Your PRISMA flowchart is filled in beautifully, and then a reviewer asks why you excluded a trial that never appeared in any database you searched. The justification is thin, and the review looks less complete than an hour ago. WisPaper, an  AI academic agent, approaches the problem from the search side rather than the reporting side.

The Flowchart Is Not the Search

A PRISMA diagram is a ledger of decisions already made, not an instrument that finds anything. It tells a reader how many records were excluded but has no way of knowing what was never screened at all. WisPaper starts upstream of that diagram, where registries get skimmed and pages of grey literature never opened.

The Trial That Lives Outside the Databases

The trial a reviewer asks about usually lives in a registry or a university repository, not in the databases you queried. Grey literature and trial registries routinely contain the negative results that flip a meta-analysis conclusion, and a checklist doesn't retrieve them for you. One unpublished null result can change a pooled estimate, and no column surfaces that trial.

Screening With a Copilot Instead of a Spreadsheet

Screening is where the spreadsheet stops being a tool and becomes a place where judgment gets logged and then lost. Two reviewers mark the same abstract differently, the reasons blur, and a week later neither can say why a paper moved from maybe to no. Across a few thousand records, you end up with exclusions nobody fully remembers making.

Checking Whether Each Citation Says What You Think

Some of this is becoming automatable. WisPaper's  AI academic agent verifies that each citation resolves to a real, published source and matches the style your journal requires, and flags entries where the reference and the source disagree. Its search reads a query as intent, not a string of terms, which matters when the missing trial uses vocabulary unlike the studies you found.

Defending Your Exclusions Line by Line

When each exclusion carries a retrievable reason tied to a specific source, the reviewer letter writes itself. You stop defending your memory and start pointing at a trail of decisions. The review becomes defensible because every box can be traced backward.

The checklist and the search finally point at the same thing. A systematic review is easier to defend when the records it excluded are visible, not vanished, and every judgment has a source behind it.