Ariel Cohen and the reservation that disappeared

Ariel Cohen had a hotel reservation and no hotel room. In 2013, he arrived in Odesa, Ukraine, on a freezing night with meetings to prepare for and discovered that his booking had disappeared. The agency that arranged it was closed. The booking tool offered no useful rescue. Somewhere, presumably, a system had done its job. Cohen was still without a bed.

There is a particular comedy to corporate travel: a company sends someone across borders to conduct important business, then leaves that person negotiating with a confirmation email. The arrangement looks orderly from a desk. At the hotel, order has become a matter of opinion.

Cohen kept thinking about the problem. Years later, as co-founder and CEO of Navan, he would return to that night when explaining why the company existed. A trip has to work for the person taking it. The software should help when the itinerary stops behaving.

Two buildings, then two companies

His preparation for that task began in enterprise technology. Born in Jerusalem and raised in Israel, Cohen entered the technology business in 1999. He later earned an Executive MBA at Northwestern University’s Kellogg School of Management. Entrepreneurship appealed to him as a shared undertaking, with the excitement and uncertainty of a ride that keeps moving.

That background matters because Cohen approached travel with experience of the tools people use at work. Corporate software has several audiences. Someone buys it. Someone configures it. Someone else must use it on a Tuesday afternoon. Those people can have quite different definitions of success.

At Mercury Interactive in the early 2000s, Cohen and Ilan Twig worked in separate buildings, M1 and M2. They heard about one another before they collaborated. Once they worked on a project together, they recognized a shared approach. Their careers continued through HP, which acquired Mercury, and through work involving Opsware, the company founded by Ben Horowitz and Marc Andreessen.

Twig moved to the United States in 2006; Cohen followed in 2009. Their eventual partnership had a history before it had a pitch. That gives the familiar phrase “co-founder chemistry” some substance: these were colleagues who had seen each other work.

Cohen held senior roles at HP from 2006 to 2010, then became a senior director of product management at Jive Software. In 2012, he and Twig founded StreamOnce. After Jive acquired it in 2013, Cohen served as Jive’s vice president of product management until 2015. His route into founding was threaded through operating jobs, product decisions and an acquisition.

The lunch before the launch

One of the relationships that carried into the next company began over hummus. In spring 2012, Lightspeed investor Arif Janmohamed invited Cohen to lunch to discuss StreamOnce. Cohen chose Oren’s Hummus in Palo Alto. A founder may be uncertain about the market and still have firm views about lunch.

Janmohamed liked the founders as well as the idea. Lightspeed invested alongside Oren Zeev. After StreamOnce was sold, Cohen and Twig came back with a concept for making expense recording less painful. Janmohamed encouraged them to consider a broader opportunity. Two months later, they returned with a business-travel idea built around rewarding employees for saving their companies money.

That sequence gives Navan’s origins a useful texture. The vanished reservation supplied an irritation. The earlier company supplied experience. The investor conversation helped widen the question. A business can emerge from several encounters without each one needing to be a moment of revelation.

StreamOnce itself had tackled a recognizable workplace difficulty: information and conversations scattered across business systems. Its ambition was to bring them into one place. The subject changed when Cohen moved into travel, but there was continuity in the work. Someone had to deal with the gaps between tools.

Forty steps, one traveler

TripActions, the company that became Navan, began in 2015. Cohen took the CEO role and Twig the technology role. The early rewards idea addressed an awkward incentive: employees choose the trip, while their employer pays for it. Give travelers a reason to consider the budget, and a cheaper booking becomes useful to both parties.

Hotels were a particular focus. Cohen explained that the team had mapped forty steps of business travel, from learning that a flight was needed to returning home. Booking was only part of the journey. Support, changes and the traveler’s preferences also belonged in the product.

In 2019, he described a roadmap shaped both by customer feedback and by the company’s own vision. He also emphasized live human support around the clock. The app could simplify a search; people still needed help when a plan came apart. Odesa had supplied a fairly convincing demonstration of the limits of office hours.

“We always put the user first.”

