Non-Disclosure Agreements

How to compare draft and executed versions for Non-Disclosure Agreements

Compare approved drafts and executed mutual NDAs using a copyable checklist, seven review steps, and a fictional example of execution-stage deviations.

October 19, 2026  ·  10 min read
Book a Demo →

Compare an executed mutual NDA against the last approved draft through clause-level text comparison and visual inspection of the complete signing package. Establish the approved baseline, then examine changes to protected information, permitted use, recipients, exceptions, duration, and return obligations from both parties’ perspectives. Record each difference with document citations, approval evidence, practical consequences, and an accountable reviewer. The deliverable is a completed version comparison checklist separating approved changes, unexplained deviations, and verification gaps, together with a decision on any disclosure restrictions or proposed corrections.

When to use this workflow

Use this workflow when commercial counsel receives a signed NDA, an in-house team discovers competing final copies, or a business team asks whether disclosures can proceed under the returned agreement. It also supports a targeted check before technical, financial, or other sensitive information is shared.

The task is to reconcile the execution package with documented approval, not renegotiate the entire NDA. An unexplained deletion may nevertheless require a disclosure decision or separate corrective document.

This guide addresses US company-to-company mutual NDAs. Rights depend on actual wording, governing law, facts, and transaction context. State-law enforceability and cross-border issues require separate analysis; the federal authorities below support only the identified trade-secret and SEC reporting points.

What you need

Collect the last substantively approved draft, returned execution copy, approval correspondence, and signing instructions. Include referenced schedules, exhibits, purpose descriptions, and incorporated documents. Preserve originals separately from comparison copies.

Also obtain:

  • Approval boundaries: the approving lawyer or delegate, conditions, and permitted administrative changes.
  • Execution evidence: available signing records, signature pages, transmittal messages, and competing counterparts.
  • Disclosure context: what each party expects to disclose, to whom, and for which evaluation or collaboration.
  • Operational constraints: representative access, backup retention, derived analyses, and destruction-certification capabilities.
  • Review instructions: the organization’s own NDA checklist, preferred wording, escalation criteria, and decision-makers.

Do not substitute the latest timestamp for approval evidence. A newer working draft may contain unauthorized changes.

Use the template

Use Non-Disclosure Agreements — Version comparison checklist as the working record. You can copy this table into your working document or spreadsheet.

ID / topic Draft cite / text Executed cite / text Difference Mutual impact Evidence / gap Action / owner Status

Assign stable row identifiers. Citations should include document ID, section, and page where available. Preserve exact excerpts sufficient to demonstrate a change; put longer passages in a working note bearing the same row ID.

“Mutual impact” captures consequences for each party as discloser and recipient. “Evidence / gap” identifies approval correspondence, missing attachments, or unreadable text. Use statuses such as verified unchanged, approved change, unresolved deviation, and verification blocked. Record “not present” only after checking the complete relevant document; missing pages do not establish absence.

Maintain a short document register above the table identifying filenames, document IDs, approval evidence, and package contents. The checklist is your working record, not a promised product module.

Step by step

1. Establish the approved baseline and execution package

Assign document IDs to the approved draft and returned signed copy. Trace approval to the actual file or identifiable attachment. Where approval was conditional—for example, on narrowing the evaluation purpose—confirm that the baseline satisfies the condition.

Inventory pages, schedules, incorporated documents, and signature counterparts. Check whether signature pages appear associated with a different version, without inferring validity or invalidity from appearance alone. Escalate uncertainty about what the parties signed.

Record: a package-verification row identifying the baseline, approval evidence, execution documents, and missing components. If the baseline remains disputed, label the comparison provisional.

2. Compare text and verify the rendered pages

Compare working copies while retaining untouched originals. For scans, treat extracted text as a review aid: inspect page images for dropped negatives, misread numbers, omitted footnotes, and handwriting. Separate layout differences from substantive changes, but investigate formatting that changes meaning, including strikeouts.

Visually inspect headers, footers, schedules, signature blocks, and inserted pages. Check whether comments or unresolved tracked changes appear in either file and determine which text the returned package actually displays.

Record: each substantive difference and verification limitation. Use one row per decision; group repetitive formatting changes only when they raise no separate substantive issue.

3. Reconcile parties, purpose, dates, and temporal coverage

Check legal names and party definitions against signature blocks. Determine whether affiliates are parties, permitted recipients, or neither. An apparent name correction may substitute the entity receiving information.

Compare the purpose and treatment of disclosures preceding signing. Separately check effective date, disclosure window, agreement term, termination mechanics, and surviving restrictions. Do not reduce these to one expiry field. The historical SAP–Sybase NDA, § 18 distinguishes its agreement term from surviving protection and use provisions; this is negotiated wording, not a default rule.

Record: changed dates and temporal qualifiers, consequences for planned or earlier disclosures, and any unresolved effectiveness question.

4. Test confidentiality coverage in both directions

Compare definitions, marking requirements, oral-disclosure treatment, exclusions, use restrictions, and residual-memory provisions. Test changes against each party’s actual information. Deleting express coverage for recipient-created analyses warrants checking the remaining definition, rather than assuming all analyses become unprotected.

Keep contractual coverage separate from statutory trade-secret status. The federal definition requires reasonable secrecy measures and qualifying independent economic value arising from secrecy; an NDA label alone does not establish those elements (18 U.S.C. § 1839(3)).

Compare exclusion-related proof requirements exactly. Check whether a deletion changes who must substantiate prior possession or independent development, without importing a preferred burden into silent language.

Record: affected information, exposure in each direction, approval evidence, and whether the issue requires commercial acceptance, legal interpretation, or both.

