Return Zero Inc.
Category: Voice / Speech AI
South Korea's leading voice AI startup specializing in STT (speech-to-text), speaker diarization, and AI voice agent technology, operating the VITO call transcription app and CALLABO meeting analysis platform. Return Zero Inc. was founded in 2018. The company is led by Lee Chamsol (이참솔). Based in Seoul, South Korea. Team size: 27. Total funding raised: $17.7M. Latest round: Series C. Key investors include Kakao Ventures; KTB Network; Atinum Investment; Hana Ventures; Company K Partners; Base Investment; The Ventures; Woori Venture Partners; Shinhan Venture Investment; Bass Investment.
- Founded
- 2018
- Headquarters
- Seoul, South Korea
- Team size
- 27
- Total funding
- $17.7M
Value proposition
Most accurate and fastest Korean speech recognition AI with proprietary E2E model, 15M+ hours of Korean voice data, 2.5x faster conversion, up to 3x cheaper per second, and 35-48% lower error rates vs competitors in phone noise environments.
Products and solutions
1) RTZR STT - B2B speech recognition API with cloud/on-premise deployment; 2) VITO - B2C Android call transcription app (1M+ users, 3B+ calls processed); 3) CALLABO - B2B AI meeting auto-recording & summarization platform
Unique value
Industry-leading Korean speech recognition accuracy (4.66% error rate vs Naver 9.09%, Google 14.11%, OpenAI 17.27%) powered by the largest Korean speech dataset (15M hours) from the VITO app with 1M+ users.
Target customer
Enterprises (financial institutions, telecom, public sector, call centers), B2B SaaS companies, and B2C mobile users needing call transcription
Industries served
Finance/Banking, Telecommunications, Public Sector/Government, Fire & Safety, Contact Centers/AICC, Enterprise Knowledge Management
Technology advantage
1) 1,500만 시간 (15M+ hours) Korean voice dataset—largest in Korea; 2) Proprietary E2E model with 4.66% CER (vs Naver 9.09%, Google 14.11%); 3) 30+ patents; 4) Proactive Voice Agent (PVA) technology combining STT + NLU; 5) Speaker diarization supporting up to 10 speakers; 6) 2.5x faster and 3x cheaper than competitors
How they differentiate
Full AI lifecycle ownership (data collection → model training → product/service delivery) unlike competitors who only offer API; proprietary E2E model 20% faster than competitors; largest Korean speech dataset (15M hours vs competitors); 30+ patents; supports both cloud and on-premise deployment with ISO 27001 certification.
Main competitors
Naver Cloud (Clova Speech); Google Cloud (STT API); OpenAI (Whisper); ETRI (Korean government AI)
Key partnerships
Ubase Group (acquirer - May 2026); Seoul National University NLP Lab (industry-academia partnership); CS Sharing (AI Contact Center co-development); MindWareWorks (Japanese AICC partnership)
Notable customers
Shinhan Bank (top Korean bank - AICC deployment); Gwangju Fire Department; MindWareWorks (Japan); ChannelTalk (B2B SaaS); Chun-Myung (fortune telling app)
Major milestones
2018.03: Company founded; 2019.04: First Korean free-speech colloquial-specific model; 2020.04: VITO app launched; 2020.12: First Korean colloquial speaker diarization tech; 2021.07: Series B $160억; 2024: CALLABO records 21,690 hours of meeting data (400% growth); 2025.11: Series C $50억 from Woori Venture Partners & Shinhan Venture Investment; 2026.05: Acquired by Ubase Group
Growth metrics
VITO app: 1M+ downloads, 3B+ calls processed, 15M+ hours of voice data; CALLABO: 21,690 hours of meeting data recorded in 2024 (400% YoY growth); Employee count: 27 (2026), peaked at 65 (2022)
Market positioning
Market leader in Korean speech recognition for enterprise use cases, particularly strong in financial AICC (AI Call Centers), with the largest proprietary Korean speech dataset. Positioned as the most accurate and cost-effective Korean STT solution compared to global big tech alternatives.
Geographic focus
South Korea (primary), expanding to Japan (MindWareWorks partnership) and global markets
Patents and IP
30+ patents (Korean), including speaker diarization, noise-robust STT, custom vocabulary recognition, and incremental language model learning technologies
About Lee Chamsol (이참솔)
Ex-Kakao Software Engineer (2011-2015); Ex-Roti-pl (로티플) Founder/CEO (2011); KAIST Computer Science graduate
Official website: https://www.rtzr.ai/