Oleg Rogynskyy and the week nobody could sell

The sales team had been grounded. For a week, the people employed to bring in money would sit with Salesforce and clean up its records. Oleg Rogynskyy was an inside sales representative at Nstein Technologies, and he hated the assignment. Somewhere between the actual conversations with customers and the software meant to record them, the business had lost track of itself. The repair required everybody to stop doing the work the software was supposed to help.

This might have remained an ordinary workplace grievance, the sort recalled over a drink with former colleagues. Then Rogynskyy started Semantria and found himself imposing the same week of administrative detention on his own team. Later, at H2O.ai, he encountered the problem again. Three companies, similar mess. Technology had advanced; the sales department was still tidying its room.

In 2016 he founded People.ai. The proposition began with an unglamorous question: could software collect the evidence of selling without requiring the seller to become its clerk? Emails, meetings and calls already left traces. If those traces could be connected reliably, a manager might understand a deal while a salesperson got on with it. A repeated irritation had become a company.

Before the spreadsheet, the party photographs

Rogynskyy grew up in Dnipro, the Ukrainian city then called Dnipropetrovsk. His early ambitions included studying abroad. He attended an English-focused school and eventually went to Boston University. There was an entrepreneurial rehearsal there, though it had little to do with the enterprise software that would later define his career: photographs from nightclubs.

His first venture began as a magazine featuring club photographs. When that failed to find its footing, he moved the idea online. Facebook was emerging from Harvard, and a feed of party pictures seemed a plausible business. It did not work out. There is something useful about this beginning: the later seller of practical software had first tried selling people the pleasure of seeing themselves having a good time.

His education was also less predictable than the AI résumé might suggest. He studied political science and international relations. In a 2017 interview, he recalled two grandmothers who were teachers, a childhood full of reading and science fiction, and a university education that taught him how to investigate a question. He also admitted nearly failing his single computer-science course.

He came to regard those studies as preparation for sales. Understanding how someone thinks, establishing a relationship and discovering a problem were useful commercial skills. His path into machine learning ran through customers. The boy interested in computers became an adult interested in what people needed computers to do.

Five questions before an answer

Semantria, founded in 2011, brought him into text analytics as an entrepreneur. The business offered cloud-based analysis of written material; Lexalytics acquired it in 2014. He later served as Lexalytics president before joining H2O.ai. The sequence put him close to language technology well before the current enthusiasm for conversational AI.

His account of building Semantria makes sales sound more like an extended apprenticeship than a performance. He and his co-founder spent about nine months building the product and securing their first customer. During that interval he talked to prospects, learned what mattered to them and fed those discoveries into the product. Selling and deciding what to build occupied the same conversation.

He described a rule of asking five questions before offering one answer. A demonstration could wait until a third or fourth meeting. It is an awkward policy for anyone who has lovingly polished a slide deck. The customer is allowed to remain the most interesting person in the room.

That habit followed him into People.ai. At Y Combinator in 2016, an early idea involved scoring salespeople for hiring. Conversations with potential buyers changed the application of the underlying data platform. The question widened from judging a candidate to understanding the work already happening inside a company.

He later told interviewer Sramana Mitra that People.ai reached 100 proof-of-concept customers during its first 90 days at the accelerator. Those were experiments with customers, a useful distinction from a mature enterprise deployment. The important movement was toward a problem buyers recognized well enough to try solving.

What the calendar knew

A customer relationship management system has a deceptively simple job: remember the customer. In practice, its memory depends on busy employees recording the right details, connecting them to the right account and returning often enough to keep them useful. Every omitted meeting makes the official picture a little less like the business outside it.

Rogynskyy saw an opportunity in information that companies already generated but struggled to assemble. People.ai captured contacts and activity from the tools revenue teams used, then organized that information for sales, marketing and customer success. The ambition extended beyond a cleaner address book to decisions about accounts, opportunities and coaching.

By 2021, the company said its AI had been trained on $239 billion in closed-won enterprise deals and 1.1 billion unique sales activities. Those are measures of the material its systems had processed, rather than a tally of revenue People.ai itself earned. They show how far the company had travelled from a sales team fixing records by hand.

