Jul 13, 2026 · 52 min
AI-native startups disrupt trillion-dollar voice and legal industries
The Trillion-Dollar Industries AI Is Disrupting: Voice, Law & the End of the Billable Hour
The rapid rise of voice AI and specialized legal models is dismantling traditional corporate structures and professional billing models.
- 1Voice AI leader ElevenLabs reached a $600 million ARR run rate using small, specialized developer teams instead of product managers.
- 2AI-native legal assistant Leya is dismantling the billable hour by shifting junior lawyers to orchestrating AI agents.
- 3Startups are building defensible data moats using specialized models that outperform legacy giants like LexisNexis.
Don't miss
Mati Staniszewski explains how ElevenLabs scaled to a $600M ARR run rate while completely bypassing the traditional product management role.
The brief
AI-native startups are scaling at breakneck speed by bypassing traditional software roles. Voice tech giant ElevenLabs reached a $600 million ARR run rate by operating entirely without product managers, relying instead on small, highly specialized developer teams.
This rapid growth is fueled by a shift in how humans interact with machines. The rise of sophisticated voice interfaces and advanced speech recognition tools like Wispr Flow is turning raw voice input into a primary driver for complex software workflows.
As voice AI scales, protecting IP has become a major challenge. ElevenLabs is navigating this by building a voice marketplace that has paid out over $22 million to voice actors, while licensing iconic voices like Darth Vader through partnerships with Disney.
The disruption is also hitting high-paying professional services. Max Junestrand, co-founder of Leya, explains how AI-native legal assistants are dismantling the traditional billable hour, shifting junior lawyers from manual research to orchestrating AI agents.
Legacy giants like LexisNexis face a steep challenge. By aggregating global case law into specialized models, AI-native startups are building defensible data moats that handle complex, end-to-end legal strategies far more efficiently than legacy software.
What was said on this episode
37 statements · 26 positive · 3 negative · 3 mixed · 5 neutral
ElevenLabs built a text-to-speech model capable of sounding human.
“We built the first text to speech model that finally could sound human.”
Listen at 1:07
ElevenLabs is building an AI communication platform spanning audio generation and speech interaction.
“we are building a communication platform for AI”
Listen at 2:33
ElevenLabs is among the leading companies for audio and interaction research.
“we are one of the leading, if not the leading place to do that”
Listen at 3:41
Embedded engineers help teams adopt AI while security-checking deployed software.
“make sure that people are adopting AI, but also there's a security check for everything they deploy”
Listen at 5:42
AI can elevate people from amateur to advanced proficiency across multiple functions.
“if you can do a little bit of all with AI, you can maybe step change from being an amateur to being advanced level”
Listen at 7:46
Using extensive AI tooling improves overall job performance.
“using a lot of tooling makes you yourself better in your job overall”
Listen at 8:58
AI interfaces will increasingly adapt dynamically to users’ behavior.
“the whole interface will change and morph depending on how you are operating with that interface”
Listen at 10:59
Voice AI will proactively provide assistance before users request it.
“will shift from reactive to proactive to help you get that help before you potentially ask for it”
Listen at 11:07
People disclose sensitive financial situations more openly to AI than humans.
“With AI, people are much more open to share what actually happened”
Listen at 14:59
Users respond more tersely to AI voice agents than to human agents.
“Usually people are more snappy with AI voice agent”
Listen at 15:13
ElevenLabs moderates generated voice and text content.
“we moderate both on the voice and text level”
Listen at 19:08
ElevenLabs can flag and block commercial or scam-related voice inputs.
“if you were to input something that would be commercial in nature or would try to scam someone, that gets flagged, we can block it”
Listen at 19:10
AI voice technology preserves emotional expression across multiple languages.
“the voice can be carried not only English, but also in Spanish and Italian and Portuguese. And you can still have exactly that element of emotions coming through.”
Listen at 20:04
ElevenLabs has paid more than $22 million to its voice-talent community.
“Today, we paid back over $22 million back to the community of talent.”
