Fireside Chat, Interview
ElevenLabs CEO: Why Voice is the Next AI Interface
Eleven Labs Company Metrics & Strategic Pivots
- The Voice Marketplace currently hosts nearly 10,000 distinct voices created by the community.
- Eleven Labs has distributed $10 million in earnings directly to community voice creators.
- The company has expanded its audio infrastructure to include text-to-speech, speech-to-text, sound effects, music generation, and a conversational agent platform.
- The organizational structure consists of approximately 20 autonomous product teams, each comprising 5 to 10 members.
- The company maintains a flat hierarchy with no formal titles, allowing lateral movement and rapid promotion based on impact rather than tenure.
- The engineering team grew from a small founding group to approximately 11 people, while the total company headcount reached 350.
Research vs. Product Execution Strategy
- Eleven Labs rejected early user requests for manual "speed sliders" to force a research-led solution where the AI dynamically adjusts pacing based on voice context.
- The internal rule dictates that if a research solution is predicted to take longer than three months, product teams are permitted to ship interim extensions or alternative models to meet deadlines.
- The company utilizes a "pre-product market fit" model for new initiatives, granting teams six months to prove viability before the product is killed or scaled.
- Enterprise customers receive "alpha" access to unstable features with clear delineation, while post-product market fit features undergo rigorous evaluation before deployment.
- Security and compliance goals target "four nines" (99.99%) to "five nines" (99.999%) availability for enterprise infrastructure.
Global Talent Acquisition & Culture
- The company operates globally with hubs in London, Warsaw, San Francisco, and New York, rather than adhering to a single headquarters model.
- Early hiring prioritized candidates with deep open-source contributions and untraditional backgrounds, including a researcher previously working in a call center.
- The leadership team actively recruits from Europe and Asia to access specialized talent pools that may be overlooked in traditional US-centric hiring.
- New hires are often placed in physical hubs to immerse them in company culture, while remote work remains an option for experienced staff.
- The company removed executive titles to eliminate internal hierarchy barriers, though leadership still exists functionally within subdivisions like research, sales, and operations.
Industry Partnerships & Legal Navigation
- Eleven Labs spent 18 months negotiating licensing agreements with major music labels (Merlin, Cobalt, and others) to legally incorporate music into their generation models.
- The music partnership strategy focused on "forcing functions" and clear timelines to overcome industry hesitation and secure commercial rights for users.
- Legal hiring shifted from risk-averse Fortune 500 counsel to partners with startup experience who can balance risk mitigation with commercial enablement.
- The company actively discourages partnerships with foundational model competitors, restricting sales to those entities despite potential immediate commission revenue.
- Early adoption in the voice acting sector was driven by the "Voice Marketplace," which incentivized creators to generate and share diverse voice assets for a revenue share.
Enterprise Transition & Incentive Structures
- The sales organization transitioned from an 80/20 split favoring engineering-led sales to a 20/80 split favoring traditional sales operations.
- The company learned that "passion-based" motivation is insufficient for large-scale growth, necessitating a shift to structured commission and quota systems.
- Sales incentives were explicitly aligned with long-term strategy, including policies that penalize selling to direct competitors even if it yields higher immediate commissions.
- Enterprise adoption cases, such as Hippocratic in healthcare, required orchestrating multiple AI layers (speech-to-text, LLMs, text-to-speech) rather than relying on single-model outputs.
- The company addresses the "demo-to-production" gap by building specialized tooling for version control, evaluation, and fine-tuning specific to enterprise deployment needs.
Notable Use Cases & Outcomes
- A deep Spanish voice created for Eleven Labs initially failed in the Spanish market but became a top-performing free voice when adopted by English-speaking users, highlighting the model's cross-lingual flexibility.
- Deutsche Telekom utilized early voice models to create notebook-style podcast experiences with localized German and English voices.
- Hippocratic deployed voice agents to handle hospital inbound calls, scheduling, and outbound patient medication reminders, integrating telephony and LLMs.
- The company's music model now provides fully licensed commercial rights for generated tracks, a critical differentiator for enterprise clients concerned with copyright.