Evidence first. Certainty only where earned.

1. Build context from the manuscript.

The analysis organizes the manuscript by chapter and draws on your intake: genre, point of view, character declarations, concerns, motifs, and what you want preserved. Clear headings and accurate intake details help the review.

2. Locate the quoted evidence.

Software checks whether cited passages can be found in the submitted version. Evidence that cannot be verified should not be presented as a confirmed source quotation. This protects the connection between a finding and your text; it does not prove the finding’s interpretation.

3. Review the claim in context.

Judgment-based candidates use AI review. The review considers whether the cited evidence supports the claim and whether nearby context or a plausible alternative explanation defeats it. Impact and confidence are recorded separately: uncertainty does not make an issue intrinsically less consequential.

4. Disclose the coverage.

Some checks are mechanical; others depend on inference, available context, and review limits. The report distinguishes reviewed findings from unresolved or unadjudicated candidates. It should also identify checks that were unavailable or incomplete.

5. Check the assembled report.

A human operator performs quality control on the assembled deliverables. This is not a separate human editorial judgment on every finding or a substitute for a professional editor’s full read.

What the public example establishes

The sample report shows how an evidence-based finding can be presented using invented test entries. It is an illustration, not generated output or a benchmark result.

A meaningful performance claim needs the actual test manuscript, an answer key, the evaluated release and settings, the complete output, and a reproducible scoring method. We are not presenting a comparative accuracy percentage or an independently validated editorial-quality claim here.

What remains your decision

Whether a repetition is purposeful, a passage earns its space, or an apparent inconsistency serves the story can require an author’s judgment. Review the passages, challenge interpretations, and use the feedback that helps your book.