The useful thought kept arriving five seconds late. Marc Bernstein would finish a sales call and, with the line already dead, remember the coaching point that belonged in the conversation. He had understood the advice in his manager's office. He had even agreed with it. Then a customer raised an objection, the clock sped up, and memory went on a coffee break. Regret, meanwhile, was perfectly punctual.
Bernstein answered this very ordinary human failure with an equally ordinary tool. He made an Excel macro. During a call, he could type a subject such as implementation and summon the talking points he had prepared. It was not elegant artificial intelligence. It was a small piece of scaffolding between knowing and doing.
Chris Kontes, a colleague at the St. Louis sales software startup TopOPPS, saw the macro and suggested removing the typing. What if speech recognition could hear the subject and surface the response automatically? That question became Balto, the company Bernstein, Kontes, and Davidson Girard founded in 2017. Its original insight still fits on a sticky note: advice is most valuable while there is time to use it.
“You don't need the perfect idea; you need to start.”
A founder made out of side quests
Bernstein did not grow up with founder as a fixed destination. He grew up near Washington, D.C., in Chevy Chase, Maryland, and began testing the edges of business early. He traded stocks in middle school and started a stock-trading club in high school. By senior year, another interest had taken over: physical training. He became a personal trainer and built DoorStep Fitness, a service that eventually reached more than 100 clients and employed five trainers.
A personal-training company is a compact business education. The product is inseparable from trust. Customers must be found, served, retained, scheduled, and encouraged to keep promises they made to themselves. Trainers must be hired and managed. Bernstein later said the company taught him the fundamentals of getting customers, caring for them, shaping a service, and watching the numbers. It also showed him the ceiling of a business that constantly needed to recruit people into a high-turnover profession.
He moved to St. Louis in 2011 to attend Washington University, where he studied entrepreneurship and marketing, with a minor in psychology. The combination now looks almost suspiciously tidy: how to build, how to persuade, and how people think. After college, he asked a mentor which skill mattered most for someone who wanted to start a company. The answer was sales. Founders sell to their first customers, to recruits, and to investors. Bernstein joined TopOPPS to learn the craft.
TopOPPS supplied more than a sales education. It supplied Kontes, the push to take the idea seriously, and the setting for the problem itself. Bernstein has described being content in his role until Kontes insisted that they dream bigger. The eventual founders pooled $210,000 of their savings, enough to operate without salaries, and began in a 75-square-foot coworking office in downtown St. Louis. The money lasted 18 months. By then they had a market-ready product and a small set of customers.
Eighty-two ways to be corrected
The founders did not simply declare contact centers their market. They let a long run of demonstrations narrow the field. Bernstein has said they delivered 82 demos before the first sale. He also described thinking about the problem as a matrix: where did getting a conversation right matter most? A good call could win revenue, retain a customer, resolve a complaint, or avoid a compliance failure. A poor call could do the reverse.
Contact centers concentrated those consequences. They also multiplied the coaching problem across hundreds or thousands of agents. A process change was not useful merely because management understood it. Ten thousand people had to remember it during ten thousand live conversations. Balto focused on that execution gap.
There is a durable product lesson in those demos. The first tool solved Bernstein's own recall problem. The market research identified where the same problem was expensive enough to command a budget. Personal annoyance became enterprise software only after repeated exposure to other people's constraints.
Balto's product expanded with the customer's operating cycle. The company added real-time quality assurance and coaching in 2021, Spanish-language real-time guidance in 2022, and a notetaking product in 2023. Guidance addressed the live moment. Quality review and insights helped teams understand what happened across calls. Coaching returned those lessons to agents. The Excel macro had become a loop.
Capital, correction, and the awkward middle
The financing followed in stages. Balto raised early seed capital, then a $10 million Series A in 2020. In August 2021 it announced a $37.5 million Series B led by Stripes, with RingCentral Ventures joining and existing investors participating. The company said the round brought its total funding to $52 million.
Funding gave the company room to hire and broaden the platform. It did not exempt Balto from the change in technology markets that arrived in 2022. Bernstein has spoken publicly about the operational reset the company made during that period, including reducing its workforce. The lesson he took from it was sharper focus and efficiency. The cheerful mythology of startup growth likes a clean rising line. Companies tend to draw in pencil.
Bernstein's way of discussing results reveals the operator underneath the storyteller. When a leader at Balto announces a new record, he asks whether the team knows what caused it. The answer is rarely clean. Still, the discipline matters. A win that cannot be explained is difficult to repeat. He has demonstrated Balto's own analysis tools by asking what separates successful calls from unsuccessful ones. The founder is also the test case.
“AI used to be an idea. Now it's a workforce.”
What remains human
By 2026, Bernstein's vocabulary had moved beyond assistance toward an AI workforce. Balto released a third generation of real-time guidance that he described as ambient agentic intelligence for frontline agents. The software is intended to notice when an agent needs an answer or resource, then act across connected systems to retrieve it. Balto also introduced an AI agent for routine calls such as booking a time or checking an account.
Yet Bernstein's public argument is not that human contact-center workers should quietly vanish. He has repeatedly drawn a line between repeatable tasks and work enriched by presence, relationships, judgment, or taste. In his version of the future, automation handles routine volume and people spend more time on difficult conversations, the ones a customer may remember. He has written that people are not horses waiting passively for a new machine to displace them. They vote, shape markets, and choose which companies receive their business.
That view makes his original problem more interesting. The Excel macro did not replace the salesperson. It gave the salesperson a better memory at the moment memory was needed. The technology changed, but the intended relationship remains recognizable: a person stays in the conversation while the machine carries some of the cognitive luggage.
Energy with an off switch
Bernstein's first paid job was camp counselor. He says it taught him the return on energy: enthusiasm changes the atmosphere, especially when work is chaotic. It is easy to see why that lesson stayed with a sales leader. Calls have weather. So do companies.
His preferred ways of getting unstuck are quieter. He takes long walks without headphones and allows boredom to create room for an idea. Or he rereads customer feedback, a corrective against the intoxicating abstractions of strategy. His hobby is chess, which he likes for the demand to anticipate patterns and remain calm under pressure. In a lighter experiment with momentum, he once wrote a screenplay in a single weekend merely to prove he could. His review of the script was unsparing. The useful part was starting.
The mix is revealing without requiring mystique. Energy matters, but so does silence. Pattern recognition matters, but the board answers back. A founder can carry a large vision about work and still return to a blunt customer comment for directions.
Balto accepted the 2026 CCW Cloud-Based CX Solution of the Year award after nine years in market. Bernstein credited the contact-center teams that trusted the company, offered hard feedback, and worked through releases. It was a fitting allocation of credit. The platform began with a rep who could not recall advice in time. It grew by listening to more reps, managers, and operators describe where work broke under pressure.
The founding insight was never really about clever phrasing. It was about the small distance between intention and action. Bernstein found a business inside that distance. The spreadsheet is long gone, but the five seconds are still there, waiting for someone to notice them before they become regret.