Amir Ashkenazi wanted to play the guitar. The lesson had moved online, and the screen supplied his instructor’s face quite efficiently. The rest of the lesson was another matter. There was sheet music to read, a metronome to follow and backing tracks to play. Getting these things into the same working space was awkward. A conversation could travel through the internet. A guitar lesson, with all its small moving parts, had more luggage.
For a founder whose earlier companies had worked on shopping and video advertising, this was a suitably unceremonious beginning. No grand announcement was required. A hobby had exposed a product problem. In 2020, Ashkenazi started Switchboard, a place where people could bring applications into a shared online room and actually use them together.
Four years of work would lead him to a harder question. He had built the technology. He had assembled the team. The product had even won a Webby. What if the original customer did not need it badly enough? The answer would produce Airtop, the company he now leads, and another change of direction in a career that has had several.
Before the guitar, a shopping basket
Ashkenazi’s first major venture was the comparison-shopping business that became Shopping.com. He co-founded it with Nahum Sharfman in Israel in the late 1990s and served as its chief technology officer. The company was known earlier as DealTime. It went public in October 2004, then became part of eBay in 2005.
Comparison shopping gave the web a practical assignment: let someone look across merchants rather than visit each shop separately. It addressed a mundane problem with substantial consequences. A customer wanted a product, a price and enough information to decide. Behind that simple request sat merchant catalogs, changing pages and the engineering work required to bring them together.
In 2000, Ashkenazi was working in New York, while the company had development operations in Israel. He described part of his job as maintaining communication between the two sides, translating business needs and keeping working relationships intact. The offices were separated by an ocean; his responsibilities crossed it repeatedly.
He also made his priorities plain. A Fifth Avenue office had little personal appeal compared with having the business in Netanya. The work excited him. The address was incidental. It is a useful early detail in the career of someone who would later argue that companies could hire and operate without gathering everyone in one building.
There was room for less dignified communication, too. He once recalled sending a strategy document to Yahoo Shopping’s Rob Solomon instead of his colleague Rob Goldman. Two similar names, one competitive roadmap. Autocomplete has probably never received a more generous briefing.
Learning to put a market online
Adap.tv followed in 2006. Ashkenazi was its co-founder and CEO, working on another market in which buying and selling could be made less cumbersome. This time the inventory was video advertising. Publishers had space to sell; advertisers wanted audiences. The company built technology to connect them.
In March 2011, Adap.tv announced a $20 million financing round led by Bessemer Venture Partners, with existing investors participating. The stated plans included international expansion and further product development. Its marketplace gave the company a way to serve both sides of a transaction through one platform.
By August 2013, Ashkenazi was discussing the prospect of television advertising moving toward programmatic transactions, where software would handle more of the buying and selling. His interest was in the machinery beneath the visible advertisement: how a market was organized, how a transaction happened and how much friction it carried.
AOL acquired Adap.tv that September. Its financial filing records a $405 million base purchase price, with a working-capital adjustment bringing total consideration to $410.6 million. Those are company transaction figures. They describe the scale of the business decision, rather than a personal fortune.
An IPO had also been a possibility. In a later recollection, Ashkenazi described weighing the acquisition against competitive pressure from Google and Facebook. A founder who had already helped take one company public chose a different route for the next. Experience supplied another option; it did not remove the need to choose.
A room that could keep its place
Switchboard’s proposal was easy to picture once someone explained it. A shared room could contain the tools a team already used. People could return to the room and find the work where they had left it. Preparing another call would no longer require rebuilding the same arrangement of tabs and links.
The technology underneath was a custom cloud browser. Sequoia’s early description included applications such as Notion, Figma and Linear. The ambition was to make the working material available on a persistent canvas, with participants able to engage with it instead of waiting for the person sharing a screen to drive.
The apparently simple room took time to build. In May 2022, Ashkenazi said the team had spent approximately $8 million on development before the beta and had five granted patents. The company had 22 employees at that point. These were the figures from an early-stage interview, when the product was still being tested.
Figures reported by Ashkenazi at the time of Switchboard’s beta.
