wanted to receive.
Finally, I changed the daily email summary. Instead of sending an email to everyone every day, I wanted it to reach only people who hadn't checked the website that day. No point telling people about things they'd already seen.
After eight hours, the moderator, two other members, and I were testing the website. And it was actually working. The testers came back with suggestions. I spent one more hour making improvements and redesigning the invitation email, which had previously been plain text. I also added two requested features: the ability to start topics within chapters and a label to make new comments easier to spot.
And that was it! After nine hours of work, we were ready to invite the rest of the book club to start using the website.
What did I learn?
I've never launched a product in such a short time. I'm used to projects that take six to nine months, where we have to navigate legal and technical constraints and work with multiple stakeholders.
This was a much less complex product, with fewer constraints, but that's precisely what made the experiment so fun. I had a real problem and a clear group of users. AI helped me turn an idea into a working MVP much faster than I could have imagined, while I still had to define the features, think through the user experience, evaluate suggestions, and make decisions.
It made me think about something we learn in design school: put things in people's hands as soon as possible. As David Kelley, founder of IDEO, put it:
“Fail faster to succeed sooner.”
After this experience, that principle has never felt more relevant.
The web app is now in people's hands. I'm observing how they naturally use it, what's working, and what needs to change. I've already made a few improvements based on how people are using it and the feedback they've shared. But that's a story for the next post. 😉
For now, I'm just happy that our Google Doc problem has turned into something real that our little community has started using.
- EXPERIENCE DESIGN
- BUILDING WITH AI
- PIANO JOURNEY