AI Might Be the Most Visible Example of Continuous Improvement in My Life — LLeanStore Journal

I used to ask AI for a better sentence. Now I ask whether I can build something I once assumed was beyond me.

Somewhere between those two questions, my plans became rather complicated. Haha.

### It started with small questions

When I first started using AI in 2023, my requests were simple. Translate this. What does this word mean? Which Excel formula should I use?

In 2024, I started brainstorming with it. I brought questions that needed more than one correct answer. I could explain an idea badly, go back, add context, and try again.

In 2025, I started coding with it.

That changed the size of the questions I was willing to ask. A problem at work could become something I might build a tool for. I still had to understand the problem, check the result, and deal with whatever broke. But I could attempt things I would previously have left as an idea in a notebook.

By 2026, I was thinking about something bigger: could this give me a fairer chance to compete?

### What does a fair chance look like?

Earning in rupiah and paying for something priced in dollars makes the distance between local income and global participation feel very real.

I remember looking at my salary and feeling that progress in rupiah did not necessarily feel like progress when I thought in dollars. A subscription might look affordable on someone else's screen and require quite a bit of consideration on mine.

That frustration became a personal question. How could I, with my background and resources, make something useful for people beyond Indonesia?

I enjoy riddles. I like following clues and figuring out why something does not work. In the office, that curiosity usually attached itself to a process: a recurring problem, confusing information, or work that seemed harder than it needed to be.

But curiosity alone does not give me every skill I need.

My books are still piled up, waiting to be read. I am not especially comfortable speaking on video. Working in another language takes effort. I cannot pretend I have the strongest academic background in every subject I want to explore.

AI gave me a way to start with the knowledge I already had and work on some of those gaps as I went.

I do not think it makes everyone's opportunities equal. I still need time, access, judgment, and people willing to try what I make. But in my own experience, it has made certain possibilities feel closer.

So when people are curious about my step back from employment, I understand. I am curious about my own track, too.

### Helpful, frustrating, and changing

I neither love nor hate AI without reservation.

It has helped me many times. It has also made blunders, sometimes with an impressive amount of confidence. I have learned to be careful with an answer that sounds complete before I have checked whether it is correct.

Some days, the conversation feels familiar, almost like talking to a friend. On other days, I wonder who this new one in front of me is. The tone changes. The explanation changes. Something that worked before needs explaining again.

Yet, looking back over these years, I can see a difference in what I am able to do with it. Translation became discussion. Discussion became code. Code became tools I could put in front of other people.

That is why AI feels like such a visible example of continuous improvement to me. I can trace the change through my own work.

Still, I need to be careful with that comparison. A new version, a longer answer, or another feature is not automatically an improvement. I need to see whether it helps with the actual task. Sometimes, the newer answer creates more checking and rework.

I wanted to understand the flow. I needed a tool. I wanted to compare workloads. Another tool. I wanted to explore a layout. Another tool, another price, another thing to learn.

Sometimes the cost was the obstacle. Sometimes it was the effort of moving information between files, rebuilding a chart, or trying to make Excel behave the way I needed.

I could spend so much time preparing to discuss an improvement that there was less time left to try it. Then came the meeting about the improvement.

I know a diagram cannot fix a process by itself. Someone still has to observe the work, talk with the people doing it, test a change, and follow through.

That is exactly why I wanted the preparation to become easier.

If a tool helps us see an uneven workload, I want to spend more time discussing that workload with the team. If it helps us draw a value stream, I want the conversation to reach the actual flow sooner.

This frustration is part of what led me to build LLeanStore. I kept thinking about the tools I wished I had while doing the work.

### “I will go back after the VSM is done”

I told a friend I would return to work after finishing the VSM tool.

Then I kept having ideas about what else we should have.

A map showed the flow, but I also wanted to examine the workload. Looking at the workload brought up questions about the layout. Each useful thing made another missing piece easier to notice.

I enjoy that part. Perhaps a little too much.