How We Revolutionized NLB Library Experience Using Generative AI
Overview
NLB, National Library Board of Singapore, is made up of a group of libraries and archives in Singapore dedicated to promoting reading, learning, information literacy, and an interest in Singapore's history and heritage. With a mission to reach and serve the majority of residents in the country, NLB sought to improve public access to information about Singapore's history and enhance user engagement through a user-friendly and interactive platform.
About National Library Board of Singapore
The National Library Board (NLB) nurtures Readers for Life, Learning Communities and a Knowledgeable Nation by promoting reading, learning and history through our network of 28 libraries across Singapore, the National Library and the National Archives of Singapore.
Challenges
NLB's extensive collection of books and archives containing Singapore's history presented a significant challenge in terms of efficiently retrieving and synthesizing information for users and readers. The traditional method of manually searching through physical books and documents was time-consuming and often resulted in a suboptimal user experience.
NLB was facing challenges in getting the general public to effectively browse reading materials concerning the history of Singapore. These challenges included:
Inefficient Search Process
Visitors have to browse through the books and articles to look for information they wish to find out about
Dull Interface
Traditional reading materials such as books and articles are not interactive by nature, which stands out as a pain point as opposed to other interactive mediums introduced through technological advancements.
Solution
To address these challenges, AXRAIL implemented a generative AI chatbot solution leveraging Amazon Bedrock and Amazon Kendra. The chatbot is based on the Claude 2 model from Anthropic, a leading foundation model (FM) in the field of generative AI of that time.
The content of NLB's books and archives was indexed into Amazon Kendra, an intelligent enterprise search service, which serves as the authoritative knowledge source for the Bedrock model. This methodology of providing relevant information to optimize the response from foundational models is commonly known as retrieval augmented generation (RAG).
When a user interacts with the chatbot, their prompt is augmented with relevant information retrieved from the Amazon Kendra knowledge base. This augmented prompt is then processed by the Bedrock model, which leverages the Claude 2 FM to generate an accurate and contextual response to the user's query.
To further enhance the chatbot's performance, AXRAIL employed advanced techniques such as few-shot prompting on the Claude 2 model using their domain-specific data. These techniques allowed the chatbot to better understand the nuances and context of Singapore's history, resulting in more accurate and relevant responses.
Outcome
By implementing this generative AI solution with Amazon Bedrock and Amazon Kendra, NLB has achieved the following measurable outcomes:
Estimated Time Saved
Previously, locating specific information within NLB's collection might take users an average of 10-15 minutes. With the chatbot, users can now retrieve relevant information almost instantly.
Enhanced Accessibility
The chatbot offers a user-friendly interface, making NLB's vast collection of historical information accessible to a wider audience. During the beta phase, up to 97% of users reported no issues with using the chatbot.
Increased User Satisfaction
Faster and more efficient information retrieval leads to a more positive user experience for NLB's patrons. In the beta phase, 82% of users agreed that the chatbot was more convenient than manually searching through physical books.
Conclusion
By leveraging the power of Amazon Bedrock and Amazon Kendra, NLB has successfully implemented a generative AI chatbot solution that addresses the challenges of providing efficient and user-friendly access to information about Singapore's history. The solution has resulted in significant time savings, enhanced accessibility, increased user satisfaction, and improved research efficiency, enabling NLB to better fulfill its mission of promoting reading, learning, and an interest in Singapore's heritage.
With a team who is AWS certified and experienced in AWS and consulting, AXRAIL has been serving clients ranging from startups to enterprise, covering multiple industries like retails, media, transportation and well-known service providers. We are proud to deliver the state-of-the-art chatbot to revolutionize the reader's experience at NLB.
Why Work With Us?
AXRAIL is the leading AWS Advanced Consulting Partner that provides cloud technology services in Singapore and Malaysia. We specialize in cloud services, including migration, B2C/B2B e-commerce platforms, and document automation processing.
Our expertise in AI, analytics, and modernization enables us to provide innovative solutions that drive growth and streamline operations. Trust AXRAIL for unparalleled expertise and commitment to excellence.
