The Trillion-Dollar Industries AI Is Disrupting: Voice, Law & the End of the Billable Hour

What It Covers

ElevenLabs CEO Mati Staniszewski and Legora CEO Max discuss how AI is reshaping voice technology and legal services. ElevenLabs reached $600M ARR in under four years with 600 employees, while Legora closes M&A deals in 12 days using its own AI tools, targeting the $1 trillion legal services market.

Key Questions Answered

• •Voice AI Revenue Velocity: ElevenLabs hit $100M ARR in 20 months, $200M in 10 more months, and $300M in 5 months after that, reaching $600M total. Founders building in audio AI should note this compression curve — each revenue milestone took roughly half the time of the previous one, signaling accelerating enterprise adoption of voice infrastructure.
• •Embedded Engineers as AI Governance: ElevenLabs places engineers inside every non-technical team — legal, talent, go-to-market — with two explicit roles: building automations and auditing AI-generated code for security. This structure prevents the common failure mode where non-developers deploy AI tools into production without understanding what runs underneath, creating hidden vulnerabilities.
• •Legal AI Market Structure: Legal services generate $1 trillion annually with only 4% captured by software — a 96% service-to-software ratio. Legora targets this gap by closing acquisitions in 12 days using its own platform in-house, demonstrating that AI-native legal tools can compress transaction timelines that traditionally take months under billable-hour incentive structures.
• •Voice Identity Monetization: ElevenLabs has paid $22M back to voice creators through its marketplace, where talent records samples, authenticates ownership, and licenses their voice across languages and use cases. This model converts a one-time recording session into recurring passive income — a replicable template for any platform sitting on top of scarce human-generated creative assets.
• •Narrow Model Strategy Over General AI: Legora deliberately avoids building general legal intelligence models, instead fine-tuning narrow models for high-volume, specific tasks like tabular contract review — 100 documents times 100 prompts equals 10,000 API calls. This approach reduces both cost and latency without the capital waste of competing against frontier labs on general reasoning benchmarks.