I didn't set out to start a company
September 6, 2026 · 2 min read
I didn't set out to start a company. I was a student building agents to do my homework.
Not to optimize anything. I didn't have work to optimize. I built them because it was fun, and because it turned out to be the fastest way I'd ever found to learn something. Mostly I used them to write code.
Then I built one that was too big to be a personal tool, and I called Santiago.
I still don't have a clean way to describe the months after that. The closest I get is existential psychosis. You watch a machine finish in seconds what should have taken you a week, and you can't unsee it. After that you start looking at businesses and noticing the same thing everywhere: hours of human attention spent on work that stopped requiring it.
So we started Axon Analytics.
We began with consulting, because everyone does. We sat inside a lot of companies in Guatemala and learned more about how a business actually runs than any classroom taught us. But we were always visitors. Never deep enough into one domain to honestly claim we understood it.
So we started building products. For a while we shipped things that worked beautifully and nobody bought. Underrated education.
What finally clicked was simple. Stop selling technology. Put it to work.
Today we run five agents. Ali, Edgar, Laura, Lupin, Clara. Each one holds a job, not a feature. They sell, analyze, buy, report, and keep customers close. And they run inside the systems each company already uses, because nobody is asking for one more login.
That's the thesis. Agentic AI as the operating layer, not a tool bolted to the side. Agents that compound revenue, collapse cost-to-serve, and absorb the operational work that was never supposed to be manual.
Within a few years every serious operation will run on a team of agents that never sleep, never lose context, and sharpen with every transaction. The companies that adopt that layer first will out-execute the ones that wait. I'm not hedging on that one.
We're still building. Next is a platform where regular people and regular businesses can use agents without it feeling technical. Simple, and honestly kind of fun. More when it's ready.
Last thing. If you're a young founder in Dallas, I want to meet you. Send me a message with what you're building and why you started. I'll tell you mine.
Originally posted on LinkedIn.
Recent Posts
- Nobody wins the AI race. That is the point.September 18, 2026 · 4 minWhy I agree with the Dreamforce keynote, what really happened at Hugging Face, and my thesis on agents.
- Building a great product is table stakes. Distribution is the moat.August 20, 2026 · 1 minWhat launching Axon Quiniela, our first consumer product, taught me about velocity, conversion, and where the real difficulty lives.
- Existential psychosis: what building with the latest models feels likeMarch 20, 2026 · 2 minBriefing agents across research, engineering, finance and validation, all running in parallel, and staring at the screen when the output lands.