Alexander Kvamme and the trouble with a full house

A restaurant can have every table booked and still have a difficult evening. Someone needs to keep track of arrivals, explain the wait and tell the next party when to come back. Software enters this scene with a promise: fewer little acts of coordination, fewer opportunities for the evening to unravel. Alexander Kvamme made an early business out of that promise. Years later, he would find another version of it inside a much larger company.

In 2011, he co-founded SeatMe, a restaurant reservation service built for the web and the iPad. Its features included texting guests when a table was ready and giving them accurate wait times. These were everyday conveniences attached to a substantial proposition: reservation management could be accessible and affordable for restaurants and nightlife businesses. An empty table is an unusually persuasive argument for better coordination. A hungry customer supplies the commentary.

Yelp announced an agreement to buy SeatMe in July 2013. Jeremy Stoppelman, Yelp’s co-founder and chief executive, saw a way to bring reservations to more of the businesses people were already discovering on Yelp. Kvamme’s stated ambition was to extend SeatMe’s reach while keeping its restaurant management system easy to use. The acquisition gave the young company access to an existing audience of diners and business owners.

The transaction closed on July 24. The consideration was approximately $2.2 million in cash and 260,901 shares of Yelp’s Class A stock, paid to SeatMe securityholders under the merger agreement. The distinction matters in any founder story: a company transaction has multiple owners, terms and moving parts. The useful biographical fact is that Kvamme had built something Yelp wanted to bring into its own business.

The next problem was inside Yelp

Working at Yelp put Kvamme close to a different coordination problem. Sales teams had information in their systems, yet managers still struggled to understand individual performance, compare it with goals and decide where coaching would help. Adding a reservations product to an existing sales operation created another layer of complexity. Having the data did not automatically make the work of managing people straightforward.

He later described the difficulty as one that persisted whether a team contained five people or thousands. Employees could miss recognition when they were doing well, or coaching when they needed help. That is an awkward condition for a business that already collects numbers about its work. The spreadsheet can be immaculate while the person represented by a row remains unsure how the week is going.

Kvamme returned to this problem with Trey Doig, the engineer who had helped build SeatMe and its successor at Yelp. Their next company, Pathlight, brought performance information, coaching and communication together for customer-facing teams. Their shared history is one of the more revealing continuities in his career. The business changed; a working partnership survived the change and became part of the next attempt.

Pathlight’s product made an employee’s own performance visible alongside the tools a manager used to discuss it. Its premise concerned the practical interval between receiving a number and knowing what to do next. A goal, an update and a coaching conversation belonged close together. Read that way, the move from restaurant reservations to team management was a change of setting for a familiar interest in making coordination less laborious.

“Craft your team.”

Alexander Kvamme, on managing people

A question worth $25 million

By April 2021, Pathlight had announced a $25 million financing led by Insight Partners, bringing its stated total raised to $35 million. Kleiner Perkins and Quiet Capital participated, alongside investors including Stoppelman, Palantir finance chief David Glazer and CAA co-founder Michael Ovitz. Stoppelman’s presence connected the new venture to Kvamme’s earlier chapter at Yelp. Quiet Capital would eventually become part of his own working life.

The company described a simple founding question: how am I doing at work, and how can I improve? It named Twilio, Earnin, Greenhouse and CLEAR among its customers. At CLEAR, more than a thousand ambassadors were using Pathlight daily, according to the financing announcement. The product was being asked to make individual progress legible across organizations large enough for a manager’s informal understanding to have limits.

Another connection became a hire. Ramon Icasiano, previously a customer at Earnin, joined as Pathlight’s first chief customer officer in January 2021. A customer moving inside the company gives the story a useful human scale. The people who understood the problem were helping shape the business built to address it. Kvamme’s growing company was accumulating relationships as well as capital.

When the conversation became the product

The next shift concerned the material being analyzed. Performance tools could show how teams were doing; customer conversations contained explanations for what was happening. By 2023, Pathlight was working with large language models to extract useful information from those exchanges. Customer service, with its accumulated calls and messages, offered a setting where language itself could become something a business could examine systematically.

