The financial analysts had a listening problem. Quarterly earnings calls kept arriving, and someone had to hear them, sort them and decide what mattered. Jeff Gallino, who already knew something about speech recognition, heard their complaint at a conference. Most business opportunities arrive with less useful dialogue.
He put two capabilities together: software that could recognize speech, and analytics that could examine what had been said across many recordings. In 2002, he founded CallMiner. A spoken conversation could become something a business could search, compare and learn from, without assigning a human being to replay every minute.
That small shift in perspective has occupied a considerable portion of his working life. Gallino is now CallMiner’s CEO and founder, having also spent years as its chief technology officer. The job has changed; the underlying question has remained wonderfully awkward. If your customers are already telling you what is wrong, why do you keep asking everyone else?
The complaint that became a company
The early opportunity was larger than earnings calls. Contact centers accumulated conversations by the thousands, but listening to recordings was laborious. Gallino’s wager was that these everyday exchanges contained information worth recovering. The customer asking for help was also describing the product, the process and the promises that had brought them to the phone.
By 2018, he was describing a familiar industry practice: reviewing roughly 1–2% of an agent’s monthly calls. CallMiner’s approach aimed to analyze every interaction. He also described the engineering challenge of processing more than 100,000 calls simultaneously in real time. The distance between a working demonstration and a working operation was measured in rather more than applause.
His attention went beyond recognizing words. Silence could reveal wasted time or a difficult handoff. Patterns could guide coaching. Feedback based on a broader set of interactions could give an agent something more useful than a verdict drawn from a few unlucky recordings.
Eighteen moves before graduation
Gallino’s own childhood offered little in the way of a fixed address. Raised by a single mother, he moved 18 times before finishing high school. He has described becoming the first person in his family to graduate from high school and then college. Those are milestones with a different weight from a product launch.
He initially wanted to be a scientist. Computer science turned out to suit his abilities, and he went to the United States Air Force Academy. His official biography lists a bachelor’s degree in computer science and 11 years in the Air Force. Work at Grant Thornton Consulting and ThinkEngine Networks followed before CallMiner.
When discussing consequential choices, he uses the phrase “enlightened self-interest.” His explanation includes thinking about the effects of a decision on other people. It is an unusually useful qualification for a founder: your own future has a habit of becoming entangled with everyone on the payroll.
The founder who changed chairs
CallMiner’s management history contains a useful interruption to the usual founder biography. Gallino ran the business during its first five years. In October 2007, the company announced Terry Leahy as president and CEO, with Gallino becoming chairman and CTO. The stated purpose was to give the founder more time to concentrate on the technology.
The arrangement put company building and product building in different hands. Leahy arrived with experience leading technology businesses; Gallino could focus on developing the software and extending its uses. An organization can ask a great deal of its founder. It can also benefit when the founder stops attempting every answer personally.
By November 2023, Gallino was again being identified in company announcements as CEO. That gives his current title some history: he has occupied both the executive chair and the technical one, and watched other people run the business he started.
“I’m giving up a title to learn.”
Jeff Gallino, on letting someone else lead, in his conversation with Kyle Carter
In a later conversation with Kyle Carter, the lesson appeared in especially plain language: giving up a title could be a way to learn. Carter’s account also highlighted the distinction between polite praise and customers who ask questions. For someone selling software to analyze conversations, paying attention to the quality of a response seems occupationally appropriate.
An idea must survive the notebook
Gallino keeps a notebook for business ideas. Every few months, he returns to the entries and grades them. Ideas that continue to interest him earn a further hearing with people in his network. It is a modest ritual for an industry that often prefers its inspiration accompanied by stage lighting.
An investor helped establish the distinction. After CallMiner’s launch, Gallino discussed other possibilities with him and learned that the investor encountered about six good ideas a day. Execution was the scarce commodity. In Gallino’s telling, a promising thought still has to become a product someone can buy.
Competition supplied another correction. He had researched the market for months without finding a business doing what he intended. Within two months of founding CallMiner, two competitors had appeared. He treated their arrival as encouragement. A second supplier can be irritating; it can also suggest that the first supplier has identified an actual market.
