Evos bags $9M to build autonomous workers that learn from companies’ most experienced humans
SiliconANGLE Mike Wheatley
Evos raised $9M to build AI workers that learn by watching experienced staff. The twist: they only go autonomous after proving they can do the job safely.
Based on reporting by SiliconANGLE, Mike Wheatley — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Evos just pulled in $9 million to push a very specific idea into smaller industrial businesses: let AI learn the job from the people who already know how to do it, then take over parts of the work itself. The startup, officially Evolutionary Operating System Inc., says its focus is the physical economy — manufacturing, freight, and logistics — where experienced workers are retiring and know-how is walking out the door with them.
The round was led by Uncork Capital and Field Ventures, with support from Crosslink Capital, Utopia Capital Management, Sandwith Ventures, and unnamed angel investors. Founder and CEO Urav Shah says the company is not trying to replace people. The pitch is more practical than that. In businesses with fewer than 100 employees, the most valuable asset is often the person who has been there for years, and that knowledge is rarely written down.
Evos calls its platform Workhorse. The system does not start by acting on its own. It watches first. Each agent is trained on a narrow task — answering broker calls, chasing missing paperwork, reconciling invoices, or building freight loads — and sits behind an approval gate while humans review, modify, or reject its decisions. Only after it has learned the role and proven itself in real use, with constant human oversight, does it get to work autonomously. If it runs into something new, a person still has to sign off.
Shah says the average agent can reach about 82% decision accuracy after 24 hours of learning, and then handle around 70% of a human worker’s daily load on its own. That sounds less like sci-fi than process automation with a memory. The company says MyFreightWorld is already using the agents to audit less-than-truckload carrier invoices against its transportation management system, logging more than 3,500 auditing decisions at 97% accuracy over its first two months and helping process more than $600,000 in invoice value.
Evos says it went live with select customers only a month ago and has already passed $1 million in annual recurring revenue. The money from this round will go toward expanding the Workhorse rollout and growing the team, with general availability planned for later this year. The real tell here is simple: software that actually does the work is still rare, and that scarcity is exactly where these investors are betting.
My take — AI-written commentary, not fact-checked reporting
This is the right kind of AI pitch: boring, constrained, and tied to a job that actually exists. The industry has spent years pretending every workflow needs a chatbot; meanwhile, the people who know how invoices, freight, and approvals really work are still doing the heavy lifting. Evos is betting that the best model is the one that shuts up, watches, and earns trust before it touches anything important. That’s refreshingly unsexy, which is usually a good sign.
Read more about this at: SiliconANGLE