🎯
Relevance scoring
Every imported study is scored against your research question and PICO criteria, so the most relevant papers surface first instead of being buried in an alphabetical list.
AI LITERATURE SCREENING
AI literature screening that assists, never decides alone.
Every study gets an AI relevance score against your research question — reviewers see it, but the include/exclude call always stays human, with blind dual review and automatic disagreement detection for teams of any size.
2
Screening Stages
Configurable
Required reviewers
3
Resolution methods
How the AI helps
Automated literature screening, with a human always at the wheel
🔍
Title & abstract screening
A fast first pass with the AI score visible alongside title and abstract — reviewers move through the queue with the highest-signal information already in front of them.
📄
Full-text screening
PDFs are auto-retrieved where available; apply your inclusion/exclusion criteria at full text with the same conflict-tracking as the first stage.
🙈
Blind independent review
Reviewers can't see each other's decisions on a study until they've submitted their own — configurable per stage, so bias from seeing a colleague's call first isn't possible.
⚡
Automatic disagreement detection
Any study where reviewers land on different decisions is flagged the moment the required number of reviewers has weighed in — nothing slips through unnoticed.
🤝
Configurable conflict resolution
Choose whether disagreements are settled by a third independent reviewer, a review manager, or consensus among the original reviewers — set once, applied consistently.
Screening software that shows its work
A relevance score with no explanation is a black box reviewers learn to ignore. EvidenceFlow's screening queue keeps the AI's role visible and advisory — every final decision, and who made it, stays in a full audit trail.
See the full systematic review workflow →FAQ
Frequently asked questions
Does the AI make screening decisions automatically?+
No — it scores relevance to help reviewers prioritize, but every include/exclude/maybe decision is made by a human reviewer.
What is blind screening?+
When enabled for a stage, a reviewer can't see any other reviewer's decision on a study until they've submitted their own decision for it — preventing anchoring bias between reviewers.
How are screening conflicts resolved?+
Configurable per project: a third reviewer who hasn't already screened that study, a designated review manager, or consensus among the original reviewers can record the resolving decision.
Can I see why the AI scored a study the way it did?+
The score reflects similarity to your stated research question and PICO criteria — it's a prioritization signal, not a final judgment, and reviewers always see the title and abstract alongside it.
Is there a limit on reviewers per project?+
No — invite as many reviewers as your team needs, by email, whether or not they have an existing account yet.
Try AI-assisted screening free
Import a batch of references and see relevance scoring, blind review, and conflict detection working together on your very first screening pass.