Introduction: As part of a comparative analysis on automation and AI tools for information retrieval by Canada’s Drug Agency, we conducted a focused literature review on how three tools (SpiderCite, Lens.org, and Microsoft Copilot) had been used in evidence synthesis. The literature review required adapting our standard search approach to capture reports based on information retrieval methods reported in the full-text. Methods: We designed strategies in PubMed Central (PMC), Google Scholar, and Europe PMC to target uses of the three tools in evidence reviews. Given functionality differences with these three databases, as well as specific challenges related to the search terms for the Lens.org and Copilot concepts, we adopted different approaches for each search. Results: In PMC and Europe PMC, we used search functions to capture keywords within the methods sections of review articles. We also used PMC’s search field for the body of the text, not including references or abstracts. In-depth searching of Google Scholar confirmed that syntax does not function correctly when used according to Google guidance, and that metadata of search results is messy and requires significant reformatting. However, tools such as Publish or Perish significantly eased the workload of working with Google Scholar. Discussion: The approaches used in this case to search non-standard fields in full-text databases and to utilize a records management tool with Google Scholar significantly focused results and reduced the screening burden for methods research. We will employ the approaches established from the literature review from this case for future methodological research at Canada’s Drug Agency.