That attitude makes his notebook more interesting than a collection of inventions. It gives an idea time to lose its initial glamour before it receives a company, a team and a bill for everything.
A listening company needs a room
There is a geographical chapter to CallMiner that a Massachusetts business address can obscure. Gallino founded it in Lee County, Florida. In 2016, the company announced a move into a 13,000-square-foot Fort Myers office, with a training center for customers and partners. Software development, research, services and marketing worked from that office.
The announcement also described a relationship with Florida Gulf Coast University. CallMiner had hired four engineering graduates, and students were working on a project intended to improve processes for business analysts. Engineers advised on the curriculum. The connection put a practical problem in front of students and brought local graduates into the business.
Gallino described his original ambitions as including a good place to work and bringing technology jobs to Southwest Florida. It is a concrete counterpart to the product story: a company needs somewhere for people to learn the craft of building it.
By 2019, its annual LISTEN conference drew 413 attendees to Hollywood, Florida. Gallino spoke in the general sessions alongside product and AI colleagues. The program included analyst workshops, industry exchanges and an escape-room challenge. Even a business devoted to examining conversations apparently understood that people occasionally benefit from being given something else to do.
Who gets heard on the next call?
Gallino’s public discussions have also looked at the person wearing the headset. In a 2020 conversation with customer-experience writer Adrian Swinscoe and CallMiner engineer Conrad Liburd, the subject was bias in customer interactions and the responsibilities of employers when callers behave abusively.
The discussion widened the usual quality-assurance question. Reviewing a call can mean asking how well an employee followed a script. It can also mean examining the behavior they were expected to tolerate, and whether the organization gave them authority to respond. A recording contains two sides of the exchange, even when the scorecard concentrates on one.
The conversation argued for empowering frontline employees and putting automation behind them to make them more responsive. That is a revealing emphasis for a software founder. Better information has value when someone can act on it. Otherwise, the person answering the phone merely becomes better informed about a problem they still cannot fix.
When the listener begins to answer
Under Gallino’s leadership, CallMiner has moved further into taking action. Its AI Assist updates in May 2025 described research and supervisor agents that divide analytical tasks into workflows. The intended user could ask questions in ordinary language and investigate customer patterns with less manual work.
Then came a more visible change. Following its acquisition of VOCALLS, CallMiner introduced OmniAgent in August 2025. The product extended its work into virtual agents handling conversations across voice, chat and email. A company built around interpreting exchanges was now supplying software to conduct them.
The connection to its earlier work was deliberate: use the information in existing interactions to decide where automation belongs. In November 2025, customer FREESTYLE selected OmniAgent for outbound collections, with plans to expand into inbound collections and other uses. Those were announced intentions, rather than a completed scorecard of results.
For Gallino, the change brings the original listening question into the design of the next conversation. An automated voice still needs to understand the reason someone called. Giving it a pleasant manner is unlikely to compensate for sending the caller back to the beginning.
The work after the demonstration
The newer partnerships show how much follows the software. In May 2026, CallMiner and Etech announced a combination of conversation intelligence and professional services. Certified analysts would augment implementation and optimization work; the partnership also anticipated virtual-agent offerings through OmniAgent.
In June, a partnership with PossibleNOW addressed contact compliance inside automated interactions. Its stated purpose included determining who could be contacted, when and through which channels. Gallino argued that automated interactions should face the same compliance standards as human ones. The ambition was to incorporate those requirements into the operation itself.
By October 14, 2026, the company was announcing expanded AI Assist access, a redesigned RealTime Command Post for supervisors and additional training resources. The emphasis included helping users validate findings through counts, comparisons and references to the underlying conversations. The latest chapter still involved making information usable by the person responsible for the next decision.
These are workaday concerns compared with an AI demonstration, which is precisely what makes them revealing. Deployment, supervision and operating rules occupy the space between a clever answer and a business that can rely on it. Gallino has been living in that space since analysts complained about their recordings.
His career now circles back to a deceptively simple habit: hear the complaint carefully before deciding what to build. The earnings calls, the agent’s recordings, the notebook and the questions from prospective customers all ask for attention before action. After more than two decades at CallMiner, he is still finding work in the things people have already said.