Check draft quotations against the supplied source pages
Compare exact quoted words and locators with the provided authority copies.
Tools for the documented steps
Practice relevance
Cross-practice idea · Useful supporting administrative method; no claim of practice-specific legal instructions.
Scope: Acrobat source-page comparison; Thomson Reuters litigation source is US-context, not automatic AU authority validation.
The problem
Compare exact quoted words and locators with the provided authority copies.
Why it matters: A quotation can be accurate but cite the wrong page, or silently change a source’s words. A literal comparison helps the lawyer find those defects without deciding whether the authority is current or supports the legal argument.
Sources & evidence
Proposed / not tested. The public source describes the problem. Documentation supports only the stated method components; no result from this recipe has been verified.
What’s been verified
Source methods and reported problems have been reviewed. We have not run the adapted workflow. The sample below describes expected behavior on synthetic data, not a customer result.
What to measure: In a synthetic pilot, count correctly found exceptions, false alarms, and unresolved inputs. No time or cost saving is assumed.
- Pain evidence: Cite Checking is THE WORST ↗Direct pain, older; proposed narrow adaptation. Cite Checking is THE WORST, 2 April 2025, by SaltyMarg4856, describes cite-checking briefs and motions as tedious. The OP's follow-up specifically describes checking every quote, including omitted punctuation and changed emphasis. Comment text was verified in the opened parent thread; direct comment fetch failed. The thread's absolute date was indexed and corroborated by its dated Reddit archive link. A self-described ten-year litigation paralegal, 29 February 2024, separately reports being assigned cite checking without prior experience. More recent corroboration: Cite-check hacks, 7 August 2026, Apprehensive-Gate207, asks how to handle a 300-page brief accurately and efficiently. That date is indexed; the opened page displayed a relative timestamp. Role on the latter account is inferred from its work description and forum.
- Pain evidence: follow-up ↗Additional self-reported public account; no independent identity or outcome verification.
- Pain evidence: self-described ten-year litigation paralegal ↗Additional self-reported public account; no independent identity or outcome verification.
- Pain evidence: Cite-check hacks ↗Additional self-reported public account; no independent identity or outcome verification.
- Method: Thomson Reuters Drafting Assistant for litigation ↗Thomson Reuters Drafting Assistant for litigation documents citation identification, authority links and an Authority Compiler that appends cited authority. These can supply a lawyer-verified source set if already licensed; they do not validate every quotation or establish that a proposition is legally supported. Python difflib documents text comparison and difference reports, a separate deterministic component for literal quote comparison. The recipe can also be piloted manually without installing anything.
- Method: Python difflib ↗Additional official capability reference. The complete adapted recipe has not been tested.
Source checked: October 15, 2026. Pain accounts are self-reported. Documentation supports components; this adapted recipe and prompt have not been tested. We wrote these prompts. The adapted workflow hasn’t been tested.
Suggested approach
Start when: A frozen draft and verified cited-source set are ready
- match each supplied citation occurrence to its supplied source/page and compare literal quote text
- Keep source references and report unknowns instead of filling gaps.
- The responsible reviewer checks the report before any separately approved action.
Limits and review
The assistant drafts or reports only. Check source records and review the output before a human performs any separately approved send or change. No automatic legal decision, money movement, account connection or publication is included.
Who reviews it: Lawyer verifies the originals, citation form, quote alterations, factual/legal context and current authority status separately. Every apparent exact match remains subject to review; OCR errors, italicized emphasis and omitted context can defeat plain-text checks.
What you need
Your firm-approved AI assistant · prepared files
- Frozen draft, verified cited-source set and an authorized reviewer; the reviewer supplies source identity and canonical reference, not AI web replacement research.
- Existing firm-approved Acrobat Studio, AI Assistant add-on, or another plan including Acrobat AI; online services/cloud storage, language and region must be permitted. No purchase, login or connection is part of this verification. The manual route works without AI.
- Preserve original evidence unchanged in restricted storage. Use synthetic material first. Prepare a separate, explicitly approved minimized AI input; aliases alone do not anonymize narrative or attachments. Do not upload raw originals or an alias key merely because another tab/file is sanitized.
