01/Why we exist
The companies that need serious AI infrastructure most couldn't access it.
The large firms ignored them. The cheap agencies couldn't build what they needed. There was nothing in the middle that combined real technical depth with the speed and operational knowledge mid-market physical businesses actually require.
Nextas was built to close that gap. It eliminates the manual operations that drain time and money from US mid-market firms, with custom AI infrastructure built to last, not built to demo.
02/How we operate
Diagnose before prescribing
We run a paid audit before we propose anything. If AI isn't the right move yet, we'll tell you.
Build for production, not for demos
Everything we build has to run in a live operation, handle exceptions, and still work in two years.
Stay and be accountable
The monthly partnership is required because we want to be financially responsible for whether it keeps working.
Speak plainly
If a system saved 18 hours a week, we say 18 hours a week. No optimism bias, no marketing language.
Work in the industries we know
Construction, logistics, manufacturing, professional services. Four industries, not fifty.
Why audit-first
42% of AI initiatives were abandoned in 2025, up from 17% the year before. Most failed because they were generic tools applied without operational context.
03/The founders do the work
Andrea (Andrew) Barbera
Runs every audit
Andrew has spent two years deploying AI systems inside US construction businesses. He knows how owners talk, what they measure, and the exact point in a growth conversation where operational chaos becomes the real problem. That's what he listens for in every audit.
Before that he co-founded and scaled a software startup, then served as Head of Startups and Venture Capital, leading a team of 9 across 25+ startups. He studied business and AI at UC Berkeley's Haas School, was part of the Silicon Valley Fellowship, and won a US collegiate startup competition.
LinkedIn ->
Nikola (Nick) Fatori
Builds every system
Nick is the one who builds it. His systems run in real operations, not demos. They plug into the software a company already uses, pull structured data out of messy documents and feeds, and run agents that act on their own.
Before Nextas, Nick spent more than two years building AI systems that actually run in production, mostly for e-commerce brands and the people running their operations. One was a €10M brand bleeding $30,000-$60,000 a quarter in ad spend on products that had quietly sold out. The ads kept running, the stock was gone. Nick's system caught it and took over the media buying. He also builds Hermes, an agent that runs like a chief of staff rather than a chatbot: it handles the reporting, works through the data, and gets ahead of problems instead of waiting to be asked. He brings the same instinct to his consulting work: take a company AI-first and cut what it's wasting. A German construction firm is one of them.
LinkedIn ->