Unveiling How Siriraj Hospital’s SiData+ Uses Typhoon LLM for Smart Knowledge Management and Admin Support

At the heart of Thailand’s largest public hospital, innovation meets real-world impact. The Siriraj Informatics and Data Innovation Center (SiData+) team at Siriraj Hospital, renowned for driving data and AI excellence in healthcare, faced a familiar challenge: how to support an ever-growing stream of internal inquiries without overwhelming their team.

Their answer? A smart, AI-powered chatbot—built on top of Typhoon.

The Challenge: Supporting Numerous Requests, Every Day

Inside Siriraj’s data innovation center, the SiData+ admin team fields a high volume of questions and requests from doctors, researchers, and hospital staff across three critical services:

• Data Governance: Oversee and control data usage within the Faculty of Medicine Siriraj Hospital in accordance with laws, regulations, requirements, and ethical standards.

• Data Management: Collect and manage data within the Faculty of Medicine Siriraj Hospital.

• Data Utilization: Research and develop beneficial data applications through Data Analytics and Data Science methodologies, such as Machine Learning, Artificial Intelligence, and Data Visualization.

The workload was relentless. Admins had to answer calls all day, look up technical details, and often repeat the same answers—wasting valuable time and talent.

The Solution: A Chatbot That’s Tailored to Siriraj’s Unique Knowledge Base

To address this, the team developed the SiData+ Admin Chatbot—a 24/7 assistant powered by Typhoon LLM and Retrieval-Augmented Generation (RAG) technology. This allowed the system to understand complex questions in Thai, fetch accurate answers from an internal knowledge base, and provide traceable sources for every response.

Why Typhoon?

Several factors led the SiData+ team to choose Typhoon LLM:

🔐 Data Sovereignty

Unlike foreign providers, Typhoon is deployed and maintained in Thailand in partnership with SCB 10X—ensuring all hospital data stays within the country and meets local data security standards.

This project places the highest priority on data security, especially given its context within a hospital. Our system follows strict security protocols in accordance with hospital data protection standards and complies fully with the Personal Data Protection Act (PDPA). The content of the questions and answers pertains only to service procedures and general operational guidelines—there is no involvement of patient personal data. This allows us to operate smoothly without compromising privacy concerns.

— Peamboon Thomchotpong, MD Clinical Ambassador, SiData+

🧠 Mastery of Thai Language Processing

From formal Thai to medical jargon, Typhoon consistently interprets questions and delivers clear, correct responses in Thai—something few global models can do natively.

💸 Cost-Effectiveness

Typhoon offers a significantly more affordable option for long-term use—especially for high-volume applications. It delivers comparable performance to proprietary models. Plus, its open-source foundation allows for deeper customization without additional licensing fees.

One of the biggest advantages of Typhoon is its adaptability to our context—whether that’s language, infrastructure, or regulatory needs.

— Tossaporn Saengja Data Scientist, SiData+

Early Impact: Faster Answers, Happier Teams

The pilot focused on the three main service areas of SiData+. Within weeks, results were clear:

• Staff could get answers 24/7 without waiting

• Admins spent less time repeating information

• The team had removed the repetitive tasks and gained more bandwidth to tackle complex tasks

“Not only did it reduce the load, but it also boosted team satisfaction,” the team noted.

What’s Next: Scaling the Vision

SiData+ plans to expand the chatbot’s role across three phases:

1. First 3 Months – Build, test, and refine the system with a small user group

2. Next 6 Months – Roll it out hospital-wide, gather feedback, and improve

3. Year 1 – Add more complex Q&A, integrate with other hospital systems, and explore new AI use cases

Ultimately, the team sees this as just the beginning. “This model could become a template for deploying AI across healthcare—making hospital operations more efficient and empowering staff with the right tools.”

Looking Ahead: Building a Thai Health AI Ecosystem

SiData+ team and SCB 10X R&D (Typhoon) team

The SiData+ team believes the future of healthcare AI in Thailand will accelerate rapidly—especially with locally developed Thai language models like Typhoon LLM. For them, the key lies in building a collaborative ecosystem where healthcare organizations, developers, and AI providers work together to solve real, domain-specific challenges.

Typhoon shows what’s possible when we collaborate with local AI experts and have the freedom to customize open-source models to fit our healthcare needs. The chatbot is just the beginning—we see opportunities to extend AI into clinical decision support, workflow optimization, and care quality improvement.

— Peamboon Thomchotpong, MD Clinical Ambassador, SiData+

Their advice to fellow developers is clear: start small, stay focused on real needs, and make data privacy non-negotiable.

You don’t need the flashiest tech. Start with solutions that reduce real burdens—like answering repeated questions—and grow from there. Local models like Typhoon give us the flexibility and control that healthcare demands.

— Tossaporn Saengja Data Scientist, SiData+

Keep Exploring

🔎 Learn more about SiData+ and how they’re transforming Thai healthcare with data and AI.

🎥 Watch the recorded panel discussion from the Typhoon 2 Unveiled event—featuring real-world use cases from organizations like TDRI, VISAI AI, and AutoX. Individual spotlight interviews coming soon.

📣 Building something with Typhoon? We’d love to hear your story. Tell us more and be featured on our platform.