Fireflies AI Co-Founder Admits Using Human 'Bots' to Deceive Early Clients

The Risky Gambit: How Fireflies AI's Co-Founder Used Humans, Not AI, to Charm Early Clients

In the fast-paced world of tech startups, especially those aiming for unicorn status, the line between ambitious hustle and ethical gray areas can become blurred. Fireflies, a company that recently snagged a $1 billion valuation this summer, found itself at the center of a controversy after its CTO and co-founder, Sam Udotong, openly shared a rather unconventional, and some would say audacious, tale of its early days on LinkedIn. This story, presented as a testament to entrepreneurial grit, sparked a firestorm of debate, with many questioning the ethical implications of his methods.

Udotong recounted a period where the promise of an AI-powered note-taking assistant was, in reality, fulfilled by two individuals operating on a shoestring budget. "We were charging $100 a month for an AI that was actually just two guys surviving on pizza," he admitted. This revelation came after the company's initial six attempts at a cryptocurrency-based food delivery service had faltered, creating immense pressure to make their next venture stick. The underlying message, as Udotong framed it, was about the potent motivation that stems from financial desperation, the kind that forces founders to innovate or face the stark reality of failure. He emphasized that the best way to validate a business idea is often to “be the product yourself.”

When 'AI' Meant Two Guys Taking Notes by Hand

The core of the narrative revolves around the early days of Fireflies, when Udotong and his co-founder decided to pivot to an AI meeting assistant. Their strategy for proving the concept was remarkably low-tech: they masqueraded as the AI itself. Prospective clients were told an AI would join their meetings to take notes. In reality, it was either Udotong or his co-founder, manually logging into calls as "Fred from Fireflies.ai." They would sit in silence, meticulously scribbling notes, and then deliver them within ten minutes of the meeting's conclusion. This hands-on approach, described as the "best prototype," allowed them to generate enough revenue to cover their modest $750 monthly rent in a cramped San Francisco living room. After facilitating over 100 such manual note-taking sessions, often to the point of exhaustion, they finally felt confident enough to invest in automating the entire process in 2017. Udotong stressed that since then, security, privacy, and data protection have been paramount in their development.

“The best prototype was two guys surviving on pizza. Validating before automating saved us from a seventh failure. Follow for more unconventional stories about how we built Fireflies into what it is today.”

Backlash Over Deceptive Practices

While Udotong intended to inspire with his story of perseverance, the response from the LinkedIn community was largely critical. Many users viewed the tactic not as clever validation but as outright deception and a potential breach of trust and privacy. Umar Aftab pointed out the fundamental violation: “Joining someone’s meeting uninvited is a privacy violation. They wanted to see a bot, not an uninvited human. You undermine trust and can face legal consequences.” He further elaborated on the ripple effect, suggesting that such admissions would lead clients to distrust the security and privacy of their meetings moving forward.

The sentiment was echoed by others. Toby Oyebade expressed a mix of admiration and concern regarding data privacy. Marco Mandarich bluntly questioned, “So it’s fraud?” while Dan 担C suggested, “Posting this is a quick way to erode trust, not build it.” Chris Hassard simply mused, “Not sure how lies are inspiring.”

A Deeper Dive into Ethical Concerns

Mauricio Idarraga, from GTM Engineering, provided a particularly thorough critique, breaking down the ethical and practical issues. He argued that framing deception as “validation” sets a dangerous precedent. “False representation. Clients were told an AI would join their meetings, but it was two founders secretly joining, listening, and taking notes. This isn’t ‘validation.’ It’s deception. Without consent. Without disclosure. Real people joined private meetings without participants knowing. This is a serious breach of privacy and, in some jurisdictions, even illegal.” Idarraga also raised critical questions about data handling: “What happened to those notes? Were they stored, deleted, or shared? If this is how the company started, how can clients trust its data handling practices now?” The core of his argument was that while “fake it till you make it” can sometimes work, transparently simulating backend processes is one thing; secretly eavesdropping on private conversations is another, fundamentally eroding trust.

Despite the backlash, Futurism reported that Udotong later claimed some early enterprise clients were aware that a human was involved and simply didn't mind. However, this clarification did little to quell the broader unease. The success of Fireflies is undeniable, but the controversy surrounding its unconventional, and ethically questionable, beginnings leaves a significant stain on its narrative. The question remains: can a company built on such a foundation truly earn lasting trust in an era where data privacy is paramount?