Siant is a technology company specializing in artificial intelligence customer service automation. Founded in 2021, the company develops conversational agents powered by large language models to handle customer inquiries for enterprises. Siant is headquartered in Singapore and focuses on deployment across Southeast Asian markets, including Indonesia, Thailand, and Vietnam.
The company emerged during a period of rapid advancement in generative AI, leveraging transformer-based architectures to build its core platform. Siant's products aim to reduce operational costs for businesses by automating routine support tasks while escalating complex issues to human agents. As of 2024, the company reports deployments with over 40 corporate clients, primarily in the retail, banking, and telecommunications sectors.
Founding and Leadership
Siant was co-founded by three engineers with backgrounds in natural language processing and software engineering. The founding team includes CEO Tan Wei Ming, previously a research engineer at Google Cloud's AI division, and CTO Priya Sharma, who worked on dialogue systems at Anthropic before returning to Singapore. The third co-founder, Ahmad Fauzi, serves as Chief Product Officer and previously led customer experience teams at a regional e-commerce firm.
The company raised an initial seed round of $3.2 million in March 2021, led by venture capital firm Sequoia Capital Southeast Asia. A Series A funding round of $12 million followed in September 2022, with participation from OpenAI's startup fund and local investors. These funds supported the expansion of the engineering team from 15 to 60 employees and the opening of a second office in Jakarta.
Technology Platform
Siant's core technology is built on fine-tuned large language models, specifically adapting open-source transformer architectures for customer service domains. The platform uses a proprietary dataset of over 500,000 annotated customer-service conversations in English, Indonesian, and Thai to train its models. This dataset includes common queries related to order tracking, billing disputes, product troubleshooting, and account management.
The system employs a hybrid architecture that combines retrieval-augmented generation with rule-based fallbacks. When a customer query is received, the model first attempts to match it against a knowledge base of approved responses. If confidence is low, the system escalates to a human agent with a suggested response. This approach reduces hallucination risks, a common challenge in generative AI applications.
Siant's platform integrates with major customer relationship management tools, including Salesforce and Zendesk, through RESTful APIs. The company also offers a no-code dashboard that allows clients to update knowledge bases and monitor performance metrics without engineering support. Average response time across the platform is reported at 1.8 seconds, with a resolution rate of 72% for fully automated interactions.
Products and Services
Siant offers three primary products. The flagship product, Siant Assist, is a multilingual chatbot deployed on websites and mobile apps. It supports real-time translation between English, Indonesian, Thai, and Vietnamese, enabling consistent service across regional operations. Siant Assist includes sentiment analysis features that flag frustrated customers for priority human intervention.
The second product, Siant Voice, launched in 2023, provides automated voice agents for phone-based support. It uses speech recognition and text-to-speech synthesis to handle simple inquiries such as balance checks and appointment scheduling. Siant Voice is currently available in English and Bahasa Indonesia, with Thai support planned for 2025.
The third offering, Siant Analytics, is a reporting tool that analyzes customer interaction data. It generates insights on common pain points, response quality, and agent performance. This product has been particularly popular with banking clients, who use it to identify recurring compliance-related questions and update their training materials accordingly.
Market Position and Impact
Siant operates in a competitive landscape that includes global players like Intercom and regional startups such as Chatbot Indonesia. The company differentiates itself through its focus on Southeast Asian languages and local business practices. For example, its models are trained to recognize informal Indonesian slang and Thai polite particles, which generic English-centric systems often mishandle.
In 2023, Siant reported a 150% year-over-year increase in annual recurring revenue, reaching $8.5 million. The company claims its automation reduces client support costs by an average of 35%. One notable deployment is with a major Indonesian telecom operator, where Siant handles over 200,000 customer interactions monthly, covering 60% of all incoming support requests.
The company has also contributed to academic research, publishing a paper on low-resource language fine-tuning at the 2023 Conference on Empirical Methods in Natural Language Processing. This work, co-authored with researchers from the University of Toronto, demonstrated techniques for improving model performance on languages with limited digital corpora.
Future Directions
Siant plans to expand into the Philippines and Malaysia in 2025, adding Filipino and Malay language support. The company is also exploring partnerships with AWS Trainium users to reduce inference costs, which currently account for 40% of operational expenses. In an interview with TechCrunch in 2024, CEO Tan Wei Ming stated that the company aims to achieve profitability by 2026 without additional external funding.
The broader trend of AI customer service automation continues to grow, with market analysts projecting the sector to reach $12 billion by 2027. Siant's focus on regional language support positions it to capture a share of this growth in Southeast Asia, where English proficiency varies significantly across demographics. However, the company faces ongoing challenges related to data privacy regulations, particularly in Thailand and Indonesia, which have introduced stricter requirements for cross-border data transfers.
As of late 2024, Siant employs 85 people across its Singapore headquarters and Jakarta office. The company maintains an active research collaboration with the Stanford AI Lab on multilingual model evaluation, and has open-sourced a benchmark dataset for Southeast Asian customer service conversations under an Apache 2.0 license.