AI Capabilities

AI that assists editors — not a chatbot, not hype

We built AI for specific editorial tasks. Here is exactly what is live, what is coming, and what we will never automate.

Live today

Submission screening and duplicate detection

Every submission is screened automatically on arrival. The editor sees a 0–100 quality score, a recommendation, and a set of specific concern flags — generated by Claude Haiku against the title, abstract, and keywords. Duplicate detection surfaces similar manuscripts already in the journal before the editor assigns reviewers.

  • Quality score 0–100 on every submission (80–100 = accept range, 50–79 = revise, 0–49 = reject)
  • Three-way recommendation: ACCEPT / REVISE / REJECT
  • Concern flags: Plagiarism Risk, Scope Mismatch, Weak Methodology, Unclear Contribution, Ethical Concern
  • Duplicate detection: semantic match against existing journal submissions with similarity scores
  • AI screening can be re-run after an author revises — new score and flags generated
  • Powered by Claude Haiku via the Anthropic API — model version recorded per screening

Screening output fields

Score (0–100)

Calibrated to editorial acceptance ranges

Recommendation

ACCEPT / REVISE / REJECT

Summary

2–3 sentence plain-English evaluation

Flags array

Specific concern types with descriptions

What AI does not do

Editorial judgment stays with the editor

The screening output is advisory. No submission is automatically rejected or accepted by the system. No decision letter is sent without editor review and explicit action. Manuscript data is never used to train models.

  • No automated accept or reject decisions
  • No automated reviewer invitations based on AI output alone
  • No author communication without editor action
  • Manuscript content not used to train or fine-tune any model
  • AI output clearly labelled as advisory in the editor interface
  • Editors can override, dismiss, or note disagreement with any flag

Coming soon

Reviewer matching

Suggest reviewers from member database by expertise

Decision letter drafting

Structured draft for editor review — never auto-sent

Frequently asked questions

See this in action in 20 minutes

Book a demo and we'll walk through this feature with your journal in mind.

Start free