Just started using Elicit for a literature review on CRM adoption in SMBs. I have a very straightforward query, something like "CRM adoption challenges small business," but it keeps returning zero papers.
I know for a fact there are relevant papers out there. I've found a couple manually. Is there a specific syntax I should be using? I'm worried I'm misunderstanding how the search works. Is it looking at full text, or just titles/abstracts? I'm coming from more traditional academic databases, so maybe my expectations are off.
That's a common point of friction when transitioning from traditional databases. Elicit primarily works with its own curated corpus and looks at the full text of papers it has processed, not just abstracts. Your query might be a bit too specific.
Try breaking it into broader components. Search for "CRM adoption" first, then use the "concepts" column or the follow-up questions feature to drill down into the small business aspect. You can also paste the DOI or title of a paper you found manually into the search, which often helps surface similar work.
The syntax is less about operators and more about conceptual matching. It's not like PubMed or Scopus.
Stay curious, stay critical.
Elicit's search is fundamentally different from traditional databases. While user793 is correct about the full-text indexing, the zero-result issue is often due to the underlying semantic matching failing to map your natural language query to the vector space representation of its corpus. It's not about syntax operators but about the density and specificity of the concept clusters in the training data.
Your query "CRM adoption challenges small business" might be creating an overly narrow semantic vector that sits in a sparse region. Try decomposing it into more atomic concepts that the model has likely encoded with stronger relationships. For instance, search for "customer relationship management software implementation" to capture the broader domain, then filter or ask follow-up questions about "SME" or "barriers." The platform often performs better with standard academic terminology over colloquial phrases like "challenges."
You can verify coverage by pasting the DOI of a known relevant paper into Elicit's search. If it returns that paper, the corpus contains it and your original query is a retrieval problem. If it doesn't return the paper, then the corpus itself may not include it, which is a separate limitation. This diagnostic step clarifies whether the issue is with your query formulation or with the available data.