Spicerack developed an AI-powered conversational search interface for the American Institute of Physics (AIP). This solution enables researchers, students, and physicists to explore AIP’s published papers using natural language queries. The system employs a Retrieval-Augmented Generation (RAG) framework, integrating large language models (LLMs) and a vector database to provide semantically relevant search results accompanied by generative summaries and rigorous citations.
Benefits & outcomes
- Reduced average search time from 12 minutes to 3 minutes (a 75% decrease).
- Enhanced content discovery by 230%.
- Increased user engagement time per session from 5 minutes to 15 minutes (a 200% increase).
- Improved citation accuracy from 72% to 96%.
- Achieved a 42% boost in subscriptions and article purchases.
- 87% of users reported finding more relevant content.
- Researchers saved an average of 9 hours per month
Notes & quotes
“Spicerack swiftly responded to our RFP, with a comprehensive proposal that exceeded expectations. Their excellent and unusual partnership provides creative expertise and valuable guidance. They’re advanced in their AI knowledge, open to new ideas and technologies, and a pleasure to work with—I can’t recommend them highly enough.” — Founder, Project Partner
The conversational interface supports features such as conversation history, result sharing, and incorporates user feedback for continuous improvement.
The user interface was crafted using Next.js and React, emphasizing an elegant and intuitive design.
The solution underscores the value of AI in enhancing research accessibility without supplanting human expertise.
Publisher / company: Spicerack (in partnership with Mathison.ai)
Date of case study: 01/01/2025
Source: https://spicerack.co.uk/case-study/conversational-search