Fireside Chat, Interview
ElevenLabs CEO: Why Voice is the Next AI Interface
- Research initiatives targeting voice speed optimization are expected to require over three months, during which product teams may deploy interim models or extensions to bridge functionality gaps if specific research commitments miss timelines.
- Eleven Labs plans to launch a fully licensed music model to accompany existing voice agents and intends to clearly distinguish between alpha models, which may lack stability, and reliable non-alpha releases for production use.
- To address enterprise needs, the company is shifting toward a 20% engineering to 80% sales resource allocation to facilitate comprehensive orchestration of speech-to-text, LLMs, and text-to-speech, while also aiming to achieve 99% or 99.99% infrastructure reliability.
- New hires from untraditional backgrounds will be integrated with traditional teams to foster cross-learning, particularly within sales, and the organization anticipates that product teams failing to demonstrate viability or prove value within a six-month window post-product market fit will face dissolution.
- The company enforces a policy prohibiting sales to foundational model competitors and encourages sales teams to reject potentially high-commission deals that are competitive or violate policy, with incentive structures adjusted to align with long-term strategic behaviors rather than acting as lagging indicators.