A new audio-focused artificial intelligence (AI) startup, WaveForms, has launched with the goal of advancing how AI understands and processes spoken language. Founded by former OpenAI engineer Alexis Conneau and ex-Google product strategist Coralie Lemaitre, the startup aims to capture nuances such as intonation and emotion, areas where many existing systems fall short.
Why It Matters
Many AI-powered chatbots rely on speech-to-text technology that can lose the subtle cues embedded in tone and inflection. WaveForms seeks to solve this challenge by creating a system that passes the “Speech Turing Test,” where users cannot distinguish between human and machine communication. “We want AI to grasp not just words but the emotional context behind them,” Conneau said.
WaveForms raised $40 million in a seed round led by Andreessen Horowitz, despite having only five employees currently.
The startup is still developing its models and acknowledges potential risks, including the possibility of users forming attachments to hyper-realistic AI personalities. “I want to believe that we’re more prepared for these challenges now,” Conneau said, reflecting on lessons learned from social media’s pitfalls.
Looking Ahead
While specific products remain under wraps, WaveForms plans to focus initially on consumer applications, with future business-to-business ventures likely. Potential use cases include education, where emotionally aware AI could transform how users interact with learning tools.
“Ultimately, this technology can touch a wide range of industries,” Conneau said. “It’s inherently horizontal, offering broad applications beyond just one sector.”