- Use synthetic inputs first. Keep real data in firm-approved accounts and processing, with appropriate access, confidentiality and retention controls. Aliases alone do not anonymize records. Never paste credentials, API keys or unnecessary client details.
How to use
Documentation-verified component path; proposed adaptation, not tested
- Literal concordance is not authority validation, legal support, citation-rule compliance or good-law status. Extracted text can lose emphasis, punctuation and surrounding context.
Check access and setup
- Actual account access, source versions/completeness, privacy approval and reviewer sign-off were not observed.
Freeze the approved draft for quote inventory
Word/Acrobat · Open the actual frozen draft DOCX/PDF
Preserve the original and record draft version and in-scope sections. Obtain explicit approval for the bounded draft text that can enter Adobe processing, using a synthetic/public draft first. If only excerpts are approved, record their scope and do not promise a whole-brief check.
- Expected output
- Approved frozen draft input and scope manifest.
- Next step
- Load this actual approved draft into PDF Spaces for the first quote-inventory pass.
- Check before continuing
- The first AI pass can propose inventory; a person must still check coverage against the whole intended draft.
Load only the approved AI input
Acrobat PDF Spaces · Acrobat homepage > PDF Spaces > Create a PDF Space > upload files or Paste copied text > Add to PDF Space
Add the approved frozen draft or explicitly bounded draft excerpt set and scope manifest. Use clear source/version labels and verify the Files panel. Do not add originals that have not separately been cleared for Adobe processing.
- Expected output
- Visible approved source set in the PDF Space; no account integration is implied.
- Next step
- Use the AI Assistant message field in this same space for the next prompt.
- Check before continuing
- Confirm every named source is actually loaded and readable. PDF Spaces supports up to 100 files, 100 MB and 600 pages per file; unsupported/password-protected/handwritten material blocks affected extraction.
Run the bounded prompt and inspect citations
Acrobat AI Assistant · The PDF Space AI Assistant message field
Paste the prompt below. Open each source-number citation, check it, and use Copy to transfer the answer into a new local Word document or text file. Correct errors using the originals before reliance.
- Expected output
- Proposed quote inventory to check before source acquisition.
- Next step
- Save the checked draft as quote-inventory-review.docx; carry source IDs, versions and reviewer corrections into the next review step.
- Check before continuing
- Human reads the intended draft scope line by line, corrects omissions/splits/quote boundaries and confirms every citation ID; unsupported inventory completeness is not accepted.
Acquire the actual cited source pages after inventory review
Word and Acrobat · Open frozen draft; open supplied source PDFs > page navigation and Select text
Using the preceding reviewed inventory, open each actual supplied cited source. Visually map its printed page label to PDF page index, and copy/check the cited page text. Retain the corrected draft citation IDs, exact quote text and editorial-change/emphasis flags. Confirm source identity and canonical reference from the lawyer-approved catalog. Missing sources or locators remain null/unresolved; do not obtain replacement authority from AI.
- Expected output
- citation inventory plus reviewer-verified source catalog and printed-page/PDF-index map.
- Next step
- Make a minimized citation-check input, optionally using the supplied JSON kit; upload/paste that actual reviewed text rather than assuming a local JSON is visible.
- Check before continuing
- Do not use a document filename or an AI citation to establish source identity; never map printed page numbers by arithmetic alone.
Load only the approved AI input
Acrobat PDF Spaces · Acrobat homepage > PDF Spaces > Create a PDF Space > upload files or Paste copied text > Add to PDF Space
Add the checked citation inventory, approved source-page text and explicit page map. Use clear source/version labels and verify the Files panel. Do not add originals that have not separately been cleared for Adobe processing.
- Expected output
- Visible approved source set in the PDF Space; no account integration is implied.
- Next step
- Use the AI Assistant message field in this same space for the next prompt.
- Check before continuing
- Confirm every named source is actually loaded and readable. PDF Spaces supports up to 100 files, 100 MB and 600 pages per file; unsupported/password-protected/handwritten material blocks affected extraction.
Run the bounded prompt and inspect citations
Acrobat AI Assistant · The PDF Space AI Assistant message field
Paste the prompt below. Open each source-number citation, check it, and use Copy to transfer the answer into a new local Word document or text file. Correct errors using the originals before reliance.
