A reviewer comes back with Major Revisions. Buried on page four of the report: “The authors appear unaware of [Author et al., 2023], which addresses this exact problem and should be discussed.”
You’ve never heard of that paper.
This flavor of rejection isn’t about sloppiness. The gap isn’t a paper you forgot to read. It’s a paper you had no way to find, because literature search is bounded by the vocabulary you already use. A machine learning researcher searching for “distributional shift” won’t surface a psychometrics paper on “test-retest reliability across populations,” even if the underlying problem is structurally identical. Neither author is searching the other’s terminology. The overlap exists; the search doesn’t reach it.
There’s also a timing problem. A paper published while your manuscript was under revision is recent enough to seem like expected reading but early enough to have been indexed and found by reviewers. The gap between your last literature update and submission can be six months.
Both produce the same outcome: a reviewer citing something you’ve never seen, in a field you follow closely, on a problem you’ve spent years studying. The work isn’t wrong. You missed the context.
All major databases return papers matching your query terms. You search for what you know to search for. Adjacent-field papers using different terminology for the same concept don’t appear. You can extend the search by following citation trails or asking colleagues, but that doesn’t resolve the central problem: you don’t know which adjacent fields to check.
Citation Gap reads your manuscript and identifies what the paper is actually claiming as a contribution, then searches for prior work that overlaps with those claims semantically rather than by keyword. A paper on a structurally similar problem in a neighboring field surfaces even if it uses different terminology. The scan also picks up recent publications that a literature review done months ago would have missed.
The report flags each potential gap with a relevance rationale so you can make the call yourself. Some findings will be genuine oversights. Others will be papers you know and decided not to cite, or work that’s adjacent but distinct enough to leave out. That’s the point: you see them before a reviewer does.
The cost of missing a relevant paper varies. Sometimes a reviewer reads it as a sign you don’t know the field and weights the submission accordingly. Sometimes it means adding a paragraph to related work. Occasionally it raises a question about novelty. Finding out which one before submission is worth knowing.