A systematic review (SR) is a well-established research method, first used in the medical field and now applied in many other fields. Over the last couple of years, we have received many questions about Keenious and its application in SRs.
Although Keenious is not specifically designed for SRs, it can still be useful at different stages in the process. This article explains briefly how Keenious can be applied in SRs and includes a step-by-step guide for using Keenious during manual searching.
Before jumping into the nitty gritty we need to clarify a few things about AI’s role in a review. Broadly speaking, AI tools for SRs fall into two main categories:
The first category involves fully automated tools, which seek to replace traditional systematic review methods by automating the entire process. The second category is supplemental tools, which support the existing SR workflow by assisting researchers in particular tasks.
Keenious falls into this second, supplemental category, as it does not aim to replace current practices. Instead, it provides something extra to complement the steps you already take in a systematic review.
Where Keenious Can Be Useful
There are many examples of where Keenious plays a helpful, supplementary role at several points in the process. Through workshops and discussions with multiple universities and librarians, as well as examining publications where authors have described using Keenious for systematic reviews, we have identified stages in the process where Keenious can be useful.
Below is a table outlining the systematic review stages (adapted from Tawfik et al., 2019) and where Keenious can contribute.
Keenious can help with discovery early on (in Stages 1-4) by finding additional relevant articles during manual searching (Stage 9) and by supporting the writing process in stages such as protocol writing (Stage 6) and manuscript writing (Stage 14). To illustrate how Keenious can be used for manual searching, we’ve prepared a practical step-by-step guide based on how it was used in a published paper.
Step-by-Step Guide for Manual Searching
We would like to highlight a real-life example of manual searching. In a protocol published in the BMC journal Systematic Reviews in early 2025, the authors describe using Keenious to find additional relevant articles that may have been missed by traditional database searches. As they write:
“... we leverage artificial intelligence in our research, specifically using Keenious Plus by transferring relevant articles into its platform to identify matching articles that are relevant to our research, helping us uncover any potentially missed articles.”
After you have screened (Stage 8) and selected a set of included articles, use the following procedure to manually search for additional literature for each of those articles:
1. Add the article to Keenious
1. Add the article to Keenious
Go to keenious.com and add the article as a PDF. Make sure you are logged in before proceeding.
2. Choose numbers of results
2. Choose numbers of results
Once the results are visible, decide how many articles you want to screen. At the bottom of the results page, you can select the number of recommended papers to display (10, 25, 50, or 100). Choose the option that best suits your needs.
3. Export the results list
3. Export the results list
Click the three-dot menu on the results page and select “Export result page.”
4. Select the right format
4. Select the right format
Choose the file format for your export. Most reference management tools and systematic review tools support RIS or BibTeX, so select the format that works for your setup.
5. Download file and upload to your library
5. Download file and upload to your library
Download and upload it into the reference management tool (RMT) or systematic review tool (SRT) you are using for your project. Example below from an Endnote library.
6. Repeat the process for each included article
6. Repeat the process for each included article
Follow the same steps for each paper you have included in your review.
7. Remove duplicates
7. Remove duplicates
If your RMT or SRT supports duplicate detection, use this feature to remove any articles that were already identified in your original database searches.
8. Screen the new articles
8. Screen the new articles
Once duplicates are removed, you can proceed with screening the new articles recommended by Keenious, following the same inclusion/exclusion criteria as in your original review (Stage 3).
By following this process, you may identify relevant articles that were not found during your initial database search, which will improve the completeness and quality of your systematic review.
There might be other ways of using Keenious for systematic reviews we are not aware of! We’re curious to hear about your experiences and ideas. Please don’t hesitate to contact us if you’ve used Keenious in a systematic review or have any suggestions.
Additionally, we hosted a webinar on using Keenious in systematic reviews with Farangis Sharifibastan, the first author of the example paper. If you’re interested in the recording, please get in touch for more details.