Thinking before diving into the deep end of this AI Experiment

When I first started my AI Experiment with Bolt.new, it felt like magic: input my idea and voilà, an app started taking shape. Friends and family had fun with the thing I made (a budget management tool), but I knew this was just the beginning and I wasn’t sure if I could – or should- take the next steps. Still not entirely sure, but I’m enjoying the journey of figuring it out.

I didn’t understand many aspects of the development process of my first AI Experiment. I was signing up for services, deploying to places I didn’t fully understand, nor did I fully grasp whether my project was of a noteworthy level of quality and security.

This series of articles is for anyone like me—a designer who dreams of creating their own scalable, secure, and feature-rich application but feels overwhelmed by the technical world that lies beyond UX design. Here’s what I’ve learned so far.

What It Takes to Build a Mature App

After that first experiment, I realized creating a robust, user-friendly app involves more than just inspiration—it demands a structured approach, improving my prompting skills and a careful selection of tools. To build an app that’s scalable, secure, and ready for monetization, you’ll need more than just code. Here’s an idea what goes into the backbone of a “mature” app and what I plan to use:

  1. Bolt.new (Design & Build – €20): Bolt.new is a platform for rapidly building and deploying apps without extensive coding; it simplifies the development process for non-developers while offering flexibility for customization.
  2. Netlify Pro (Hosting & Serverless Functions – €19): Netlify hosts your app on fast, reliable servers and provides serverless functions, enabling backend operations without needing a dedicated server.
  3. Supabase (Database€19): Supabase is a scalable database solution for storing and managing app data, such as user accounts, product inventories, or any other dynamic content.
  4. Sentry (Monitoring€10): Sentry monitors your app for bugs and performance issues, helping you quickly identify and fix errors before they impact users.
  5. Plausible Analytics (Analytics€9): A privacy-focused, GDPR-compliant analytics tool that tracks user behavior without collecting personal data, aligning with my commitment to privacy by design.
  6. GitHub (Version Control€4): GitHub allows you to track changes in your app’s code, collaborate with others, and safely roll back to previous versions when needed.
  7. Stripe (e.g. 2.9% + €0.30 per transaction): Stripe processes online payments and subscriptions with transparent fees, offering a scalable, developer-friendly solution for businesses of all sizes.

So it’s basically around € 100 a month just to have the thing, not counting the fees from accepting payments. You get a whole new level of appreciation for developers of applications, that have to leverage these monthly costs against providing a valuable product or service to a user. And rapid development, might not be about learning and failing fast to help a user. Especially when you have a growing user base, these costs will rocket sky high, how one manages to make a living from this, is currently beyond my limited capacity of understanding. I’m just a designer, dabbling with learning a new skill, okay?

A Designer’s Perspective: Where the Struggles Begin

As someone used to designing complex enterprise tools, I didn’t think much of it. Just naively prompted my idea and what’s the magic happen. I could have left it to that, but no, I had to open Pandora’s box, and now I’m knee-deep in deployments, databases, and monitoring services. What started as a fun experiment has turned into a crash course in modern software development, and honestly, it’s been invaluable for connecting the dots as a designer.

Signing up for services I didn’t know was fun to see a result, when I didn’t understand how to integrate Bolt and Netlify, I ended up breaking my deployment. That was a fun learning opportunity. I had a huge “aha” moment, and all past conversations with developers all started to make so much more sense. Finally, I started to connect the dots, realizing how to connect it all and that I was using tokens and burning bandwidth. I now knew I needed to smart up and become a bit more sophisticated in my actions. I don’t mind paying tuition fees, but I don’t like waste.

Key Lessons for Fellow Designers

  • Understand the Ecosystem: Before signing up for services, take time to map out what your app needs. For example, hosting platforms often bundle features like analytics or version control, which can save costs.
  • Think Long-Term: Building a scalable app means considering future updates, security, and performance. What works for a small project might not scale for a larger audience. For example, choosing a database like Supabase ensures scalability if you gain thousands of users, whereas simpler solutions might crumble under growth.
  • Leverage AI, But Don’t Rely on It: Tools like Bolt.new are fantastic for getting started, but understanding the “why” behind decisions (like where to deploy or how to set up a database) is critical for long-term success.

Why This AI Experiment Matters

Learning to build an app has been fun, a bit humbling, challenging, and incredibly rewarding. By understanding what it takes to create a mature application, I’m improving myself as a designer working in cross-function development teams.

Next up in this series is a deep dive into the art of ‘naive prompting.’ You may call me Queen of Naive Prompting or Mrs. ‘Let Me Ask ChatGPT.’ In full transparency, this article concept was co-written with ChatGPT, and I spent two hours adding my own insights and personality. Having an AI writing buddy isn’t just efficient—it’s like having a brainstorming partner who never runs out of energy (or complains).

Thanks for reading! You’re amazing, you know that, right?