Sea Voice
Của Sea VoiceWe build localized voice AI that listens, speaks, and responds like a Southeast Asian to supercharge enterprise workflows. Our proprietary models are trained to understand regional accents and mixed languages, and are built on a flexible infrastructure that integrates with any LLM.
- Giai đoạn
- Đang thí điểm
- Ngành
- Ngân hàng và tài chính
- Viễn thông
- Năng lượng
- Triển khai
- Đám mây (SaaS)
Giá trị mang lại
Unlike pre-LLM players, we’ve built our system with the assumption that new and better LLMs will launch every 3-6 months. Our modular infrastructure and evaluation model is designed to rapidly adopt these improvements - meaning every new LLM release makes our product stronger, not obsolete. Among post-LLM startups, what sets us apart is execution velocity and strategic clarity. As a repeat founder, Kai Song brings a rare blend of hands-on product intuition and consulting-grade client strategy. Her experience at McKinsey and GuruLab gives her a strong bias for speed, customer alignment, and knowing how to prioritize core features that drive impact. Currently, we operate like McK consultants on startup iteration cycles. Jay complements this with strong technical leadership, building and integrating each component of the AI stack with precision. His background in data science, systems architecture, and full-stack development ensures we not only ship fast, but build robust, scalable AI solutions that work in real-world environments. Our experience organising AI education programs and AI competitions in the region also give us unparalleled access to emerging AI talent that we engage to work on our projects.
- Based on product interviews with mid-to-large call center operators and recruitment agencies, we've identified several key pain points faced by enterprises in Southeast Asia that handle high volumes of customer or candidate interactions:
- High operational costs and reliance on human labor: Managing large frontline teams is expensive and resource-intensive. Skilled staff are hard to retain, and turnover leads to repeated training cycles and loss of institutional knowledge.
- Inconsistent service quality: Human-led interactions often vary in accuracy, tone, and professionalism. Quality assurance is typically manual - reviewing only 1 in 10 calls - which adds further operational costs and still leaves gaps in oversight.
- Limited capacity to engage at scale: Many organizations simply don't have the resources to meaningfully engage every customer or applicant. As a result, promising leads or candidates are often overlooked, and user experience suffers.
- Our product interviewees also expressed frustration with off-the-shelf AI solutions that struggle with Southeast Asian accents, local speech patterns, and code-mixed languages (e.g., English with Malay or Tagalog). This mismatch leads to miscommunication and reduced effectiveness, making automation efforts difficult to implement.
Đang mở deck…
Trang 1
Khách hàng phù hợp nhất
- Our ideal customers are mid-to-large enterprises operating in Southeast Asia, particularly in industries such as telecommunications and financial services.
- Key use cases include:
- First-round recruitment screening: Companies that receive a high volume of job applications but lack the resources to manually screen candidates, yet still want to make quality hires.
- Outbound sales and lead qualification: Companies running high-volume sales operations that require scalable, consistent outreach and qualification conversations.
- Training simulation: Companies looking to simulate realistic customer service or sales calls for onboarding and upskilling staff.
- Debt collection: Companies using voice communication for repayment reminders and collection follow-ups.
- Inbound order-taking: Companies handling high-frequency, repetitive call center tasks
Khách hàng tiêu biểu đang trả phí
- University of Cambridge We have not obtained clearance from the other customers for testimonials
We are currently operating in stealth mode and laser-focused on deploying high-impact AI projects and proof-of-concepts. In just three months, we have successfully delivered projects with the following partners: •University of Cambridge (Research Institute): Built an agentic AI system to classify over 15,000 codes from handwritten notes. To ensure precision, we developed a custom LLM evaluation platform, achieving over 95% accuracy. The project was a major success, resulting in an invitation to be the sole RFP candidate for inclusion in their upcoming grant proposal. • UK Care Group: Developed a voice-based AI triage assistant for NHS 111 (virtual GP), designed to streamline patient assessment and routing. The solution was showcased to members of UK Parliament at the House of Commons, highlighting its potential to improve healthcare delivery. • Multinational Bank: Designed and deployed an AI-powered interviewer to screen over 500 fresh graduate candidates. The system enabled the client to successfully hire 10 top-performing graduates within three weeks, all demonstrating strong academic credentials, project experience, and behavioural fit. • Recruitment Agency: Created a personalized AI interviewer for experienced hires, capable of generating context-aware behavioural and technical questions tailored to both the candidate’s profile and job requirements. The platform also includes automated performance evaluation to support recruiter decision-making. These rapid project deliveries demonstrate our team’s ability to work closely with enterprise client - aligning stakeholders, understanding complex workflows, and executing with speed and precision to deliver tangible business value.