The CEO and co-founder of Aquant has spent a decade teaching software the dialect that field technicians use when something expensive stops working. The vocabulary is unusually specific. So is he.
Aquant does not sell chatbots. Aquant sells the closest thing a company can buy to the retiring technician who has been fixing your CT scanners since 1998. The company is a generative AI platform for field service, and Shahar Chen has been running it since 2016, which is longer than most people have been saying the phrase "generative AI" out loud.
The problem Aquant solves is boring in the way that expensive problems tend to be. A hospital owns an MRI machine. The MRI machine breaks. The technician who knew, from twenty years of repairs, that a particular error code usually means the coolant line is bent - that technician is retiring. His replacement has been on the job nine months. Somewhere inside the company's ticketing system there are ten thousand pages of notes that describe exactly the fix, written in a private language of abbreviations, part numbers, and idioms that nobody bothered to standardize. Aquant reads that language. Then it reads it back, out loud, to the nine-month technician holding a wrench in a hospital basement.
That is the pitch. Chen has been giving it, in one form or another, since before OpenAI was famous.
"AI is not here to replace anyone. AI is here to help us be the best we can be at what we do." — Shahar Chen, Authority Magazine
Most founder biographies are compressed. Chen's is not. He grew up in Israel with two hobbies - playing soccer, which he did badly and admits it, and taking computers apart to see what was inside. Only one of those hobbies became a career.
He studied computer science at Bar-Ilan University. His first engineering job was writing network-monitoring software for Mekorot, Israel's national water utility, which is a workplace where "unplanned downtime" has a slightly more literal meaning than in most SaaS companies. Then he joined ClickSoftware, a field service management company, and stayed for approximately fourteen years.
Fourteen years is a long time for anyone. It is an extremely long time in software. During it, Chen was, in rough order: a Technical Support Engineer, then a Senior Support Engineer, then a Solution Specialist, then Director of the Solution Center, then, briefly, an Account Executive. He was on the front line of hundreds of service deployments. He read the tickets. He watched customers describe the same problems in five different vocabularies. He watched the industry age.
Around 2016, he called his eventual co-founder Assaf Melochna, a former Israeli intelligence officer with his own service-industry résumé, and proposed building the company that Chen had spent fourteen years wishing existed. They incorporated Aquant in New York. They have been co-signing everything since.
Software Engineer at Mekorot, Israel's national water company. Writes network monitoring applications for a system that, if it fails, people notice.
Joins ClickSoftware. Cycles through Technical Support Engineer, Senior Support Engineer, Solution Specialist, Director of the Solution Center, and Account Executive. Watches the field service industry from the ticket queue up.
Co-founds Aquant in New York with Assaf Melochna. The premise: expert service knowledge is trapped inside people who are about to retire, and a machine could learn it if anyone bothered.
Aquant announces $70 million Series C. Total funding crosses $110M. Customer roster spans medical devices, printing, food equipment, industrial machinery, and utilities.
Aquant ships Service Co-Pilot, its generative AI product. Chen debuts it at Field Service Palm Springs and later says he had been worried they moved too fast. Competitors were still theorizing.
Aquant expands offline-capable AI, multi-agent architectures, and integrations with Salesforce and other CRM/ERP systems. Chen keeps writing at AI Business and Forbes.
"If the first version of your product doesn't make you slightly embarrassed, you've probably launched too late." — Reid Hoffman, quoted by Chen when asked how he ships
"AI is not here to replace anyone. AI is here to help us be the best we can be at what we do."On the value proposition
"If the first version of your product doesn't make you slightly embarrassed, you've probably launched too late."On shipping (via Reid Hoffman)
"In conservative industries, no one buys the model. They buy the confidence that the model won't embarrass them."On selling AI into hospitals and factories
Chen grew up playing soccer (badly, he says) and taking apart computers. Only one of those scaled. The other, presumably, is why he does not talk about soccer in interviews.
Chen's tenure at ClickSoftware was not a delay before founding Aquant. It was the research. Every service ticket he read became a data point in a future product spec.
He and Assaf Melochna, he says, instinctively know which of them should answer a given customer question. This is either lucky or a lot of practice. Probably both.
Aquant's headquarters is 147 W 24th Street in Manhattan. This is unusual for an Israeli-founder company and probably deliberate. The customers are here.
Asked what he would do outside Aquant, Chen has said he would like to start a Universal Access to Education movement. It rhymes with what he does for a living: making expertise portable.
Chen has said he was worried Aquant moved too fast with Service Co-Pilot. Then he attended the Field Service Palm Springs conference and found competitors still describing generative AI in future tense.
Aquant ingests a company's service data - the tickets, the schematics, the technician notes, the error-code taxonomies, the equipment histories - and trains service-specific models that answer questions like "what is wrong with this machine and what should I do about it." The models speak the customer's internal vocabulary, not a generic English. They work in medical devices, printing, food equipment, industrial machinery, aerospace, utilities, and automotive. They work offline, which matters when you are inside a factory or a submarine or a rural clinic. They integrate with Salesforce and other systems of record.
Investors have put in $112.6M. Roughly 170 people work there. The publicly reported annual revenue is about $26.5M. The customer stories are quieter than most AI companies would tolerate, which appears to be the point. Chen has said, more than once, that trust is the actual currency in conservative industries. Aquant sells trust, delivered as software.
"Trust is the currency of enterprise AI. In conservative industries, no one buys the model. They buy the confidence that the model won't embarrass them." — Chen, on selling into medical devices and utilities
In a rare personal answer during a 2023 interview, Chen said that if he could start any movement, it would be one for universal access to education. He did not say when. He also did not say he had a plan. It is worth noting, though, that the throughline is consistent: Aquant is, in a somewhat industrial way, an education product. It captures what someone knows and delivers it to someone who does not. Chen's day job and his stated aspiration are, at bottom, the same problem.