Fireflies.ai: The Startup That Faked Its AI Until It Became a Real $1 Billion One

Key Facts

• Who founded Fireflies.ai: Krish Ramineni (CEO) and Sam Udotong (CTO), starting around 2016, both eventually recognized on Forbes' 30 Under 30 list.
• What was the "fake it till you make it" story: Before building any real AI, the founders manually dialed into over 100 client meetings pretending to be an AI notetaker named "Fred," silently taking handwritten notes and emailing them afterward.
• Who uses Fireflies.ai: Over 500,000 organizations and 16-20 million people, including Uber, Nike, Netflix, Delta, and Expedia, with usage across 75% of Fortune 500 companies.

Before There Was Any AI, There Were Just Two Guys Named Fred

Fireflies.ai's origin story, revealed publicly nearly a decade after the company's founding, is among the most candid and unusual founder confessions documented anywhere in this CrackTheDeck series. Co-founder Sam Udotong disclosed in a widely shared LinkedIn post that he and CEO Krish Ramineni manually took notes for over 100 client meetings while pretending to be an AI bot, deliberately deceiving their earliest customers about how the product actually worked. Business Insider's coverage of the viral revelation quotes Udotong directly describing the mechanics: "When customers scheduled a meeting, we'd manually dial in as 'Fred from Fireflies.ai.' We'd sit there silently, take detailed notes, and send them 10 minutes later".

Hindustan Times' coverage of the same story adds the founders' own characterization of what they were actually doing at the time — describing themselves bluntly as "two 'broke guys' pretending to be AI". Ramineni's own account clarifies the specific nature of the deception to prospective customers: he told those early clients "there would be some human involvement and oversight, but did not specify that the process was actually entirely manual" — a deliberately ambiguous framing that let customers believe they were testing a genuinely automated product, when in practice every single note was being generated by hand.

A Deliberate Strategy, Not Just Early-Stage Desperation

That "fake it till you make it" period reflected genuine financial necessity as much as strategic testing. Hindustan Times' reporting is explicit on this point: "when Udotong and Ramineni settled on the idea of an AI notetaker, they did not have the funds to actually deploy AI" — meaning the manual approach wasn't simply a clever validation experiment layered on top of an already-functioning product, but literally the only version of the product the founders could afford to offer at that early stage. That constraint forced the founders to directly test the core underlying hypothesis — that people would genuinely value and pay for automated meeting notes — before investing scarce resources into building the actual automation technology to deliver that value at scale.

The company's own founding narrative, as recounted by AssemblyAI's interview with Ramineni, frames the underlying problem the founders identified as "rapidly becoming universally felt: ensuring that attendees were getting maximum value out of virtual meetings and automating the cumbersome task of notetaking". Krish's own stated founding philosophy, documented in a separate profile of his journey, reinforces that problem-first approach directly: "Rather than starting with a technology and attempting to find a problem for it, he advocates for identifying a substantial, real-world problem first," warning explicitly against "artificially creating a problem that doesn't naturally exist" — a philosophy the manual "Fred" period embodied almost literally, testing real customer demand for the solution years before the underlying AI technology to deliver it at scale had matured enough to be practical.

Timing: Riding the Explosion of Speech Recognition and LLMs

Fireflies.ai's eventual transition from manual notetaking to genuine AI automation was substantially enabled by broader technological progress well outside the company's own control. AssemblyAI's profile of the company notes directly that "the recent explosion of technological advancements around automatic speech recognition, LLMs, and other frontier models is what ultimately enabled Fireflies.ai to accelerate its growth and become a dominant market leader in the AI meeting assistant space" — a pattern echoed throughout this CrackTheDeck series, where companies founded years before generative AI's mainstream breakout found their original vision suddenly, dramatically more achievable once the underlying model capabilities finally caught up to what their founders had originally imagined.

Business Insider's coverage adds a further specific accelerant to Fireflies' growth: the company's current $1 billion valuation is "largely thanks to the rise of virtual meetings during the pandemic" — the sudden, forced global shift to remote work and video conferencing in 2020 created exactly the kind of massive behavioral change and pain point (endless virtual meetings, difficult to track and follow up on effectively) that Fireflies' product was uniquely positioned to address.

Reaching Unicorn Status Without a Mega Funding Round

Fireflies.ai's June 2025 unicorn milestone stands out sharply against the capital-intensive fundraising patterns dominating most of this report. TechFundingNews' coverage is explicit that the $1 billion valuation came "following its first tender offer" rather than a traditional new funding round — a tender offer being a transaction structure where existing shareholders, typically long-tenured early employees, sell some of their shares to new or existing investors, providing them liquidity without the company itself raising fresh primary capital. Halisi Consults' analysis frames the significance of that structure directly: "In June 2025, Fireflies reached unicorn status not by raising another funding round, but through a secondary sale that let early employees cash out".

