Evos raises $9 million for AI workers trained by industrial employees

Evos emerged from stealth on October 14, 2026, with $9 million in pre-seed funding for AI operators that learn jobs from experienced employees. It is targeting small and mid-sized industrial businesses, where the company wants software to carry out operational work inside existing systems—not simply help staff organize it.

Uncork Capital and Field Ventures co-led the round, joined by Crosslink Capital, Utopia Capital Management, Sandwith Ventures and angel investors. In its funding announcement, Evos said the money would support hiring and bringing its operators to businesses that make, move and power the economy. Its platform, called Workhorse, has not yet formally launched.

Learning through the approval gate

Evos describes its operators as learning from behavior: the decisions employees make and the patterns behind them, rather than written instructions alone. Each operator receives a function to learn. Examples described by SiliconANGLE include answering broker calls, chasing missing paperwork, reconciling invoices and building freight loads.

SiliconANGLE’s Mike Wheatley reported that an operator first works behind an approval gate. People review its proposed decisions and actions, then approve, change or reject them. Tasks it completes correctly become training data for the AI model. The system is designed to learn how a business handles the job, with employees correcting it as it goes.

According to Evos, an operator takes on a function fully autonomously only when its decisions consistently match the team’s. Wheatley also reported that unfamiliar situations still require human approval. Businesses can choose their level of oversight, from signing off on every step to allowing an operator to run independently.

Speed is another part of the pitch. Evos calls the platform self-serve and describes a sequence of one hour with company experts, a working operator the next day and measurable results within a month. These are the company’s deployment promises, rather than a timetable independently demonstrated across customers.

Preserving judgment before it leaves

Founder and CEO Urav Shah says industrial businesses are losing experienced employees faster than they can replace them. Much of their expertise, he says, has never been written down. Evos aims to capture that judgment and use it in daily operations. Shah told SiliconANGLE he wants to address the labor shortage, not replace people.

Shah also gave SiliconANGLE a more specific performance claim: an average agent reaches roughly 82% decision accuracy after 24 hours of learning. At that point, he said, it can handle around 70% of a person’s daily workload autonomously. Those claims describe partial coverage of a worker’s day, not flawless execution or independence on every task.

An invoice audit is the clearest customer example

Freight brokerage MyFreightWorld uses an Evos operator to compare carrier invoices with its transportation management system. Evos says it made nearly 3,500 audit decisions at 97% accuracy in the first two months. That figure covers one workflow over two months, unlike Shah’s claim about an average agent after one day.

Evos says the operator closed out more than $600,000 in invoice value in the last five weeks. That is invoice value handled, not a stated amount of savings. In the announcement, MyFreightWorld president Mike Chalfant credited Evos with improving billing accuracy and efficiency, and said the company wanted to add more functionality.

Evos says it passed $1 million in annual recurring revenue four weeks after going live. The funding therefore accompanies claimed early commercial traction, rather than a completed public launch. The company says Workhorse will formally launch in the coming weeks.