In August that year, People.ai announced a $100 million Series D at a $1.1 billion post-money valuation. Akkadian Ventures and Mubadala Capital co-led the round, with existing investors including ICONIQ Capital and Lightspeed participating. Earlier backing included Andreessen Horowitz. The business born from administrative frustration had acquired an unusually expensive audience.

The chili recipe that stopped the roadmap

More data brought a more delicate question: which information belonged in the system at all? Around 2017, a customer reported that an unexpected edge case had shared somebody’s chili recipe with colleagues. Rogynskyy recounted the episode the following year. Dinner had wandered into the office software, and the company took the warning seriously.

People.ai stopped feature development and concentrated on a privacy protection layer. His telling is candid about the distraction of shipping products quickly and the need to change priorities. A recipe was a comparatively modest exposure, but it revealed a mechanism that could mishandle something more consequential. The response was to interrupt the roadmap before a worse example arrived.

His stated management practices also made room for uncomfortable information. Employees were expected to spend time with customers, and board members held sessions with the team while the CEO stayed out of the room. For a founder who praised listening, the empty chair was a concrete way to practise it.

“I always felt like I was herding cats.”Oleg Rogynskyy, on managing CRM data, 2018

The distance to Ukraine

Ukraine remained part of his working life. People.ai had a team in Kyiv, and he mentored Ukrainian founders. After Russia’s full-scale invasion in 2022, his connections acquired a different urgency. In a public account of the company’s preparations, he described planning that had begun the previous November and moving most affected team members to locations in the European Union and western Ukraine.

He discussed mobilizing technology for Ukrainians, including connectivity, in a BreakLine conversation that year. UFORCE’s account of his career credits him with helping secure commercial satellite intelligence and communications early in the war. In August 2022, President Volodymyr Zelensky awarded him Ukraine’s Order of Merit, third class. The presidential decree names the People AI founder among the recipients.

He also founded Freedom Fund to invest in Ukrainian defense technology. His network of founders, engineers and investors now reached into a field where procurement, production and deployment carried a different weight. The transition grew out of his ties to Ukraine and work already under way, rather than a sudden discovery that hardware was fashionable.

In 2025, he passed the People.ai CEO role to Jason Ambrose. The company described a planned succession after the two had worked alongside each other on strategy. Rogynskyy publicly welcomed continuing the relationship from a board and strategic perspective. The distinction matters to his story today: he founded People.ai and led it for nearly a decade; he now leads UFORCE.

A different kind of feedback

UFORCE brings together Ukrainian-developed unmanned systems for maritime, aerial and ground use. Its leadership includes board chairman Oleksiy Honcharuk, a former Ukrainian prime minister, and former British defence secretary Ben Wallace. In March 2026, the company publicly emerged with $50 million in funding at a valuation of about $1 billion. Its corporate base is in Britain, with the international structure intended to help attract investment and expand production.

Rogynskyy’s work has moved into factories, partnerships and the practical differences between operating environments. In August 2026, UFORCE announced the MV7, its first UK-developed autonomous sea drone, designed for Atlantic temperatures and sea conditions. A system proven in one body of water still needs engineering for another. Experience supplies a starting point, followed by fresh work.

In a September interview, he discussed recruiting Formula One specialists and maintaining research and development teams abroad while keeping the largest share of staff in Ukraine. The interest in motorsport engineers has a practical explanation: engines, sensors and structural materials. A racing résumé can turn out to have applications well beyond a racing circuit.

There is a recognizable concern running through these otherwise different businesses: the gap between an official plan and what the user encounters. At People.ai, that meant asking what a seller actually did all day. At UFORCE, it means adapting systems to conditions that refuse to cooperate with a presentation.

The younger Rogynskyy had wanted to put party photographs online. The salesman wanted his week back. The founder learned to ask enough questions to discover what somebody else wanted. Today the questions concern Ukraine, allied countries and the work of building at scale. His career has taken a long route from that room full of grounded salespeople. The calendar, at least, has become more interesting.