Listen at 21:03
Voice AI can restore speech for people who lost their voices to ALS or throat cancer.
“our most important work was actually working with people that lost their voice due to ALS due to throat cancer and working on bringing that voice back”
Listen at 21:49
AI enabled live interactive Darth Vader characters in Fortnite.
“Fortnite, so Epic Games Fortnite launched Darth Vader, which people and players could interact with live”
Listen at 23:37
ElevenLabs has outperformed frontier AI companies on several voice-model capabilities.
“We've been able to outcompete them on voice models, both on text to speech, speech to text, on the turn taking on music.”
Listen at 26:45
For voice-model research, architecture matters more than model scale.
“it's on the research side. It's the architecture that matters, not the scale.”
Listen at 27:04
High-quality voice models require extensive human labeling of audio data.
“you build an internal team of over 1,000 contractors that label all those audio assets to make them good”
Listen at 27:19
ElevenLabs expects voice conversations to feel humanlike within the year.
“we hope this year we'll do that same thing for voice, where any conversation feels like you are speaking with another human”
Listen at 30:18
The company has sustained 50% quarter-over-quarter growth for seven quarters.
“it's sustained 50% quarter over quarter for the last seven quarters”
Listen at 31:51
Annual spending on legal services is approximately one trillion dollars.
“It's a trillion dollars every year into legal services”
Listen at 34:25
Legal-technology software spending is approximately $40 billion annually.
“the software spend into legal technology is about 40 billion”
Listen at 34:30
Legal-technology software will expand substantially relative to legal-services revenue.
“The software piece should be much bigger than that.”
Listen at 34:41
Legora’s tool supported in-house diligence for acquisitions, including a 12-day transaction.
“We acquired four businesses so far this year. We did the diligence in house with our own tool. And the fastest transaction we did was 12 days from LOI to closing.”
Listen at 36:48
Technology is completely transforming the legal-services industry.
“one of the biggest industries in the world now is being completely transformed and reshapen as a consequence of the tech”
Listen at 37:41
Kirkland Ellis generates approximately $10 billion in annual revenue.
“Kirkland ellis turns around $10 billion a year.”
Listen at 38:10
AI poses both an existential threat and opportunity to major law firms.
“when something like AI comes along, that poses existential threat and existential opportunity”
Listen at 38:24
Legal AI can perform substantial legal work and reshape junior-lawyer roles.
“This can do a lot of the work. And so it's really reshaping what it also means to be a junior lawyer going into this occupation.”
Listen at 39:26
Junior legal jobs will remain, but their tasks will change.
“The job will exist, the tasks will be different.”
Listen at 39:45
Legora can immediately provide approximately 80%-accurate answers across foreign jurisdictions.
“they can get an 80% accurate response immediately that they can start working out of”
Listen at 42:04
Legacy legal-information providers struggle to become AI-native businesses.
“some of the existing providers and the sort of legacy players have a really hard time pivoting into becoming AI native businesses”
Listen at 43:28
A complete legal-research solution requires comprehensive legal data.
“You cannot build a legal research solution that doesn't have all of the data.”
Listen at 45:44
Legal AI agents can combine evidence and cases to perform end-to-end case work.
“they can now start to do really intelligent case strategy and they can actually start to combine the witness statements, the cases, and they can really do end to end work”
Listen at 46:57
Claude’s legal offering generates leads for Legora after users encounter its limitations.
“it drives a lot of initial usage there. And then you hit the ceiling or you understand how shallow it is, and then you Call us. So it's actually a big pipeline generator for us.”
Listen at 48:12
Narrow, scaled models can reduce costs and latency for specialized legal tasks.
“I do believe in very narrow models for narrow use cases that you also drive a lot of scaling, so you can drive both cost and latency down.”
Listen at 48:59
VPC deployment consumes time and slows product roadmaps and execution.
“deploying in a VPC is very time consuming and it creates a lot of dependencies which slow down your roadmap and the execution forward”
Listen at 50:56
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
