Short description: In this text, I share my journey of building a small booking application for my friends’ business with the help of Chat GPT. I describe the decision-making process, coding challenges, and key lessons I learned along the way. Overall, it’s a personal account of my experience with AI-driven development, offering insights and takeaways for those interested in similar projects.
Background
Couple months ago friends of mine informed me about starting their small business and asked about some advices regarding their social media, website … Knowing how much this means for them I offered my help in setting up everything needed. Gradually, from starting with a basic website and offering simple advice, we eventually arrived at the idea of creating a small booking app where their potential customers could book services they offer.
Out of my own curiosity and growing interest for AI tools all around me, instead of implementing one of the solutions that are already available on the market I decided to build my own small booking app with the help of Chat GPT.
Disclaimer: While I’m not a developer, I am working in IT industry for little over 7 years now and have some knowledge on technologies and programming concepts. This knowledge has proven to be immensely valuable during this small yet enjoyable journey.
Just for some context before I go deeper in my experiences of working with ChatGPT this is the list of features built:
- Simple UI where customers can select services they want to book, check availability on certain dates, choose a time slot and enter customer data needed for successful booking
- Logic for single and group appointments
- Simple logic of multiple employees available per appointment
- Integrated mail client sending transactional emails upon successful booking confirmation
- Small admin panel where admin users can have an overview of made bookings and manage their availability
Start of the journey
Of course I didn’t dive into coding right away. Even though the app was intended as a one-off project for my friends, being the proper product manager that I am, ChatGPT and I (referred to as ‘we’ from here on) spent way too much time on coming up with the catchy name.
After some back-and-forth, we settled on a term I genuinely liked, while ChatGPT remained neutral. That bothered me a bit since I get hyped up beyond any reasoning for smallest of things. Soon I will find out that ChatGPT is always neutral and we will never get to the same level of excitement over anything. Therefore I made decision for us both and the name we picked is Busy Bee Reservations!
Our good friend Adobe Firefly AI jumped in and created this super cute logo:
Coding
I started with small prompts:
- “Please build a small booking app for me using PHP, JS and MySQL.“
- “Can you fix the code so it works?“
- “It doesn’t work, can you make it work?“
That didn’t end up quite as well as I hoped for and is the sole reason why this article will have bit more paragraphs.
Key learnings:
- Effective Prompts: I had best results with small and thought out prompts. In a similar manner that you would break up tasks in pieces when building software with humans, you need to do it with AI as well but more, a lot more. Context is the key and more information you provide, output will be better. Focus on iterations and small steps. It made my life easier.
- Avoiding Prompt-Answer Loops:There is a big potential at any moment to get into the prompt-answer loop without engaging your critical thinking. Think about the solutions provided to you and whether they make sense. AI is there to assist you and enhance your work. It is not there to read your mind.
- Debugging Skills: Even though ChatGPT is amazing we stumbled upon numerous scenarios where I felt like we are in endless loop of reiterating the same problem over and over again without finding new solutions. In such cases, it’s important to engage your own problem-solving skills and adjust your approach. Plain reiterating of same challenge will not get you far
- Providing Context: ChatGPT is stateless and doesn’t hold data on your previous conversations, even though it seems like it does. That’s why it’s called AI, it picks up fast. Don’t be afraid of providing more context to it even though you might believe it already has it
- Thoughtful Code and Database Structure: Think about your code and database structure and which problem are you trying to solve. I’ve spent a biggest amount of time on building a feature that was already there (basically I could have commented out one field from my query to achieve the same end result)
Final Takeaways
ChatGPT is amazing technology. While I understand there is hundreds of booking apps on the market and this whole experience could have been avoided, I believe it wasn’t waste of time. I’ve learned a lot during the process and on certain moments was genuinely amazed by the potential of it.
It excelled at brainstorming, providing relevant data for consideration and is now my favorite tool I use on daily basis.
On other note, if there is someone reading this article asking themselves whether AI will replace developers, I can with absolute certainty say that will not happen, at least not in near future. I see AI as a tool that should make developer life’s easier and finally allow them to focus on bigger picture and ideas. It will take bit more time for technology to come to desired level, but it will get there.
And finally, for everyone that sticked with us until the end, I’m super proud to announce that our MVP is now live and actively used by customers. Here is a small GIF of Busy Bee Reservations in action.
And a small admin panel:
Leave a Reply