The desk was empty. This is not generally how legends announce themselves. There was no embossed welcome packet, no tasteful succulent, no laptop waiting at a flattering angle. Joel Grossman had talked his way into a summer at Location Labs while earning his MBA, even offering to work for free. Then he arrived. “I showed up and found an empty desk, and that was the beginning.”
It was a useful beginning because it contained the whole assignment: discover what matters, make yourself valuable and do not wait for the furniture to explain the job. Grossman stayed for roughly a decade. He moved from product into marketing, sales, business development, people operations and finally the chief operating officer’s seat. Location Labs grew from a small venture-backed company, still searching for a durable business, into a mobile-safety provider with more than 200 employees. AVG bought it in 2014 for a reported $220 million.
The acquisition is the shiny number. The stickier fact is 95 percent, Location Labs’ reported employee-retention rate. Grossman became a student of why people joined, why they stayed and how a company could grow without turning itself into a handsome brochure about a workplace that did not exist. His answers were strikingly unornamental: tell candidates the truth, look for rate of growth instead of polish, invite disagreement before a decision, and notice good work before the person doing it has to ask.
Before the phone learned to answer
Grossman’s first chapter was at Microsoft, where he worked from 1999 to 2004 as a lead program manager. His projects included Windows XP, MSN Explorer and SPOT, the company’s early attempt to put live information on connected objects. A SPOT watch could receive news, weather, messages and calendar updates over FM-radio subcarriers. It was the sort of future that arrived early, wearing a large wristband.
The important thing was not whether the watch conquered the world. It placed Grossman near a problem that would follow him: a communication system is only as useful as its handoff. When MSN Direct suffered an outage in 2004, he appeared in the product’s message board to say the operations team was investigating. Two decades before AI agents began politely picking up dealership calls, he was already standing where software, networks and impatient humans collide.
He left Microsoft for Harvard Business School, graduating with distinction in 2006. A classmate introduced him to Tasso Roumeliotis, the founder of what was then a lean mobile-location startup. Grossman wanted a Bay Area technology internship and, more particularly, a business model he could believe in. The empty desk followed.
A company that knew what it was
At Location Labs, Grossman helped build a culture that treated candor as operating infrastructure. Job listings were expected to describe the place honestly, including the absence of elaborate training wheels. Interviews tested company values. New interviewers trained beside experienced ones. Feedback was calibrated. The point was not to make every candidate want the job; it was to let the right ones recognize it.
He liked the “diamond in the rough,” but his metaphor was mathematical rather than sentimental. Put time on one axis and skill on the other. The attractive candidate was the steep curve. One person he cited had been a financial adviser, taught himself to code, entered Location Labs in quality assurance and grew into an engineering leader. More than 30 percent of the company’s 2015 hires had begun their careers in a different field.
This was also a theory of Grossman’s own career. A computer-science graduate became a product manager, then an operator responsible for functions well beyond product. His authority grew because the job grew, and because he kept giving parts of it away. Location Labs organized work around “missions.” Leaders pitched the problems; employees chose where to invest themselves. Code acquired a purpose beyond being code.
For teamwork, the company gave out a Spartan Shield award. A shield in a phalanx protected the person beside its carrier, so the honor went to an employee who stepped outside a narrow job description to help another team. The imagery was a little grand for office life, which is precisely why it worked. Most trophies celebrate the person holding them. This one celebrated the colleague they had covered.
The 95 percent retention figure did not come from kombucha or slogans. Grossman tied happiness to productive agency. Teams defined goals and presented their progress in miniature board meetings. Leaders surveyed employees for anonymous town-hall questions. Management admitted when internal communication had failed. Raises and promotions could arrive before the annual calendar gave permission. Treating individuals as individuals is harder to forecast in a spreadsheet. It is easier to remember as a person.
The second act narrows its aim
After Location Labs, Grossman, Roumeliotis, Andy Ruff and Steven Ginn assembled again. In late 2016 they formed NumberAI, soon known by its product name, Numa. The first idea was broad: turn the ordinary business phone number into an AI receptionist for millions of small businesses. A caller or texter could reach the same familiar number, while software routed the conversation and automated simple work.
The team raised a $1.6 million seed round in 2017. It also learned that “small business” was less a market than a continent. A bar, a salon and a repair shop could share a phone system while sharing almost nothing about how work moved after the call. Numa killed the horizontal version and concentrated on automotive dealerships, where the volume was heavy, the software stack was fragmented and a missed response had an invoice attached to it.
A dealership is a theater of divided attention. Service advisers stand with customers, check parts, talk to technicians and watch phones flash. Sales and service use different systems. A voicemail may be sincere, urgent and commercially valuable, then disappear into the small administrative Bermuda Triangle between departments. Numa’s software answers and texts, books appointments, surfaces customer sentiment and keeps conversations visible to staff.
The focus produced measurable momentum. Numa ranked No. 168 on the 2024 Inc. 5000 after reporting 2,248 percent revenue growth over three years. In October 2024 it announced a $32 million Series B led by Touring Capital and Mitsui, bringing the company’s stated total funding to $48 million. At the time it served about 600 dealerships in the United States and Canada. By April 2026, after acquiring the AI sales platform Ficus, Numa said its footprint had passed 1,300 rooftops and its revenue had grown 200 percent.
The expansion continued in August 2026 with the acquisition of SocketTime, a fixed-operations platform, and FordDirect’s selection of Numa as a strategic engagement AI vendor for Ford and Lincoln retailers. The company that once tried to make a phone number smarter now describes a larger ambition: connect sales, service and retention without asking a dealership to discard every system it already uses.
The durable advantage is a memory
Grossman is not the loudest figure in Numa’s public story. Roumeliotis is its chief executive and most frequent quoted voice. That distinction suits the pattern. Grossman’s work has repeatedly concerned the connective material of a company: how product meets sales, how a candidate meets a culture, how a caller reaches the person who can help, how colleagues protect one another’s flank.
His board service at The Seven Hills School supplies another quiet continuity. The school describes him as a technology entrepreneur and parent; he and his wife, Holly, have three children connected to it. Operating, in this view, is not merely arranging resources. It is tending to institutions where people are meant to grow.
There is a fashionable story about serial founders: they exit, rest briefly and return with sharper instincts and a larger address book. Grossman’s version is less cinematic and more revealing. Several people came back together. They tried something wide. They admitted it was too wide. They found a dense, unfashionable operational problem and narrowed until the product fit the work.
The empty desk has become a useful emblem. It offered neither status nor instruction, only a place from which to begin. Grossman filled it, then spent years building the conditions in which other people could fill their own. Numa’s AI may answer in seconds. The company behind it was built on a much slower technology: colleagues learning, disagreeing, remembering and deciding to begin together once more.