Logispot

We started with two people.

One of us knew logistics inside out. The other knew software engineering.

We didn't start with an idea like “Let's build something with AI.” We started with a problem we saw in logistics and wanted to solve.

First, we did customer development. We talked to potential users, asked about their current processes, their problems, and what they were already paying for. After enough conversations, we had a clearer idea of what was actually worth building.

Then we built the MVP.

It wasn't beautiful. It wasn't full of features. It just solved the main problem we wanted to test.

And then came the less exciting part: finding users.

We sent emails. We did LinkedIn outreach. We followed up. A lot. Some people replied, most didn't. Some agreed to try the product, some gave us feedback, and some told us things we didn't really want to hear.

But this part was extremely useful. Every conversation helped us understand what people actually needed instead of what we thought they needed.

Next, we started experimenting with AI coding agents. At first, we just used agents as coding assistants. Then we thought: what if we stop treating them as assistants and start treating them as a development team?

So we started building our own AI software development team. Now, humans navigate the work. AI agents write code. Other agents review and verify it. Humans check the important user cases.

So we release features and fix bugs much faster than a human team of 5-7 developers. We continue our outreach and think about other selling strategies...

To be continued... 🚀