Sam Udotong revealed that he and his cofounder at Fireflies.ai, Krish Ramineni, painstakingly took notes for well over a hundred business meetings by hand while deliberately presenting themselves as an artificial intelligence assistant named Fred. This revelation, shared on LinkedIn as part of the AI startup’s founding story, immediately attracted widespread attention and curiosity because it exemplified a bold ‘fake it till you make it’ philosophy, a phrase often used to describe early-stage entrepreneurs who simulate a finished product to test their ideas in the real world.
As the company’s Chief Technology Officer, Udotong explained that Fireflies.ai—the platform that would later become synonymous with automated note-taking for virtual meetings—initially relied entirely on human labor. The very first customers in 2017, he confessed, were unknowingly interacting not with a sophisticated machine-learning model, but with real people performing every task manually behind the scenes. Udotong’s honest post spread rapidly across LinkedIn, garnering several thousand reactions and generating hundreds of comments. Some readers applauded the ingenuity and scrappy resourcefulness of the founders, seeing it as a testament to entrepreneurial tenacity. Others, however, expressed ambivalence or skepticism, questioning the ethical implications of appearing to offer an automated system that, at least in its earliest days, did not yet exist.
The story is even more striking considering Fireflies.ai’s meteoric rise. Today, the company is valued at around one billion dollars, its success fueled largely by the global shift toward digital communication and virtual collaboration during the pandemic. Yet, at the time the concept was conceived, Ramineni and Udotong were operating under severe financial constraints. As Ramineni later told Business Insider, the pair had their backs ‘against the wall.’ Facing dwindling funds and the high cost of living in San Francisco, they knew that validating their idea before writing even a single line of code was essential to survival. They set out to confirm there was genuine demand for AI-driven meeting transcription by manually providing the service themselves—even if doing so meant playing the part of an AI entity.
At that point, Ramineni was living on personal savings saved from a short period working at Microsoft, while Udotong had never held a full-time professional role. To gauge potential demand, they reached out to acquaintances in the tech community to propose a simple offer: for $100 a month, the founders would ensure that all meeting notes were accurately documented and delivered promptly. Though they disclosed that ‘some human involvement’ would occur in the process, neither explicitly stated that there was no automation whatsoever. The two then attended clients’ meetings—often back-to-back—taking meticulous notes by hand and delivering detailed summaries within twenty-four hours. Ramineni later reflected that they barely had enough revenue coming in to cover rent for Udotong’s apartment, yet they quickly realized there was enormous appetite for what they were building.
However, as human beings, they could not replicate the infinite scalability of an algorithm. After about a hundred sessions, the effort became grueling. The founders struggled with exhaustion, burnout, and the stress of managing overlapping meeting schedules. Recognizing that their manual approach could not continue indefinitely, they began focusing on developing the actual automated system that would later define Fireflies.ai’s product. Importantly, they made no approaches to institutional investors until they had completed substantial progress toward building a truly automated tool—ensuring that their temporary ‘pretend AI’ phase never overlapped with investor funding.
By late 2018, full-scale software development was well underway, supported by modest financial contributions from angel investors. The manual stage was complete, and their vision of an AI-driven note-taking platform was finally taking tangible form. A year later, Fireflies.ai began demonstrating its beta version to institutional investors, performing live product demos and testing the system in real-world scenarios. This period culminated in the company raising a seed round exceeding four million dollars, which allowed it to scale rapidly. According to Ramineni, when addressing those investors, they openly explained how their earliest users had actually received notes written directly by the founders themselves. This honest disclosure impressed potential backers, who recognized that such hands-on experimentation had allowed the founders to deeply understand user needs before committing significant resources to product development.
Entrepreneurship experts have described Fireflies.ai’s approach as a classic example of a strategy known as ‘pretotyping.’ Tim Weiss, a professor of management and entrepreneurship at Imperial College London, explained to Business Insider that pretotyping is a common but somewhat controversial early-stage tactic. Essentially, an entrepreneur pretends a product exists to observe how potential customers interact with it, thus generating insights before expensive development begins. While such a practice can be useful in testing assumptions, Weiss noted that it is generally acceptable only at the earliest stages of startup formation, before the line between simulation and misrepresentation becomes problematic.
To some observers, Fireflies.ai’s story simply brings to light what many founders do quietly: simulate automation as a way to confirm genuine user interest. Although the startup may have been vague with its earliest customers, it maintained integrity with investors by presenting only proven products to them. Kevin Werbach, a professor of legal studies and business ethics at the Wharton School, commented that within the technology ecosystem, ‘fake it till you make it’ has become an almost revered tradition. When executed responsibly, he suggested, it can resemble Steve Jobs’ legendary ‘reality distortion field’—the power to will a vision into existence by persuading others of its possibility. But when this boundary is crossed, it can veer into harmful deception, as illustrated by high-profile failures such as that of Elizabeth Holmes.
Werbach added that most entrepreneurial situations fall somewhere in between the two extremes—neither entirely harmless nor overtly fraudulent. The reference to Jobs’ ‘reality distortion field’ reminds us of his extraordinary charisma, which enabled him to make even flawed technology appear seamless. During the 2007 iPhone debut, for example, internal reports later revealed that the demo units were prone to crashes unless handled in a very specific sequence, and even the display of cellular signal bars was hard-coded for the presentation. Yet this act of managed illusion was ultimately in service of a product that did transform the industry.
The practice of augmenting human assistance to simulate AI performance is not isolated to Fireflies.ai. Facebook’s personal assistant project, known as ‘M,’ operated in a similar way before it was discontinued in 2018. Many of its complex responses were actually written by humans who trained the system while masquerading behind an automated interface. The project never progressed beyond private testing phases, but it serves to illustrate a broader theme—the blurred line between early human involvement and eventual automation in AI development.
Today, Fireflies.ai has transcended its experimental beginnings. According to Ramineni, its platform has processed over two billion minutes of meeting audio and produced summaries for twenty million individual users—a staggering accomplishment equivalent to roughly four thousand years of continuous work if performed by humans on an eight-hour schedule. Reflecting on this transformation, Ramineni acknowledged that healthy skepticism toward AI remains important. He emphasized that while resourceful improvisation has its place in innovation, transparency becomes essential once external funding and commercialization are involved. ‘You can’t fake it till you make it forever,’ he said. ‘That’s not how it works.’ Through perseverance, humility, and disciplined validation, Fireflies.ai ultimately transformed its early improvisation into genuine technological achievement, turning a small act of creative survival into a billion-dollar success story.
Sourse: https://www.businessinsider.com/startup-story-fake-it-till-you-make-it-fireflies-ai-2025-11