The public beta opened in May 2022. In 2023, the product won the Webby Award in productivity and collaboration. Ashkenazi accepted it for the team. The customary five-word speech was neatly economical: “Nothing Great Is Accomplished Alone.” A collaboration company had managed to keep its acceptance speech collaborative.
Remote by choice, with the laundry nearby
Switchboard was Ashkenazi’s first fully remote company. He wrote that future companies he founded would be remote by choice. Hiring beyond the area around him could bring different perspectives into the product, and money that would have gone toward an office could instead support employees and team gatherings.
His account of remote work had ordinary details alongside the policy. He liked people, but could do without office politics and the mystery of an old tuna sandwich in the communal fridge. Working at home introduced its own reminder: change the laundry. The future of work still had household chores.
It also asked more of the manager. Without constant face time, he wrote, leaders had to become better at delegation and deciding which details required their involvement. The company’s arrangement therefore tested its founder as well as its software. Tools could help people gather around work; trust still needed someone to practice it.
That distinction matters in his story. Ashkenazi had been dealing with communication across locations since the Shopping.com years. Switchboard gave him a chance to build a product around the subject and to run the company under the same conditions. He was a user of the problem he was trying to solve.
The award stayed. The company changed.
Recognition did not settle the question of demand. Switchboard moved toward asynchronous collaboration as it came out of beta in 2023, emphasizing fewer meetings and more time for focused work. The product was already changing as the team learned how people wanted to use it.
Later, Ashkenazi offered a blunter account of the original premise. Applications where collaboration mattered often already supported it. In applications without that capability, people did not care enough to pay for an extra layer. The proposed solution made sense to its builders, but the need was less compelling in practice.
The cloud browsers had another possible user: AI agents. Technology developed so people could work together across applications could help software operate those applications. Switchboard’s direction changed, and Airtop emerged. Sequoia’s company description now identifies it as the former Switchboard and describes browser automation, including interaction with sites that require complex authentication.
In June 2025, Ashkenazi wrote about the decision, placing it roughly nine months earlier. He described the feeling of abandoning work the team cared about, alongside his confidence that the change had been right. Both belonged in the account. A pivot can be rational and still feel expensive.
“When you stop believing your own story, it’s time to pivot!”
Amir Ashkenazi, in his public founder AMA
Three weeks. Three months. Three years.
His published pivot framework asks founders to examine customer pull, market feedback, willingness to pay and whether the core assumptions are holding up. It also includes an uncomfortable check on sunk costs: would you begin this same business again with the knowledge you now possess?
He sets out three horizons. What could one major change achieve in three weeks? What is a realistic outcome after three more months of serious effort? Where does the current path lead after three years? The questions force the next experiment and the longer trajectory into the same conversation.
He has described pivoting each of his six companies, with results ranging from successful changes to failures that cost years and millions. That admission gives the framework some weight. It comes from someone who can point to earlier exits and still talk about being wrong.
For Airtop, a familiar engineering concern has returned: reliability. Ashkenazi has linked the problem of broken browser automations to his Shopping.com experience. He argues for using AI heavily when building an automation, then running repeatable code with less reasoning required on each execution. The goal is work that keeps working after the demonstration ends.
The shopping question that came back
Airtop has also brought him back to an unfinished ambition from Shopping.com. He has written that the earlier business helped answer where to buy, while the question of what to buy remained less complete. A product-review automation offered another attempt: collect reviews and organize the pros and cons around a buyer’s particular criteria.
There is a smaller example, too. For a company Secret Santa, he described two Airtop agents. One asked employees how they would spend a free day, week and year. Another used those answers to choose an Amazon gift and add it to a cart. An office ritual had become a browser task, with personal preferences at its starting point.
From merchant catalogs to shared rooms to gift shopping, the work keeps encountering the same practical obstacle: useful information and actions are scattered across the web. Ashkenazi keeps trying to arrange them so someone can get on with the task. Sometimes that someone is a shopper. Sometimes it is a colleague. Now it may be an agent.
The guitar lesson remains a fitting place to begin. A person wanted to do something modest, and the tools made it difficult. Ashkenazi built a company around that difficulty, then changed his answer when the business asked a different question. Keeping time is useful in music. Knowing when to change the arrangement has proved useful elsewhere.