Posted by Apple Podcasts ¦ Amazon Music ¦ Spotify
Hi, and welcome to Try AI for Growth, a baby podcast out of Make Space for Growth. Here, I share short and sometimes surprising stories of how I’ve used technology to tackle everyday challenges — at home, at work, in business. I’m your host, Sara Vicente Barreto, and today I want to tell you about my first experience building an app with Lovable.
Starting with… no ideas
Now, I should start by saying this was not some grand, strategic, perfectly planned experiment. Quite the opposite. It started with me sitting in a conference workshop, surrounded by people already typing away, and suddenly having absolutely no idea what to build. Which is slightly ironic, because I am usually not short of ideas. If anything, my problem is often having too many. But there I was.
We had spent the day reflecting, learning, talking about AI, tools, possibilities, the future of work, all the things that usually make my brain light up. And then came the practical session: now go and build your first Lovable app. My mind went completely blank.
Calling a friend
Everyone around me seemed to be starting. People had concepts. They had use cases. They had prompts. They looked focused. And I felt that very familiar pressure of: I should know what I am doing by now. I looked to the friend next to me, and she stared back, hoping I would come up with something for us. The pressure built further.
So, naturally, I turned to my thinking partner. I asked ChatGPT: Given all you know about me, what could I build? The answer was surprisingly kind. It basically said:
“You are not short of ideas. You are just too full of them and don’t know where to start.”
Which, honestly, felt a bit accurate. My shoulders relaxed as I read through its complex ideas. I wanted something simpler, so we started narrowing it down.
Re-design before you design
Instead of thinking, “What is a cool app I could build?”, I shifted the question to: “What process could be redesigned?” And that shifted my perspective. Because suddenly I wasn’t trying to invent something impressive for the sake of the workshop. I was looking at real friction. Real work. Real bottlenecks.
And one idea stood out. At the charity, one of the things we constantly need is better data collection from the field. Field officers collect information from families, children, activities, visits, outcomes, challenges, and case studies. They do it on paper, then they type it into Excel, then they send it to us through Google Drive. After that, we clean it for typos and language, then we import it into Salesforce, then we send it to sponsors, then we may do a case study or two. It is a long and bumpy voyage for this data. The truth is, more time is spent handling data than extracting real information. And rarely do we have the chance to look at it in an aggregate manner, define red flags and follow-ups, or consistently extract case studies.
I often wonder how the process can be different. So here was my chance. I started building an app that would help collect better data from the field, and then allow that information to be extracted into different functions: database updates, marketing content, sponsor emails, and project reporting. Rather than thinking of the task of “cleaning up data”, I took a full view of the process instead. That changed everything.
Lovable was not just helping me automate one tiny task. It was helping me imagine a better process.
Yet again, I was hooked.
The first attempt
For my first attempt, I did not go straight into Lovable with a half-formed idea. I stayed with ChatGPT to create a more complete prompt. I wanted to describe the app properly: who it was for, what data it needed to collect, what outputs I wanted, and what kind of user experience would make sense. Moreover, ChatGPT knows my work at the charity, understands a lot of my reports and tone, and I was not yet ready to leave familiar ground.
Could I have done all the experimenting directly inside Lovable? Probably. And I suspect next time I will. But for my first attempt, I wanted to give myself a better starting point. I wasn’t fully convinced that my prompt, on its own, would be good enough.
And perhaps that is one of the first little lessons here: sometimes using AI well is not about using one tool. It is about knowing which tool can help you think, and which tool can help you build. ChatGPT helped me shape the idea. Lovable helped me turn it into something real.
What surprised me was how quickly Lovable started to understand the direction of the app. It wasn’t just waiting for me to specify every single button, line, rule, and screen. It was interpreting the context. Suggesting additions I may want.
That matters. Because for non-coders like me, the barrier is not only “I don’t know how to code.” It is also “I don’t know what the app should technically include.” I may understand the problem. I may understand the users. But I don’t always know the mechanics of making it work.
Lovable bridges that gap. I was amazed.
Ready for more
So amazed that I was ready for more as soon as I got home from the conference. When I got back, I sat down with my son, and we built two more apps. The first was a fun experiment for his spelling homework. Every week, he has to practice spelling words. And every week, he hates writing a story with them. Not dislikes. Hates. So we built an app where he could enter his spelling words and generate funny stories using them to spark his creativity.
Suddenly, the dreaded homework became a game. The app could take a short list of words and create a silly story around them. I did not need to provide hundreds of examples of funny stories. I did not need to pre-write the humour. The app interpreted the task from the small set of inputs we defined. And he loved it. Not just because the app was useful, but because he had helped create it.
When we shared it with the family WhatsApp, his cousin in the US came back with his story. That is when the penny dropped for him. Can other people play my game? Have we just done this in 7 minutes? I have a game?
