From TIM and Plenitude to Euronext, two decades building innovation with enterprises.
Hoplo didn't start with AI agents. We started, over twenty years ago, by building software and innovation alongside some of the largest companies in Italy and Europe, telco, energy, financial markets, insurance and media.
The same instinct, new tools
That history is why the forward-deployed model comes naturally to us: we've always worked inside our clients' operations, for the long run. Vertical AI agents are simply the most powerful tool we've had to do it with.
Twenty years in regulated industries leaves a way of thinking that is hard to acquire quickly. Anyone who has worked with telco, energy, financial markets and insurance knows the constraint is almost never technical: it is that a system has to be explainable to a supervisor, survive a change in regulation, and keep working once the person who commissioned it has moved on.
What long projects teach you
Staying alongside a company for years means seeing what happens to software after launch, which is the part almost nobody talks about:
- Integrations that looked solid crack when a system changes on the other side, and it always does.
- Processes accrete: an exception is created for one special case, and five years later it has become the rule without anyone ever deciding so.
- The people who use a system daily know more about it than whoever specified it, and the fastest way to get it wrong is not to ask them.
Why we didn't start with agents
The skills that make an agent useful inside a company — understanding a process, integrating with systems that don't talk to each other, holding up under compliance requirements, staying when things get complicated — are the same ones the work required before. The tool changed, the craft didn't.
A more powerful tool, the same discipline
Vertical agents do something traditional software couldn't: handle work that requires reading, interpreting and deciding, not just executing rules written in advance. That's a genuine step change, and it's why the company is built around it today.
But the extra power raises the bar on everything else. A system that decides needs verifiable sources, guardrails and someone who answers when it gets things wrong — precisely the things twenty years inside client operations taught us to treat as the first consideration, not the last.