ElevenLabs builds AI models that turn text into speech that sounds human. Its technology is used in customer service, with companies like Klarna and Deutsche Telekom. It also sells to creators for audiobooks, dubbing, and music.
The company is four years old and is now valued at 22 billion dollars. It reports 600 million dollars in annual recurring revenue, with most of it coming from enterprise clients.
ElevenLabs faces competition from other companies that train their own voice models. Some customers, like Decagon, use models trained on ElevenLabs' technology.
The company offers different AI models for various uses. For simple information calls, open-source models work well. For critical tasks like banking, frontier models are preferred.
Different governments use different models based on their needs. For example, Poland uses models for healthcare, with data kept locally.
The CEO believes companies should tell customers when they talk to an AI. In the future, most calls may be with AI agents. Customers may prefer this if it’s transparent.
Margins are not discussed in detail. The company focuses on proving value and may accept lower margins to grow market share.
ElevenLabs trains its models on customer data, often with human annotation. They also create custom models for specific client needs.
Why it matters
Advances in voice AI can improve service quality and reduce costs for businesses.
What to do
Monitor new voice AI models and consider how they can improve your customer interactions.