Ariel Cohen, 2019

That sentence is easy to print and harder to accommodate. A traveler wants a convenient room. The finance team wants spending control. A hotel wants its inventory presented properly. Every extra rule and every extra option can add another decision. The work is in making those interests meet without appointing the traveler as the unpaid coordinator.

The receipt follows you home

Cohen’s product ambitions extended into the administrative tail of a journey. Travel produces payments, receipts and records. A flight may last an hour; its paperwork can enjoy a much longer visit.

In 2020, the company launched the expense platform originally called TripActions Liquid, later Navan Expense. It brought policy controls and receipt capture into the same wider operation as travel. The expansion also returned to the subject of that early pitch to Janmohamed: the expense process had been there from the beginning.

Geographic expansion followed. Acquisitions including Reed & Mackay in the UK, Comtravo in Germany, Resia in Sweden and Tripeur in India broadened the business. For Cohen, this meant operating across markets with their own travel habits and requirements. A product can fit on a phone while the organization behind it becomes increasingly complicated.

The company also worked on allowing customers to keep existing corporate cards. Navan Connect, introduced in 2023, linked eligible cards to its expense system. It is an unglamorous kind of flexibility, which may be precisely why it matters. A finance team can want better software without wanting to change every banking relationship at the same time.

A new name, an old irritation

On February 7, 2023, TripActions became Navan. The name joined “navigate” and “avant.” Travel, corporate cards and expense management were brought together in a single app. The business had grown beyond what its original name comfortably described.

The change coincided with generative AI becoming part of the company’s product infrastructure and its assistant, Ava. Cohen’s public discussion of enterprise software became increasingly pointed. He questioned established vendors’ pace of innovation and talked about the advantages a startup could retain as it grew.

There is a useful pressure in that argument. A founder who criticizes slow incumbents eventually has a large organization of his own to move. Growth creates meetings, dependencies and perfectly reasonable explanations for waiting. The impatience that helps launch a company has to survive the company’s success.

In a January 2024 conversation with Lightspeed’s Michael Mignano, Cohen discussed shifting a business with roughly 3,000 people toward AI. The discussion ranged from personalized bookings to expenses and the mechanics of changing direction at scale. By then, the question involved an existing operation with customers to serve while it learned new ways of working.

His more recent description of travel support combines Ava with human agents. The intention is to preserve the traveler’s context across the handoff. Anyone who has repeated a booking reference to several people will understand the appeal. The technology has to carry the conversation forward.

The bell and the next booking

Navan listed on Nasdaq on October 30, 2025, under NAVN. Cohen remained CEO and board chair, a role he had held since September 2022. The listing placed a decade of product work and company building inside the routines of a public business.

A bell is an agreeable milestone. Quarterly reporting is a recurring appointment. On September 9, 2026, Navan reported revenue of $233 million for its second quarter of fiscal 2027, up 35% from a year earlier. Gross booking volume reached $3.0 billion, up 45%. Those figures measure different things: bookings passing through the platform are much larger than the revenue the company records.

The same update described an OpenTable partnership for Navan Edge, bringing restaurant discovery and reservations into the AI travel assistant. Dining might seem a detour from an expense platform. Follow the trip itself, though, and dinner is another arrangement somebody has to make.

Cohen’s ambitions keep returning to that practical territory. Bookings, changes, payments and reservations all consume attention. Removing some of those demands gives a traveler more room for the purpose of the journey. The test happens in ordinary use, after the announcement and away from the stage.

What free time?

Asked in 2019 what he did in his free time, Cohen replied, “What free time?” He then explained that he loved travel and found mini-sabbaticals useful for thinking and reflection. It is an amusing answer from someone building a travel business: given a break from work, he still wanted to go somewhere.

That preference leaves the story with a person rather than a ticker symbol. Cohen enjoys the thing whose complications he has spent years examining. Travel can offer a chance to learn and pause. It can also leave you at a hotel trying to establish whether a reservation exists.

His company has grown far beyond that first missing room. The room remains a good test of the whole undertaking. When someone arrives, tired and preparing for tomorrow, the booking should mean something. Ideally, it should mean a key.