Legacy annotation firms failed at audio emotion labeling, forcing in-house creation
Mati Staniszewski · No Priors Ep. 143 | With ElevenLabs Co-Founder Mati Staniszewski · Dec 11, 2025 · at 17:15
ElevenLabs CEO Mati Staniszewski discusses the challenges of dataset creation and speech emotion labeling with Sarah Guo on No Priors.
0:00 / 0:28exact quote · 28.9s
▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark
“The understanding of how you describe audio data is still lagging in the industry. Like, when we initially started, we, of course, went into the traditional players for them to help us label not only what was said, so, like, transcription, but also how it was said. Like, what are the emotions used, accent. And most people just weren't able to do that work effectively because you kind of need to hear and have, like, a little bit of a skill set of, like, how would I describe this specific delivery? So we needed to create data ourselves.”
quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →
More from Mati Staniszewski
Audio AI progress requires architectural breakthroughs rather than raw compute scale
“The main part that I think is different in audio space is that you don't need the scale as much as you need the, Architectural breakthroughs, the model breakthroughs to really make a dent.”
Voice conversational AI will pass the Turing test within a year
“I think this is, is good for customer support or customer experience, but it's still a way away from like conversation like we have and like passing that Turing test. So I think this is still like a, at least a year, like within a year. And then you will have …”
Staniszewski: AI research is only a head start; ecosystems create lasting value
“The thing that will really give that long-term value is the ecosystem that you create around, whether that's the run and distribution, whether that's the collection of voices you can have, the collection of integrations you can build, the workflows that you ca…”
Staniszewski: Cascaded voice architectures will dominate enterprise for 1–2 years
“And of course, depending on the use case, if you are, like, enterprise-reliable use case, the cascaded approach is the approach for the next year to...”
ElevenLabs reached $300M ARR, split evenly between self-serve and enterprise
“We are at three hundred million in, in, in ARR, which is roughly fifty-fifty between SelfServe, so a lot of subscription and creators using our creative platform, and then Approaching 50% on the enterprise side using our agents platform work, and that's on the…”