A LinkedIn analysis of the milestone captures just how unusual this capital efficiency is relative to typical AI startup fundraising patterns, with the headline framing reading: "Fireflies.ai: $1B Valuation with $19M Funding, Quietly" — directly contrasting the company's approach against founders elsewhere in the industry who "think: 'AI = higher valuation'" and raise "$200M and still don't" achieve comparable outcomes. Independent tracking from Latka confirms the underlying figures precisely: Fireflies.ai has raised just $19.1 million in total funding across three rounds, most recently a $14 million Series A led by Khosla Ventures in 2021 — meaning the company achieved a thousandfold-plus return on invested capital at its $1 billion valuation, an efficiency ratio almost unmatched among the other companies profiled throughout this series, many of which have raised hundreds of millions or billions of dollars to reach comparable valuations.

From "AI Notetaker" to "AI Teammate"

Fireflies' own June 2025 announcement of its unicorn milestone coincided deliberately with a significant product repositioning. The company's blog post describes itself as "the #1 AI teammate for meetings used by people at 75% of Fortune 500 companies" — a deliberate shift in framing from a passive "notetaker," simply transcribing and summarizing what happens in a meeting, to an active "teammate," participating meaningfully in the conversation itself. That shift was embodied concretely in the simultaneous launch of "Talk to Fireflies," a feature enabling voice-activated interaction with the AI during live meetings, triggered by saying "Hey Fireflies" or typing "/ff" in the meeting chat, available across Zoom, Google Meet, and Microsoft Teams in over 60 languages.

Notably, that feature was built through "a first-of-its-kind partnership" integrating Perplexity's web search capabilities directly into the meeting experience, letting users "ask questions and get real-time answers sourced from the internet—all without leaving their conversations" — a detail that connects Fireflies' own growth story directly to the broader AI infrastructure ecosystem this CrackTheDeck series has profiled throughout, illustrating how even a company built around a relatively narrow, well-defined use case like meeting transcription increasingly finds it valuable to integrate capabilities, like real-time web search, developed by other specialized AI companies rather than building every underlying capability itself.

Scale: 16 Million People, 500,000 Organizations, and More Audio Than Spotify

Fireflies.ai's usage metrics by 2026 reflect genuinely massive scale for a company that raised so little outside capital. AssemblyAI's interview with Ramineni states that "since its founding in 2016, the AI notetaker platform Fireflies.ai has taken notes for over 16 million people and 300,000 organizations," adding the striking comparative detail that "today, more audio is created on the Fireflies.ai platform than on Spotify each day". Startup Intros' independent profile of the company puts the organizational figure even higher, describing usage across "over 100,000 organizations globally" with named corporate clients "including Uber, Nike, and Netflix," and total reach "used across 500,000+ companies globally". Ramineni's own most recent public figures, shared in late 2025, described the platform's AI having processed "over 2 billion meeting minutes" for 20 million people — a volume Ramineni calculated as equivalent to roughly 4,000 years of continuous meetings if measured against a standard eight-hour workday.

Growing Competition in an Increasingly Crowded Category

Fireflies' continued success has not gone unchallenged. Ramineni's own January 2026 public commentary captures the intensifying competitive pressure directly: "Every time I breathe, there's a new AI notetaker... When we started Fireflies, there were maybe two of us in the entire category" — a candid acknowledgment that the AI meeting assistant category, once a genuinely novel and differentiated niche when Fireflies first pioneered it, has since become one of the more heavily contested segments of the broader AI application landscape, with numerous well-funded competitors having since entered the space.

A Fitting Final Chapter for This Series: Radical Capital Efficiency in an Era of Mega-Rounds

Fireflies.ai's story offers a genuinely distinctive counterpoint to close out this batch of the CrackTheDeck AI Unicorns series. Where companies like Fireworks AI, Baseten, and Modal reached comparable or greater valuations through hundreds of millions or billions of dollars in venture capital, and companies like Dream Security and Zhiyuan Robotics moved at similarly capital-intensive paces, Fireflies built a billion-dollar company on just $19 million — starting, quite literally, with its two founders manually pretending to be the very AI product they would eventually spend years building for real. That trajectory serves as a useful reminder that even in an industry increasingly defined by ever-larger funding rounds and ever-faster valuation escalation, durable product-market fit, patiently earned over nearly a decade of iteration, can still produce genuinely extraordinary outcomes without requiring the same scale of outside capital that has become the default expectation across so much of the rest of the AI unicorn landscape.

FAQ

Source: CrackTheDeck Research.