KMITL transforms prospective student guidance with HAQM Bedrock

King Mongkut’s Institute of Technology Ladkrabang (KMITL), one of Thailand’s leading research and educational institutions, has been at the forefront of technological innovation in higher education for over 60 years.

Every year, KMITL hosts a major educational exhibition and several open houses featuring innovation and research from various sectors that attract thousands of prospective students seeking guidance about their academic future. These events traditionally require significant staff resources, with faculty members and academic advisors spending countless hours providing program information, career guidance, and addressing applicants’ questions.

Through a strategic collaboration with HAQM Web Services (AWS) and DailiTech Co. Ltd, an AWS Partner, KMITL developed and delivered a prospective students virtual assistant “Knowva,” a generative AI application powered by HAQM Bedrock and HAQM Bedrock Knowledge Bases.

HAQM Bedrock is a fully managed service that offers a choice of high-performing foundation models (FMs) from leading AI companies like AI21 Labs, Anthropic, Cohere, Meta, Stability AI, and HAQM via a single API, along with a broad set of capabilities you need to build generative AI applications with security, privacy, and responsible AI.

The Knowva application allows prospective students to ask the questions related to program information, degree requirements, admission processes, and career prospects. This innovative generative AI solution enables KMITL to:

• Provide instant, 24/7 responses to prospective students’ inquiries
• Offer personalized guidance based on individual student interests and academic backgrounds
• Deliver consistent and accurate information across all interactions
• Scale support services to handle thousands of concurrent users during peak periods
• Support multilingual communication in both Thai and English

Figure 1: Knowva, an intelligent prospective student assistant from KMITL.

The challenge: Scaling prospective student guidance

Before implementing Knowva, KMITL faced significant challenges in managing prospective students’ inquiries every year during their open houses and educational exhibition. The traditional approach of relying on faculty members and staff to provide guidance was becoming increasingly unsustainable as the number of prospective students and inquiries grew every year.

“Our faculty members were spending significant time repeatedly answering similar questions about program requirements, career prospects, and admission processes,” said Associate Professor Komsan Maleesee, D.Eng. President at KMITL. “Additionally, our sources of documents are spread across different departments and exist in various versions. We needed a solution that could handle these routine inquiries efficiently, picking up content from various sources and combining them, while allowing our staff to focus on more complex student needs.”

KMITL faced several critical challenges:

Time sensitivity and resource constraints

• Only two weeks to deploy a solution before the major educational exhibition
• Faculty members overwhelmed with repetitive inquiries during peak periods
• Limited staff availability to handle increasing volumes of prospective students’ questions

Information management and consistency

• Program information scattered across different departments
• Multiple versions of documents leading to potential inconsistencies
• Need for unified, accurate responses across all programs and departments

Scalability and accessibility

• More than 6,000 prospective students during exhibitions and open houses
• Need for 24/7 availability of accurate program information
• Requirement for quick response times
• Limited ability to scale human resources during events

The solution: Building an AI advisor with HAQM Bedrock

To address these challenges, KMITL selected and worked with AWS to build an AI advisory solution, which harnesses the agility, scalability, and resiliency of the AWS Cloud. Moreover, HAQM Bedrock allows KMITL to try and test different FMs from leading AI companies. Ultimately, KMITL leverages Anthropic Claude to provide natural, context-aware responses in Thai to prospective students’ inquiries.

Secure foundation and development

The implementation began with establishing a secure AWS environment. DailiTech, as an AWS partner, provided expertise in:

• Implementing AWS Landing Zone, a well-architected, multi-account AWS environment that is a starting point from which the workloads and applications are deployed, following security best practices
• Setting up proper network segmentation and access controls
• Configuring monitoring and logging capabilities

With the secure foundation in place, KMITL’s team can focus on building the Knowva application by:

• Developing the chat interface for prospective students
• Integrating with Anthropic Claude on HAQM Bedrock using API
• Consolidating the program information and structuring the Knowledge Base contents

Figure 2: High-level architecture of Knowva. The major components are HAQM Bedrock, and HAQM Bedrock Knowledge Bases.

The results: Elevating admission guidance through AI

The implementation of Knowva has delivered significant improvements in KMITL’s ability to support prospective students while optimizing staff resources. Key achievements include:

Operational impact

• Successfully handled more than 6,000 student inquiries during the educational exhibition and scaled up to hundreds of concurrent inquiries
• Reduced waiting and response time from hours to minutes
• Enabled 24/7 availability of accurate program information

Enhanced student experience

• Provided instant, accurate responses to common admission queries
• Delivered personalized guidance based on individual interests
• Enabled students to access information at their convenience
• Increased student satisfaction by 60 percent compared to the previous year

Staff and resource optimization

• Freed faculty members from repetitive inquiry handling
• Centralized program information management
• Enabled staff to focus on handling complex student needs

“Knowva has transformed how we engage with prospective students, we’ve seen a significant increase in student satisfaction and a more efficient use of our staff resources with AWS technology,” commented Associate Professor Komsan Maleesee, D.Eng. President at KMITL.

Beyond Knowva: KMITL’s AI roadmap

After the successful launch of Knowva, KMITL is planning to expand the use of the generative AI solution to include:

• Virtual AI tutoring services for current students
• Campus assistant capabilities
• Integration with additional university services

KMITL is also exploring the use of multi-agent collaboration and HAQM Bedrock Guardrails features to improve the solution’s capabilities and safety.

To learn more and get started, contact your AWS account team or the AWS Public Sector team.