A footnote can make a conclusion look supported even when the cited material establishes something narrower. Gemini can help organize a footnote claimcheck, but the reviewer must inspect the cited source and the exact conclusion. This workflow audits the relationship between document claims and their footnotes rather than checking citation formatting alone.
Primary product documentation checked on 8 October 2026. The worked examples are fictional editorial examples, not claims of hands-on testing inside Gemini or Perplexity.
Extract claims and their footnote links
List consequential conclusions in the document and identify the footnotes attached to each. Capture the full claim, not only the sentence fragment near the marker. A paragraph may combine several assertions while citing one source at the end. Split those assertions into separate review items when they need different evidence.
In a fictional strategy document, a paragraph claims that a tool reduces costs, improves accuracy, and is widely adopted. Its footnote links to a vendor announcement describing a new feature. The link may support feature availability while offering no evidence for the other claims. The audit should reveal that mismatch instead of treating the presence of a footnote as sufficient.
Open and characterize the original source
Record source title, publisher, publication date, evidence type, and the relevant passage or table location. Distinguish vendor statements, research findings, opinion, and secondary reporting. Ask Gemini to help summarize selected source text, but verify the surrounding context directly. Do not rely on a search snippet or an AI-generated citation description as the source itself.
Check whether the cited page has changed since the document was written. A current page may omit a detail that appeared in an earlier version. Use a preserved authorized copy or dated archive where available, and label its role. An inaccessible source is an evidence gap, not automatic proof that the claim is false.
Compare the claim with the evidence scope
Inspect population, setting, measurement, time period, and degree of certainty. A laboratory result may not support an operational guarantee; a pilot may not support an industry-wide conclusion. Ask Gemini to propose a support assessment for each claim, then review that assessment against the original evidence.
Use clear outcomes: supported as written, supported only with qualification, unsupported by this source, contradicted, or unable to verify. A qualified outcome should state the missing qualification. For the fictional cost claim, the source may report savings for one customer under specific conditions. The document should not convert that into a general promise for every organization.
Draft narrow corrections and evidence requests
When the source supports a narrower conclusion, write a proposed revision that preserves its boundaries. When evidence is absent, request the specific kind of source needed or remove the unsupported claim through the document’s review process. Do not ask Gemini to find a convenient citation after deciding that the conclusion must remain unchanged. That approach encourages selective evidence gathering.
Keep the original claim, assessment, proposed wording, and approval state in the audit register. A correction might change widely adopted to available to eligible users if that is what the source actually establishes. Explain why the wording changed so reviewers can judge the substance rather than seeing a mysterious editorial edit.
Retest citations after revision
Check that edited claims still connect to the correct footnotes and that numbering or cross-references survived document changes. Inspect repeated conclusions in the executive summary and conclusion section. A corrected body paragraph does not fix an unsupported headline elsewhere. Use claim identifiers to trace the change through the document.
Then review a sample without seeing the audit outcome first. Read the source and decide independently what it supports. Compare that judgment with the register. Disagreement reveals where the assessment needs clearer reasoning. Save the reviewed document and audit together so future writers can reuse the evidence without repeating the original overstatement.
Validate the evidence and hand over the result
Run the review in two passes. First, check the source records without reading the generated conclusions. Select a small but representative sample: a clear record, a contradictory record, an incomplete record, and a record that changed after the review began. Write down the expected treatment of each. This gives you a reference for judging the output instead of accepting a plausible narrative because it looks organized.
Then inspect the generated result against those expectations. Open every source supporting a consequential conclusion. Check the surrounding paragraph, the relevant table heading, and any qualifying note. A link to the correct document is only the beginning of verification. The document may support a narrower claim, describe a different version, or use a different unit. Record a correction beside the affected item rather than quietly changing its meaning.
Count the inputs and outputs where a one-to-one relationship is expected. Explain every omitted record and every merged record. When several source items become one conclusion, retain the individual references so the reviewer can reverse that grouping. When one source item produces several actions, preserve the connection so a later source correction can update all affected actions. Never interpret an empty output as evidence that there are no problems until you have checked source coverage.
Keep proposed wording separate from approved wording. An AI system can suggest how to express a finding, but the source owner still needs to confirm the underlying facts when records disagree. Use explicit states such as supported, disputed, missing evidence, and awaiting owner confirmation. Avoid a single confidence score that disguises those different reasons for uncertainty. A disputed statement with two strong sources needs a different next step from a statement with no accessible source.
Before handover, ask a colleague to use the result without your explanation. Can they locate the original evidence, identify unresolved items, and understand what changed since the last version? Fix the places where they need to guess. Save the source manifest, the prompt, the reviewed result, and the correction history together. A useful deliverable should remain understandable after the conversation that produced it has ended, and it should make the next review cheaper because the evidence trail is still intact.
Frequently asked questions
Does a working source link prove support?
No. The source must substantiate the particular claim and its scope.
Can one footnote support several claims?
Yes, but each assertion must be checked individually against the evidence.
What if the source is inaccessible?
Mark the claim unable to verify and identify the missing evidence. Do not assume support or falsity.
Should unsupported claims receive new citations automatically?
First review the conclusion itself and define the evidence needed. Avoid searching only for confirmation.
What should be checked after revision?
Retest footnote links, numbering, repeated claims, and the relationship between the new wording and its evidence.
