Advancing research with AI: New features in the Web of Science Research Assistant

New Web of Science Research Assistant capabilities streamline search and discovery and help researchers explore the landscape through interactive visualizations and simple analytics.

As rapid advancements in AI technology reshape the academic research landscape, Clarivate is committed to creating responsibly designed solutions that empower researchers to work smarter and more efficiently. In September 2024 we launched the Web of Science Research Assistant, a generative AI-powered tool designed to help researchers get more out of their interactions with Web of Science Core Collection data. Since that time, we have delivered several valuable enhancements based on deep analyses of user behavior and conversations with institutional partners.

Streamlining search with new features

The Web of Science Research Assistant helps researchers search more efficiently and extract key takeaways from the literature faster. A thorough beta phase preceded the assistant’s launch, during which we engaged librarians and researchers to shape features that tackle real research challenges today. Post-launch analysis of user feedback revealed opportunities to further optimize the search experience.

Several recent enhancements give users more flexibility and control over their queries while improving the relevancy and accuracy of returned results:

Editable queries enable users to quickly edit and rerun a search to refine results and discover content even more relevant to your task.
Structured responses group introductions, key themes and conclusions so that users can easily locate the information most relevant to you.
Intelligent semantic search automatically leverages the Web of Science knowledge graph to expand upon the terms in each query, helping users capture all relevant literature.
Contextual search memory makes it easy to expand or refine the information provided in a previous response.

Search memory allows users to follow up on queries within a session by retaining the context of the ongoing conversation, and it can handle multiple conversation threads throughout a session.

As we implement search memory, we remain committed to the highest standards of user privacy and security. Across the Academic AI Platform, we do not share or pass any publisher content, library-owned materials or user data to large language models (LLMs) for any purpose. Only the content that users input into a query is transmitted to the LLM, which is accessed through a private setup. To protect sensitive information, no additional data is shared during this process.

Delivering high-value visualizations and analytics

In addition to helping users search with ease, the Web of Science Research Assistant surfaces meaningful connections between people and papers at a researcher’s point of need. Some questions are best answered through visual representations of data. The assistant recognizes these questions and proactively serves up interactive maps and simple analytics alongside referenced responses. Since launch, we have released new capabilities that deliver insights faster:

Simple analytics help users compare institutional and regional levels of output on a topic to understand the landscape.
Visualizations on demand enable users to directly request topic maps and publication trend graphs on your subject of interest.
Co-author maps reveal collaborative relationships between researchers in an interactive visualization users can explore.
Profile summaries provide a more extensive view of a researcher’s contributions and impact.

Figure 1: Web of Science Research Assistant co-author map

As researchers learn more about a new topic area or develop strategies for collaborative research projects, they need information about the key players in a field. The assistant delivers pre-generated summaries of claimed Web of Science Researcher Profiles or Highly Cited Researchers. These summaries leverage profile data to deliver information about fellow researchers, including professional background, research areas, key publications, research impact, trending research interests and impact beyond academia.

Guided by the research community

By listening to feedback, analyzing customer needs and iterating quickly, we have delivered impactful features and capabilities that help our users research smarter. We are excited to see how the early Web of Science Research Assistant enhancements will empower researchers to unearth areas of opportunity and make new discoveries. Community engagement has guided every step of our development process to date, and librarians and researchers continue to help identify the most significant opportunities for AI technology to improve the researcher experience.

Take a quick tour of the recent Web of Science Research Assistant enhancements.

Research