New York-based Evolutionary Operating System, or Evos, has emerged from stealth with $9 million in pre-seed funding to build AI operators for small and mid-sized industrial businesses.
The round was co-led by Uncork Capital and Field Ventures, with participation from Crosslink Capital, Utopia Capital Management, Sandwith Ventures and angel investors.
Founded in 2026 and led by CEO Urav Shah, the company is targeting businesses in freight, logistics, manufacturing and other physical industries where operational expertise often remains concentrated in experienced employees.
Unlike industrial robotics companies, Evos is not building machines that perform physical tasks. Its operators are software agents designed to carry out business processes inside the digital systems those companies already use.
Evos says its platform learns from how experienced employees make decisions, identifying patterns in operational behavior rather than relying only on written instructions. Customers can initially require human approval for every decision and gradually increase autonomy as the system demonstrates consistent performance.
The company reports crossing $1 million in annual recurring revenue within four weeks of going live.
One early deployment involves freight brokerage MyFreightWorld, where an Evos operator audits freight carrier invoices against transportation management system records. According to the company, the operator made nearly 3,500 audit decisions in two months at 97% accuracy and processed more than $600,000 in invoice value over a five-week period.
Evos also claims that customers can deploy a working operator within roughly 24 hours after an initial discovery process, rather than undertaking lengthy software implementation projects.
The company plans to expand through its Workhorse platform, which will allow businesses to create autonomous operators for additional operational functions.
The opportunity comes from a persistent problem in industrial businesses: critical knowledge is often embedded in employees rather than documented in software. When those employees leave, their expertise can be difficult to replace.
Evos is betting that AI can capture some of that institutional knowledge and use it to perform recurring work, rather than merely assist employees with recommendations.
The business now needs to demonstrate that its reported deployment speed and accuracy hold across different customers, systems and operational conditions. A successful invoice audit is a useful starting point, but broader autonomy requires reliable exception handling, clear accountability and evidence that the software can perform complex work without constant supervision.