Sam Liang, co-founder and CEO of Otter.ai, has spent a decade building a company around a core frustration: most human voice knowledge throughout history has been lost because there was no way to capture it at scale. Shakespeare and Darwin never left voice notes. Otter.ai started in 2016 as a transcription tool, evolved into an AI meeting assistant, and is now building a meeting-centric enterprise knowledge base. The company has over 35 million users and exceeded $100 million in ARR.
The Original Bet That Made People Uncomfortable
In 2016, Sam’s vision was to record every meeting and make the contents shareable across teams. Both ideas were deeply countercultural at the time. Being recorded felt invasive, and meeting notes were traditionally private, handwritten, and personal. Sam anticipated that culture and mindset would shift over time, and that early adopters would demonstrate enough value to pull others in. He deliberately chose a problem most people weren’t yet convinced by, believing that’s where the biggest opportunity lies.
Why Otter Built Its Own Speech Recognition Instead of Using APIs
From day one, Sam insisted on building proprietary speech recognition technology rather than relying on third-party APIs. His reasoning was strategic and economic. Third-party APIs would have capped how much free service Otter could offer and created a cost structure that limited scale. More importantly, owning the technology meant Otter could tackle problems no API vendor was solving, like modeling multi-speaker conversation dynamics and leveraging hundreds of millions of voice data points to understand how people actually talk in meetings. Sam’s view is that if a product is easy to build, it’s easy for a hundred competitors to replicate, so deep technology is the only durable differentiator.
Voice as the Next Primary Interface for Enterprise Intelligence
Sam believes voice is replacing text as the primary interface for business communication. He points to the historical shift from email to Slack as precedent: when a new tool becomes good enough, behavior changes permanently. He argues that talking is fundamentally easier than writing, and AI is already making it practical to dictate documents, emails, and posts. Within a few years, he expects keyboard use to decline sharply as AI handles the writing. Otter’s long-term bet is that voice becomes the central layer for enterprise intelligence, with AI capturing, organizing, and making searchable the knowledge that flows through meetings.
The Scale of the Remaining Opportunity
Despite 35 million users and $100 million in ARR, Sam estimates that 95% or more of the world has not yet adopted a tool like Otter. He frames the company’s trajectory over a 10-year horizon, not quarter to quarter, and compares building a startup to running a marathon, except harder. He has personally run 11 marathons and credits that discipline with helping him manage stress and persist through challenges. His core advice: most people give up too fast, and if you’re building something genuinely difficult, the hardships are expected, not a signal to stop.
