AI Tools from Elsevier
Scopus AI
"Scopus AI" is a tool, or set of tools, that generates research summaries from the abstracts of scholarly content within the Scopus database.
"Scopus AI" is available in: Scopus
- Select the "Scopus AI" tab above the search box.
Enter a natural language query to generate an easy-to-digest summary of your topic, including references.
- Example: how do fireflies produce light
- Click the "Temporary conversation" toggle if you do not wish Scopus to maintain a list of your searches for your future use.
- The "Deep research" feature will be addressed at the bottom of this page.
The research summary will include citation numbers within the text corresponding to the numbered items in the reference list in the right-hand sidebar.
Clicking on a reference number in the text, or an item in the reference list will reveal the document abstract.
The "Copilot" feature shows you the keywords and search syntax that was generated from your original query.
- Click on the "Show Copilot steps" link to reveal the keyword search string that was generated based on your natural language query.
Tip: Copy the search string (and modify if needed), and enter it into a standard Scopus search to compare standard search results, and to retrieve older articles (before 2003) if applicable to your topic. You can also use this search string in other databases.
Additional features include:
- Expanded Summary — offers more depth and a wider range of perspectives, supporting further exploration of your topic.
- Concept Map — a visual depiction of your topic area, and related topics and subtopics.
- Topic Experts — identifies authors active in your topic area.
- Emerging Themes — helps to identify "consistent," "rising," and "novel" themes related to your topic by scanning Scopus documents from the last two years.
- "Go deeper" questions — three questions designed to help you drill down into your topic, and to broaden your understanding.
Scopus AI: Deep Research
Scopus AI: Deep Research uses a different type of AI—Agentic AI—to produce a multi-page research report based on your search query.
The Deep Research report includes the following sections:
- Overview of key findings
- Suggested research questions and study methodologies
- Study scope including time period and disciplines used
- Areas for suggested further research
- Discussion and future research directions
- Conclusion and recommendations
- References for findings
Deep Research is intended as a tool to save time with preliminary research.
"... the tool is carefully designed to avoid mimicking important areas that require human creativity and input, for example a traditional literature review. That’s because we don’t want to replace the work of researchers, we want to support it, with insights that will guide you toward your next discovery." — Six tips for using Scopus AI Deep Research.
View a sample deep research report here:
Click the "Deep Research" button to generate the multi-page report based on your query.
Tip: This report may take several minutes to generate.
More information on "Scopus AI"
More information on "Scopus AI" — Deep Research
Six tips for using Scopus AI Deep Research
Reading Assistant
Elsevier's "Reading Assistant" tool provides document summaries and answers to questions posed about specific articles within the ScienceDirect database.
"Reading Assistant" is available in: ScienceDirect
- Sign in to your ScienceDirect account to gain access to the full "Reading Assistant" features.
- Tip: Sign in "via your organization" and find University of San Francisco to use your myUSF login.
- Perform a search in ScienceDirect to retrieve a set of results.
Click on a title from the results list to bring up the full record, and the "Reading Assistant" entry box will appear on the left.
- You can ask a question regarding the item you've chosen, or simply ask the tool to summarize the article.
- Tip: You are allotted 5 articles per month to use with the Reading Assistant. Additional usage can be purchased.
This is an example of an article summary, with numbered deep link references to source material within the article.
The Reading Assistant also suggests "Questions you could ask" about the content of the article, and "Actions you could take" such as "summarize this article," or "summarize experiments" which would briefly outline the experimental process.
This is an example of the "Summarize experiments" option.
More information: Getting Started with the Reading Assistant
Under the Hood
What powers Elsevier's AI tools?
Elsevier’s “Scopus AI” in the Scopus database is powered by a retrieval-augmented generation (RAG) architecture that combines advanced vector search, reranking (using models like miniLM), and large language models (primarily OpenAI’s GPT hosted on Azure and proprietary Elsevier LLMs). The system retrieves and ranks relevant abstracts and metadata from Scopus, then generates summaries and insights solely from this trusted content, ensuring outputs are transparent and cite sources. Privacy is prioritized, with strict controls on data usage and adherence to GDPR.
The “Reading Assistant” in ScienceDirect leverages generative AI—again, a RAG approach—focused on full-text peer-reviewed articles and book chapters. It allows users to chat directly with a single document, ask questions, summarize sections, and receive cited, passage-level insights. All answers are grounded in the actual text of the selected document, helping users quickly extract, compare, and verify information while maintaining citation integrity and accuracy.
AI-assisted, human-verified text