Kvamme also wrote about the inconvenient parts of putting these models into enterprises. His October 2023 essay examined consistency, auditability, predictability and testing. Models that can produce different outputs from the same input create difficulties for business software. A promising demonstration still leaves engineering work to do. Companies need ways to evaluate what happened and confidence that a result can be used responsibly in the next step of a workflow.

He discussed ways of managing that variability, including combining approaches and adding rules after a model produces an answer. Each involved trade-offs. This is a revealing detail in a career often summarized through acquisitions: he was spending time on the unglamorous conditions that determine whether a product can function reliably. The quality of the machinery behind an answer matters when someone else must act on it.

In January 2024, Pathlight became Echo AI. The company, incorporated as TTSF, introduced Conversation Actions and expanded the kinds of feedback its platform could ingest. Alongside calls, emails and texts came material such as app reviews, surveys and chatbot exchanges. The scope of the product was widening from the work of a customer-facing team toward information other parts of a company could use.

Conversation Actions connected detected signals with activity in outside systems. A customer’s interest in a product could lead to a sales offer through a marketing service. Mentions of a competitor could trigger a notification when they crossed a threshold. The operational idea was to shorten the distance between hearing something useful and responding to it. Software was being given another part of the coordination job.

Calabrio announced its acquisition of Echo AI on December 11, 2024. It positioned the technology within its plans for conversation insights, quality management and agent coaching. For Kvamme, the announcement marked another acquisition of a company he had co-founded. The products were different, and so were the buyers. Both businesses had been built around making the exchanges between companies and customers easier to manage.

Before the dashboards, Room 9

Kvamme grew up in Menlo Park with three siblings and attended Phillips Brooks School. In a 2023 alumni interview, he described a connection that began on his first day there, when he was three: he met the friend who would later be the best man at his wedding. Some professional biographies take several pages to establish that someone values relationships. This one has a preschool attendance record.

He also remembered a teacher who baked a cake and pretended to sit on it, teaching the class that everyone makes mistakes. It is a pleasantly undignified lesson to carry into adulthood. Returning to the school as a parent, he looked for the same attention to each child that he remembered from his own years there. The interview identified him as both a school parent and a trustee.

His education later took him to Duke University, which his LinkedIn profile lists from 2006 to 2010. The school story supplies a different kind of chronology from the familiar sequence of companies and financing rounds: the persistence of a place, a friendship and an idea of what a supportive institution should feel like.

The books that refuse to leave

Away from company announcements, Kvamme keeps a reading list organized around what stays in his head. It is a more personal filter than a ranking of books everyone ought to buy. He remembers the honey in A Gentleman in Moscow, Enoch Root’s jungle memo in Cryptonomicon and the craft of choosing a word in John McPhee’s Draft No. 4.

Science fiction occupies plenty of shelf space, including Alastair Reynolds, Neal Stephenson and Vernor Vinge. Business and engineering histories appear too: Shoe Dog, Skunk Works and The Wright Brothers. His notes are short, sometimes only a detail. They make room for a reader who notices an image, a mechanism or a sentence and keeps it long after the surrounding plot has faded.

Another company, the same collaborator

His personal website now describes investment work with Quiet Capital and names companies including Taalas, Chroma and Fauna Robotics. It also says he and Doig are building a new company in AI inference, funded by Quiet Capital. That update, dated August 2026, places him back in the activity of making a company while he also backs other founders.

The same site links to small things he has made, including a guitar chord builder. Guitar, cooking and parenting sit among his interests. There is room for the large undertaking and the small useful object on the same page. The website introduces itself with deliberately modest expectations, calling itself sparse and infrequently updated. After years of explaining products, he appears quite content to let a personal homepage remain a homepage.

A restaurant table, a manager’s question, a customer’s message: these are modest places to begin a technology business. Kvamme’s career shows how much work can gather around them. The new inference venture opens a further chapter with a familiar collaborator. His next product will have its own particulars. The recurring question remains practical enough to ask without a pitch deck: where is the work getting stuck, and what would help it move?