Lars Hammar, CC BY-NC-SA 2.0,via Flickr
Now you have completed the systematic search, you may want to look for additional resources through citation searching, a grey literature search, visiting organisation websites or using AI discovery tools. You will record these on your Prisma Flow Diagram as "studies identified via other methods."
There may be non-academic organisations that have produced research in your area of interest - for example, charities, think tanks, NGOs or IGOs. You can find more information on the grey literature pages. You may not be able to search these sources in a systematic way, but you can still include their work in your review if it meets your criteria.
While you can not use Google Scholar for your systematic search, you can opt to run a version of your search and screen the first few pages of results for ones that meet your criteria.
This term refers to searching through a specific journal - a task that would have been done by hand before online journals. Our database coverage is very broad, but you may find there is a particular journal in your area of interest that you want to search, but it's not indexed in any of the databases that you are using.
You will probably find that you can't run your full search strategy on a journal's website, but you can still include relevant articles found by searching a specific journal in your review.
If you aren't sure how to check if a journal is indexed in your choice of databases, ask your subject librarian.
A citation network refers to the papers that an article has referenced and papers that is has been cited in. By exploring these networks, you may find more research that is relevant to your literature review. When you view the article on the journal page, or via a database, you will see links to both the references (the papers it cited) and citations (the papers that cited it).
There are also AI tools that will help you explore these network - they can not create the comprehensive search that is required for a systematic search, but they have potential for supporting supplementary searches. You can learn more about the role of AI in information finding via our guide to AI and the Library. Before using any AI tools, please follow guidance from your school and the University's academic integrity guidelines.
Any AI tool mentioned in this guide is not an endorsement or recommendation of its use by the University of Bristol or the library.
Key function: takes a specific article and then identifies similar papers via its co-citation network. It selects the "most closely related" papers and displays them as a graph or list.
Data source: Semantic Scholar
Method: Co-citation and bibliographic coupling
Cost: An individual account is free for up to five analyses per month or a single month's access can be purchased for a small fee.
Key function: generates citation maps based on other works which cite a given source paper. Our browser plug-in, LibKey, is integrated with Litmaps.
Data source: Crossref, Semantic Scholar and OpenAlex.
Method: You can choose an algorithm for your search. "Shared citations and references" and "common authors" use co-citation and bibliographic coupling whereas "similar text" uses GenAI to analyse semantic connections within titles and abstracts.
Cost: An individual account is free with limited searching or a single month's access can be purchased at a student rate.
Key function: takes a single paper or group of papers (e.g. a reference list from EndNote, Mendeley, or Zotero) and generates recommendations. It shows a graph representing citations between the source paper and the recommendations, identifying the most 'connected'.
Data source: OpenAlex and Semantic Scholar
Method: The recommendation engine is by "citation networks as well as some additional magic!", therefore not transparent. The search engine is by PubMed and Semantic Scholar.
Cost: Free for individuals.
You may also wish to try using an AI discovery tool as part of your supplementary searches. GenAI-powered discovery tools use the power of Large Language Models to find connections between academic articles and may help you find previously undiscovered papers. The University of Bristol does not recommend or hold a subscription to any of the following products, but you may find them useful supplementary tools for your research.
You can learn more about the role of AI in information finding via our guide to AI and the Library. Before using any AI tools, please follow guidance from your school and the University's academic integrity guidelines.
Key function: A search engine to find academic articles. It has additional features such as 'Upload Papers' where you can chat with a PDF, 'Consensus Meter' which "lets you quickly see scientific consensus and gain contextual insights", and 'Pro Analysis' which uses ChatGPT-type functionality to draft and format work.
Data source: Semantic Scholar.
Method: Large language models (LLMs) and vector searching. Analyses a question or a phrase with OpenAI’s GPT-4 model to generate a one-sentence summary of "the top 10 studies", with links to these articles.
Cost: Limited access is available to individuals for free together with the option of a less restricted account for a monthly fee.
Key function: A search engine to find academic articles. Additional features include adding columns to the results table (for further analysing), searching citation networks, uploading and chatting with PDF's and generating research reports.
Data source: Semantic Scholar
Method: Analyses a question or a phrase via LLM-powered semantic searching to generate a summary of the 8 most relevant articles and a table containing the papers.
Cost: Limited access is available to individuals for free together with the option of a less restricted account for a monthly fee.
Key function: Generates a list of related articles based on an uploaded PDF. You can search the citation networks and results can be exported to reference management software.
Data source: OpenAlex.
Method: Analyses the content of a document using large language models and then generates a list of relevant articles.
Cost: Limited access is available to individuals for free together with the option of a less restricted account for a monthly fee.
Key function: A search engine to find academic articles. It has additional features such as various summaries (e.g. TL;DR, research methods, limitations and conclusions), Zotero plugin, chat with uploaded PDF and 'My Notebooks'.
Data source: Semantic Scholar, OpenAlex, Google Scholar, and "other trusted repositories".
Method: Analyses a question or a phrase via LLM-powered semantic searching to generate a summary of the top 5 most relevant articles and links to the papers.
Cost: Limited access is available to individuals for free together with the option of a less restricted account for a monthly fee.
Key function: A search engine to find academic articles. It has additional features such as scite Assistant (ChatGPT to answer questions and analyse uploaded PDF's), and 'Dashboards' where you can track trends and a Zotero plugin.
Data source: "We have a dataset of over 1.2 billion Citation Statements and 187 million full-text articles we access through indexing agreements with publishers and Open Access content."
Method: Analyses a keyword/DOI/article title (via Search Citation Statements) or a question/phrase (via scite Assistant) using LLM-powered semantic searching of its “smart citation index”. It categorises citations based on context and highlights supporting or contrasting evidence for a referenced work.
Cost: 7-day free trial, then a monthly fee. No free option.
Key function: A search engine to find academic articles. Additional features include TL:DR summaries, integration with Zotero and 'Semantic Reader' for selected papers (an augmented reader which is more accessible and contextual).
Data source: Web indexing and partnerships with scientific journals, indexes, and content providers.
Method: Uses machine learning techniques to analyse semantic connections between research articles, generating results of relevant articles based on the search query.
Cost: Free.
The University of Bristol suggests using:
If you log into Microsoft Copilot using your university email you will see it has Protected status meaning that your chat history and data are not used to train AIs. Logging in also increases the number of possible responses from Copilot.
Alternative and free LLMs include:
Next step: Writing up your review
Photo by Hanna Balan on Unsplash
Learn how developing your library and research skills can strengthen your AI literacy and explore the opportunities and limitations of AI research tools.
