Search experiences break down quickly when content grows more complex and multilingual. This customer story blog post shows how Northwestern University Libraries rebuilt search using generative AI on AWS to better connect users with rich text, audio, and visual content. Read the story to see why the team chose AWS for flexibility, cost transparency, and fast iteration, then reach out to Technologent to talk about applying similar approaches to your AI search or discovery projects.
What is the purpose of Northwestern University's new AI search tool?
Northwestern University aimed to enhance the accessibility and usability of its extensive digital collections by creating a multilingual generative AI search tool. This tool is designed to provide a more intuitive and inclusive search experience, particularly for users who struggle with traditional search methods that rely on keywords and Boolean logic.
How does the AI search tool improve user experience?
The AI search tool allows users to search by concept rather than just keywords, making it more approachable for non-experts. It supports multilingual queries, providing responses in the user's language even when the underlying metadata is in English. Additionally, it surfaces content related to emerging and underrepresented topics, improving discovery and accessibility.
What technologies did Northwestern use to build the AI search tool?
Northwestern University leveraged several AWS services to build the AI search tool, including Amazon Bedrock for experimenting with foundation models, Amazon OpenSearch Service for semantic search capabilities, and AWS Lambda for backend processing. This combination allowed for a flexible, scalable, and efficient architecture that supported rapid development and deployment.
Sign in with LinkedIn