I used one of my own papers (co-authored with Dr. Anna Walter ) to test which AI tools are currently best suited for literature research in the context of scientific work.
How I proceeded: there is only one research question in our paper. I copied it and used it as input in the tools.
Why did I do this? I wanted to test whether the tools could find our paper, which was published as open access by Wiley.
Here are my current findings (as of September 16, 2025):
My clear favorites are still Consensus and Elicit.
Why?
Consensus has achieved an absolute highlight with its new Deep Search feature. In just a few seconds, I not only get an overview of the research field, but Consensus also prepares a real short paper for me, with an introduction, methods, search strategy, key results, discussion, research gaps, and open research questions. For now, the free version allows for three comprehensive Deep Searches based on 50 papers, but 25 shorter Smart Searches based on 20 papers.
Elicit provides me with a quick overview of truly relevant articles. In the free version, it generates a clear and understandable answer to my research question in seconds, based on 4 papers. In the $12/month Plus version, it even covers 8 papers and allows you to extract data from 50 papers in tabular form – be it summaries, research designs, results, or limitations. The research reports are an absolute highlight. Whether they can replace a comprehensive systematic literature review will certainly be discussed more frequently in the scientific community in the coming months.