He was immediately keen to do another one, this time a football gaming app, focused on tactics. He explained the main concept, and the app did the rest. All in under 10 minutes. Lovable then suggested improvements such as adding scoring to a game or creating a leaderboard. It understood the type of thing we were building and started proposing features that made sense. My son was ecstatic. He wanted me to send it to everyone. He could not wait to go to school and say he had done an app.
That was probably my favourite part of the whole experiment. Not the charity app. Not the workflow improvement. Not even the speed. It was seeing a child realise: I can make something other people can use. With no coding knowledge.
A new way of thinking about technology
This creates a very different relationship with technology. Not just consuming. Not just watching. Not just playing someone else’s game. But making.
And yes, these were simple apps. I am not pretending we built the next Salesforce or Football Manager at the dinner table. But that is not the point.
“The point is that the distance between idea and prototype has collapsed. “
And when that happens, it changes how you think.
Instead of asking, “Could someone build this one day?”, you start asking, “Could I test this today?” That is a big shift. It does not have to be an end product, but it gives you immense freedom and also huge flexibility in testing out ideas.
Lessons Learnt
So, what did I learn?
- Broaden your horizons.
When we think about AI, we often think about single tasks. Write this email. Summarise this document. Create this image. Draft this post. And all of that is useful. I have been bringing quite a few of these to the podcast. But this Lovable experiment brings us to something bigger. Not necessarily bigger in complexity, but bigger in process. What is a workflow that frustrates you? What information do you collect repeatedly? Where does data get stuck? What do you keep copying from one place to another? What would be easier if there were a small interface around it? That is where app-building becomes interesting.
- Work on your idea.
The quality of the starting prompt matters. Even if Lovable can guide you, the better you can explain the user, the problem, the inputs, the outputs, and the desired experience, the better your first version will be. This is not about writing perfect prompts. It is about thinking clearly. Who is this for? What are they trying to do? What should happen first? What should the app produce? What would make it easy to use? If you can answer those questions, you are already much closer to building something useful.
- Let the app guide you.
This was a funny one for me, because I like to have a plan. But Lovable is built for people who are not coders. It expects you to iterate. It expects you to try, publish, test, and adjust. And it starts suggesting improvements. Add scoring. Add a leaderboard. Improve the flow. Make the interface clearer. That guidance is part of the process. You don’t need to know everything upfront.
As a bonus, stay connected and build it with someone else. Doing this with my son changed the energy completely. It made the experiment playful. It also showed me how quickly AI tools can become collaborative tools — not just between you and the AI, but between you and another person. A child can describe the game they want. A parent can help shape the prompt. The tool can build a first version. Then everyone reacts to what appears. It is a very similar process when you want to design an app with a teammate with specific expertise you may not have. It brings a different type of perspective and creativity.
Don’t dismiss simple apps
The spelling story app was simple. But it addressed a pain point. He still has to do the story on his own, but he can go to the app after it is done to get extra ideas or see what other alternatives would be. The football app was simple. But it gave my son a feeling of agency and delight. The field officer data app was only a prototype. But it opened up a completely new way of thinking about reporting and data flows. And we are now engaging in a process of redesigning how all these processes can operate in the presence of AI.
“Simple does not mean useless. Sometimes simple is exactly what gets you started.”
So, what can you try?
If you have never used Lovable or a similar app-building tool, don’t start by asking, “What impressive app should I create?” Start smaller and more honestly. Think of one process that annoys you.
- Maybe it is collecting information from clients.
- Maybe it is preparing a weekly report.
- Maybe it is helping your child practise homework.
- Maybe it is turning notes from the field into useful outputs.
- Maybe it is a game, a tracker, a checklist, a form, a dashboard.
Then describe it in plain English. The user, the inputs, the outputs. What would make it functional, enjoyable or easier? And if you get stuck before you even begin, do what I did: ask the AI you are more familiar with to help you find the idea hiding inside the chaos. Because sometimes you are not short of ideas. You are just too full of them.
This experience makes it clear that AI is no longer only helping us write, summarise, or brainstorm. It is starting to help us build. And for people like me, who are full of operational problems but not full of coding skills, that is exciting.
It also reminded me that the first version does not need to be perfect. It just needs to exist. Once it exists, you can react to it. Improve it. Show it to someone. Laugh at what doesn’t work. Build again. And maybe, in under ten minutes, you end up with something that makes your child excited to practise spelling. Or something that allows your co-workers to suddenly see the possibilities. Both count.
Thanks for joining me on this episode of Try AI for Growth. If you try building your first app with AI, I would love to hear what you create — especially if it starts as a tiny experiment and turns into something unexpectedly useful.
Until next time, keep experimenting, keep building, and keep having fun.