Two founders of Fireflies.ai, Sam Udotong and Krish Ramineni, revealed that they initially pretended their AI-powered meeting note-taking product was fully automated when it was actually run by humans. The story, shared on LinkedIn by Udotong, the company’s CTO, quickly went viral, sparking debate over the ethics of the “fake it till you make it” approach in startup culture. In 2017, with little money and no product, the two co-founders created a fictional AI bot named Fred. They manually attended over 100 virtual meetings, took handwritten notes, and delivered them to early customers within a day—while claiming the work was being done by an AI. Ramineni, the CEO, said they were desperate to validate their idea before writing a single line of code. “We had just enough money to pay for the rent where Sam was staying, and we found incredible demand,” he told Business Insider. The tactic worked. They charged $100 per month and quickly built a user base, even though the process was exhausting and required them to juggle back-to-back meetings. The human team could not be everywhere at once, and the stress of being double-booked eventually led them to develop a real AI solution. They did not seek investment until they had a working product. By late 2018, they had stopped manual note-taking and were focused on building the automated system. In 2019, they conducted live demos for institutional investors, which led to a $4 million seed round. Ramineni said investors were impressed not by the idea alone, but by the founders’ method of validating the problem first through real user interaction. Experts have mixed views on the approach. Tim Weiss, a professor at Imperial College London, described it as “pretotyping”—a common early-stage strategy to test demand before building. He noted it’s acceptable in the beginning but becomes risky if continued past the validation phase. Kevin Werbach, a professor at the Wharton School, called the tactic a “hallowed element” of startup culture. He compared it to Steve Jobs’ “reality distortion field”—the ability to convince others that impossible goals are achievable. Jobs famously used carefully scripted demos to hide technical flaws in early iPhones. Similarly, Facebook’s AI assistant M relied on human operators behind the scenes to handle complex tasks, though it never launched publicly. Ramineni acknowledged the ethical gray area. “It’s fair to have skepticism around AI,” he said. “You have to be transparent, especially when raising funds or building the product. You can’t fake it till you make it forever.” Today, Fireflies.ai is valued at $1 billion and has processed over 2 billion meeting minutes for more than 20 million users—equivalent to roughly 4,000 years of meetings. The founders say their early deception was a necessary step to prove demand, but one they wouldn’t repeat.