💡 AI Did Not Give Me New Ideas. It Gave Me a Way to Build Them.

Over the last few years we have talked a lot about artificial intelligence.

That it will replace jobs. That it will write code. That it will make companies more efficient.

Many of those things will probably happen.

But, at least for me, the biggest change has been something else.

AI drastically shortened the distance between having an idea and turning it into something real.

And I believe that transformation is still in its early stages.

Plenty of ideas, few realistic ways to build them

I have always been a curious person.

I like talking with people from different industries, understanding how their businesses work, looking for opportunities, and constantly asking myself how things could be done better.

For years I had many ideas. What I did not have was a realistic way to build them.

Creating an application, automating a process, or developing a product meant learning new technologies, finding time after work, and often depending on people with skills I did not have.

AI did not remove that effort. But it made trying to build stop feeling impossible.

A background different from development

I have never considered myself a developer. My career has been tied to another world.

For more than ten years I have led complex projects in telecommunications, infrastructure, and critical platforms: coordinating distributed teams, managing risks, planning deployments, and getting important changes into production without affecting millions of users.

Looking back, I think that experience turned out to be far more useful for working with AI than I imagined.

Because I never saw it only as a tool to generate text or write code.

I saw it as a new way to solve problems.

From handmade copy-paste to building what I need

I remember perfectly the first times I used ChatGPT: I asked for a fragment of code, pasted it into Visual Studio Code, it failed, I copied the error, and asked again. It was a slow and fairly handmade process.

Then Cursor, GitHub Copilot, Claude, Codex, and many other tools appeared. Each one reduced the friction a little more.

But the real change was never the models.

It was stopping asking what software already existed to solve a problem and starting to ask something different: what if I build the solution myself?

NeuraPRO: from numbers to decisions

One of the first projects where I really understood that change was NeuraPRO.

The initial idea was to help small businesses understand their business better. At first I was obsessed with building the application. Over time I realized the real value was somewhere else.

I remember the moment when, after entering sales, costs, expenses, inventory, and other data, the platform started automatically generating indicators, charts, and financial statements.

For the first time I could clearly see how much a business was actually making, where it was losing money, which products had better margins, where waste existed, and which processes consumed time without creating value.

I was no longer looking only at numbers. I was looking at decisions.

That project made me understand that data, on its own, has little value. What matters is turning it into useful information to make better decisions.

Automating an income statement, VAT calculation, or preparing information for a bank does not only save time. It also helps a company understand much better how it works.

And that way of thinking ended up influencing practically everything I built afterward.

The team Planner

Something similar happened at work.

For years I assumed that certain tasks simply “were that way”: planning deployments, reviewing engineers’ workload, linking activities to sprints and releases, moving dozens of tasks when a change was delayed.

All that information existed. But it was scattered across Jira, calendars, spreadsheets, and different conversations.

Until one day I asked myself the same question.

What if I build exactly the tool I need?

That is how a Planner for my team was born.

Today I can see each engineer’s workload, create an activity, assign an owner, and let the system do the rest: tasks are created automatically, story points, sprint, and release are assigned. If a deployment date changes, I no longer have to reorganize everything manually. I adjust the plan and the rest follows the change.

I did not build that tool because I wanted to become a developer.

I built it because I no longer wanted to keep wasting time solving the same problem over and over.

I discovered that I like writing

Curiously, AI also changed something I never expected: it made me discover that I like writing.

A few years ago I would never have imagined having a blog, a newsletter, or dedicating part of my time to sharing experiences about technology.

Writing forced me to organize my ideas better, to question them, and to explain them in a way other people could understand.

AI helps me a lot in that process: structuring a text, finding a better way to express an idea, detecting when something is unclear.

But the experiences are still mine. The opinions too. The mistakes too.

AI does not think for me. It does not make decisions for me either.

It simply gave me the support I needed to turn thoughts that used to stay in a notebook or in my head into something other people can actually read.

And I believe that has also been part of the journey. If you want to follow along, you can find me on LinkedIn.

Judgment, discipline, and mistakes

Not everything has been success.

I also broke repositories. I lost hours following the wrong paths. I received a fairly painful bill for consuming far more tokens than I imagined. I built projects I ended up pausing.

And I confirmed something I had already learned many years earlier working in production: projects rarely fail because of one big decision. They usually fail because someone forgot a dependency, because a validation failed, or because a small detail ended up breaking everything.

AI accelerates development.

But it still does not replace judgment, discipline, experience, or the ability to truly understand a problem before trying to solve it.

A small ecosystem

Looking back I realize I did not build a single project.

Without noticing, I ended up building a small ecosystem: applications, internal tools, a blog, a newsletter, and a personal brand that a few years ago I would not even have imagined developing.

All of them were born from exactly the same place.

An idea. And the possibility, for the first time, of trying to turn it into something real.

Does AI worry me?

Sometimes people ask me if AI worries me. The truth is not really.

What worries me much more is stopping learning.

Because if there is one thing I have discovered over these years, it is that this technology changes at a speed that is hard to keep up with.

Just a few years ago I could work perfectly well without AI. Today, honestly, I do not know how I would go back to doing it.

Not because I stopped thinking. Quite the opposite. Because it is now part of practically everything I do: learning, analyzing, writing, building, and solving problems.

If it disappeared tomorrow, I would feel like I was taking an enormous step backward.

And that makes me think: if in just a few years we already feel this dependence, what will happen when artificial intelligence is integrated into practically every profession? Will we become too dependent? Will it be a risk? Or are we simply living through a transition similar to what happened with the Internet or smartphones?

I do not have the answer.

The only thing I know is that AI did not give me new ideas.

It gave me the possibility of turning many more of them into reality.

The advantage is not having AI. It is knowing which problem to solve with it.

And I suspect that change is only just beginning.

Related reading:

• NeuraPRO: My MVP Worked. My Customers Didn’t.
• When Jira Is Not Enough
• The AI Bill

---

✍️ Claudio from ViaMind

“Dare to imagine, create and transform.”

---

• Previous
  🌐 The Invisible Work Behind the Internet: Change, Infrastructure, and Judgment

---