AI Tools and Resources

Beginning in March 2025 the USF Libraries catalog, powered by Primo and hosted by the Florida Virtual Campus Library Services , will be powered with an AI research assistant.  To access the assistant within the catalog, click on the ‘Research Assistant’ option in the top menu bar.  Users must be logged in with their USF NetID to use the Research Assistant.

By entering your research question the Research Assistant’ will create a complex Boolean search string excluding newspapers and retracted documents.  The first five results will be analyzed by a Large Language Model (LLM) to create an overview of the results with in-line citations.  The overview is based on the search results and should not contain extraneous material.   

Each search is unique.  Clicking try again will give you a slightly different overview or different top-five sources.   

Clicking ‘view all results’ will take you to the full list of search results and the option to filter or refine using the search facets in the catalog.

What is the Primo AI Research Assistant?

• The Primo AI Research Assistant is a generative AI-powered tool developed by Ex Libris (part of Clarivate) designed to improve library discovery. It allows users to make research queries with everyday language and provides AI-generated summaries with references based on the Ex Libris Central Discovery Index (CDI). The tool is intended to make research more intuitive and accessible.

How does the Primo AI Research Assistant work?

• The tool converts a user's natural language question into a Boolean query using a large language model (LLM). This query is then sent to the Central Discovery Index (CDI) to retrieve results. The top results (up to 30) are then analyzed, and the five most relevant sources are identified based on their abstracts and metadata. The LLM then creates a concise summary with inline references based on those top five results, which is then delivered to the user.

What are the key strengths of the Primo AI Research Assistant?

• The primary strengths include its natural language query support, which allows users to search without advanced search skills, AI-generated summaries with inline references that provide quick access to relevant information, dynamic and reusable search strings that can be refined for further research, and multilingual support for over 50 languages.

What are the limitations of the Primo AI Research Assistant?

• Limitations include potential access challenges for users who may encounter resources they cannot directly access, limiting the search functionality to "OR" boolean modifiers, the tool's limited depth of analysis which relies only on abstracts and metadata, the risk of institutional search caps being imposed, and the inherent training biases of large language models which can affect the accuracy and neutrality of results.

How does the Primo AI Research Assistant ensure user privacy?

• The tool is designed with a privacy-centric approach. User inputs and licensed content are processed securely, and no personal data is stored or used in training the AI models, ensuring user privacy and data protection.

What is the Central Discovery Index (CDI) and why is it important to the tool?

• The Central Discovery Index (CDI) is a vast database containing over 5 billion records. The Primo AI Research Assistant utilizes the CDI as its primary source for retrieving relevant search results. The quality and breadth of the CDI directly influence the effectiveness of the AI assistant's summaries and references.

Is the Primo AI Research Assistant similar to other AI research tools like Elicit or SciSpace?

• While all these tools leverage AI to assist in research, a key difference is that the Primo AI Research Assistant's analysis is limited to abstracts and associated metadata. Tools like Elicit and SciSpace may perform deeper analysis by looking at the full text of the source, potentially leading to more comprehensive results.

Can the search strings generated by the Primo AI Research Assistant be modified or reused?

• Yes, the tool generates search strings based on user queries. These search strings can be reused or refined by the user for further research. This allows users to build on the initial search and explore the topic in more detail.