- Expected output
- Literal quote/source concordance exceptions and counts of checked/unresolved/manual-review rows.
- Next step
- Save the checked draft as quote-concordance-review.docx; carry source IDs, versions and reviewer corrections into the next review step.
- Check before continuing
- Check every apparent match visually for emphasis, punctuation and context; missing source/page evidence must remain unresolved.
Human review and final handoff
Final human reviewResponsible staff reviewer and lawyer/process owner · Local reviewed report beside the frozen source set
Lawyer checks original pages, canonical reference, citation form, quote alterations/context and authority status separately. Any draft change invalidates the old comparison for affected citations.
- Expected output
- Reviewed clerical concordance report tied to frozen draft/source versions.
- Next step
- No automatic correction, replacement authority or filing submission.
- Check before continuing
- All unresolved or unreadable evidence stays open. No inferred approval or external action follows from AI output.
What you’ll get
citation-check.csv: draft_version, citation_id, draft_locator, source_id, cited_page_label, observed_page_label, status, reference_difference, evidence, reviewer_action; coverage counts for checked, unresolved and manual-review rows. Derive every ID and finding from supplied input.
Inputs and a sample test Synthetic example · not a measured result
Input files
- One JSON object with these fields, no additional fields required: - `draft_version`: nonempty string - `citations`: array of `{citation_id: string, draft_locator: string, source_id: string, cited_reference: string, cited_page_label: string, quoted_text: string, has_editorial_changes: boolean}` - `sources`: array of `{source_id: string, canonical_reference: string, verified_source_url: string, reviewer_verified: boolean}`; URL may be empty for a synthetic fixture - `source_pages`: array of `{source_id: string, page_label: string, pdf_page_1based: positive integer, text: string, text_visually_verified: boolean}` Keys are case-sensitive. `(source_id,page_label)` must be unique. Do not convert printed page labels into PDF page indexes. Duplicate/missing keys stop that row. Empty quote means citation-only and remains a lawyer-review item. Normalization permits only CRLF-to-LF and runs of whitespace to one space; preserve punctuation, capitalization and every number.
Synthetic example
Synthetic inputs are included in the starter kit. citation_check.example.json
Expected output
Q1=quote_mismatch: page label 10/PDF page 2 says green, not blue. Q2=found_on_other_page: label 11/PDF page 3. Q3=unresolved because S2 is absent. Three rows accounted for; two compared and one unresolved. No replacement authority or legal-support judgment.
Test a failure
Lawyer verifies the originals, citation form, quote alterations, factual/legal context and current authority status separately. Every apparent exact match remains subject to review; OCR errors, italicized emphasis and omitted context can defeat plain-text checks.
Try with the sample files
Follow the tool steps above and use the inputs named in each step. Where provided, native-* files are blank scaffolds and NATIVE-INPUT-MAPPING explains their fields. Original examples and the optional full review prompt use a separate prepared-input schema. Keep them separate unless the mapping explicitly connects them. Downloading a kit does not connect accounts or prove the workflow works.
Download inputs and checklist 6 files
Instructions, blank templates and separate synthetic examples; these files do not install an automation. Blank templates still need verified source data.
- Download starter kit All inputs, synthetic examples and review notes in one ZIP. No automation is installed.
- Start here: get the sources and prepare the review Documentation-checked acquisition, explicit manual mapping, exact prompts and human review limits
- citation_check blank input Preserve nested input structure and validate supplied field types
- citation_check synthetic fixture Preserve nested input structure and validate supplied field types
- citation_check JSON Schema Preserve nested input structure and validate supplied field types
- Setup and fixture review guide Recipe-specific setup, input rules, offline prompt, expected result and boundaries
Optional full review prompt for prepared inputs
Prepared-input review prompt with Your firm-approved AI assistant · prepared files
Proposed / not tested · editorial prompt. This prompt uses the original prepared-input schema shown in the sample test. Step prompts may use different tables. Follow NATIVE-INPUT-MAPPING where supplied; do not mix the original examples with native-step inputs without mapping and checking every field.