Hey everyone. I’ve been digging into research tools lately, partly because our RevOps team is always looking for better ways to systematize literature for market analysis, and partly out of personal curiosity. I know Elicit is a big name here, but for something as structured and exhaustive as a systematic review in psychology, I'm wondering what the full toolkit looks like.
From what I've gathered, tools like Rayyan are almost a staple for the screening phase, handling those thousands of abstracts with AI-powered deduplication and blind reviewing. But the actual data extraction and synthesis is where it gets tricky. I've seen folks use Covidence alongside Rayyan, and then maybe something like ASReview for the initial title/abstract screening to prioritize relevant papers.
My question is about the end-to-end workflow. If Elicit is great at finding and summarizing concepts from PDFs, how does it plug into the more rigorous, PRISMA-guided stages of a systematic review? Are researchers using Elicit alongside these other tools, or is there a shift towards more all-in-one platforms? I'm particularly curious about handling grey literature and data quality—psychology reviews often need to pull from diverse sources, not just journal databases.
Also, any experiences with tools that handle bias assessment well? That's a huge part of the psychology field, and automating even parts of that risk-of-bias evaluation would be a game-changer.
You've identified the core tension really well. Elicit and tools like ASReview function brilliantly for discovery and rapid prioritization, which maps perfectly to that initial, overwhelming search phase. However, they aren't designed to enforce the methodological rigor required for the subsequent stages of a true systematic review.
The integration point is usually manual and happens at the import/export stage. Researchers will use Elicit to generate a broad set of potentially relevant papers, then export that list to a tool like Covidence or Rayyan for the formal, audited screening process. Covidence, in particular, becomes the system of record because it rigidly structures the PRISMA flow, dual-reviewer conflicts, and the data extraction forms. Elicit doesn't replace that; it feeds into it.
Regarding all-in-one platforms, I haven't seen one that masters both the generative AI aspect and the strict compliance tracking. Most teams I've observed are using a best-of-breed stack: an AI discovery tool, a dedicated screening platform, and then separate software like JBI Sumari or even Excel for synthesis. Grey literature remains a manual challenge, as those sources often bypass the automated discovery tools entirely.
Method over hype