5. Compare disclosure permissions and operational duties

Check representative categories, need-to-know conditions, downstream confidentiality duties, and responsibility for representatives. Match changes against intended advisers, contractors, affiliates, and collaboration arrangements.

Review compelled disclosure separately from voluntary protected reporting. The SEC’s Whistleblower Protections guidance explains that Rule 21F-17(a) prohibits impeding individuals’ direct reports to the SEC about possible securities-law violations, including through confidentiality agreements. This protection extends beyond employment relationships. Flag changed approval, notice, or cooperation requirements that could conflict with reporting rights; the guidance is not a universal disclosure exception.

Compare return or destruction triggers, certification, retained-copy exceptions, and continuing restrictions. The SAP–SuccessFactors NDA, § 9 illustrates request-triggered duties and specified retention exceptions, not blanket retention permission.

Record: changed duties, affected teams, feasibility questions, and resolution owners.

6. Reconcile deviations with actual approval

Locate evidence authorizing each exact change. An email saying “looks fine” may be ambiguous where several versions circulated; identify its attachment and conditions.

Prioritize deviations by information sensitivity, imminent disclosure, operational feasibility, and consequences. Keep standstills, IP provisions, remedies, governing law, and dispute provisions visible rather than burying them among formatting changes. Verify definitions and cross-references affected by substitutions.

Choose a proposed disposition: accept as documented, seek clarification, recommend a disclosure pause, or prepare counsel-approved corrective documentation. Approval history and the legal effect of the signed agreement are separate questions.

Record: supporting approval, rationale, responsible lawyer, and required business decision. Do not close a deviation merely because no objection has surfaced.

7. Close the comparison without changing the signed record

Verify every changed passage against both sources. Record targeted checks of unchanged provisions on which the business will rely, such as permission to share with a particular adviser category.

Prepare a closeout summary identifying reconciled changes, unresolved deviations, and any recommended disclosure limitations. If correction is needed, prepare a separate proposed redline or corrective document. Do not overwrite the executed original or present a working redline as signed text.

Record: reviewer, final status, disposition evidence, and location of proposed corrections. For assigned open issues, include the next decision point—for example, before contractor access. Completion means resolved or expressly assigned, not universally accepted.

Worked example

Fictional scenario: Alder Circuit LLC and Meridian Lens Inc. negotiate a mutual NDA for an optical inspection evaluation. Alder will disclose circuit drawings; Meridian will disclose calibration methods. Alder plans to involve an outside technical contractor. Counsel approves D04, which expressly permits technical contractors subject to the representative conditions. The returned signed PDF is E01.

The comparison identifies a deleted recipient category, deleted analyses wording, and a missing purpose schedule. The transmittal says only “signed copy attached.” No approval of these differences has been located. All parties, documents, and excerpts below are fictional.

ID / topic Draft cite / text Executed cite / text Difference Mutual impact Evidence / gap Action / owner Status
V01 / Recipients D04 § 3, p. 2: “employees, advisers, and technical contractors” E01 § 3, p. 2: “employees and advisers” Contractor category deleted Alder’s planned access is uncertain; Meridian also loses express contractor permission No approval found; whether “advisers” covers this contractor is unresolved Counsel: assess remaining wording before access Unresolved deviation
V02 / Analyses D04 § 1, p. 1: “including analyses derived from that information” E01 § 1, p. 1: phrase not present Express analyses coverage deleted Each party’s evaluation notes may contain the other’s information Deletion verified; remaining definition requires interpretation Counsel: assess scope and propose clarification if needed Unresolved deviation
V03 / Purpose D04 Schedule A, p. 5: “evaluation of the optical inspection project” E01 § 2, p. 1: “the purpose in Schedule A”; schedule unavailable Referenced purpose cannot be verified Neither party can confidently confirm permitted project use E01 contains four pages; executed schedule not supplied Deal owner: obtain schedule and signing-package evidence Verification blocked

The findings change the immediate recommendation: hold contractor access while counsel assesses V01 and obtain the missing schedule before confirming permitted use. V02 requires interpretation, not a conclusion that every derived document is unprotected. A later-supplied schedule must be connected to the signing package; finding a matching draft schedule alone does not resolve V03.

Common mistakes

  • Using the newest file as the baseline. Trace substantive approval and its conditions to an identifiable document.
  • Assuming mutual wording creates equal exposure. Test each change against both parties’ information and intended recipients.
  • Accepting a clean text comparison. Inspect rendered pages, schedules, signatures, and handwritten changes.
  • Treating unavailable text as deleted. Distinguish missing pages and unreadable scans from verified absence.
  • Repairing the signed file. Preserve it and prepare any proposed correction separately, with its approval status clearly identified.

Final quality check

  • Baseline approval is tied to an identifiable file and conditions.
  • Both packages have stable document IDs and component inventories.
  • Changed excerpts and section/page citations match the originals.
  • Definitions, cross-references, schedules, and signature blocks were inspected.
  • Each substantive difference has a two-direction impact assessment.
  • Legal propositions remain within their cited primary sources’ scope.
  • Every open issue has an owner, next action, decision point, and explicit status.
  • The closeout distinguishes executed text from proposed corrections and recommended disclosure limits.

Complete the workflow with DecoverAI

DecoverAI is the Legal Intelligence Platform for litigation and transactions. Its contract review capabilities support review and redlining in Word: choose the represented party or neither and provide deal-specific instructions. Findings quote clauses, rank severity, locate text, and identify missing language as not present. Lawyers can edit or insert suggestions as ordinary tracked changes in the signed-in Word user’s name, or dismiss findings with reasons. Use a separate working copy for proposed corrections and independently verify the executed package. To discuss this NDA comparison workflow, book a demo.

Frequently asked questions

Which draft should be the comparison baseline?

Use the last demonstrably approved substantive draft, including conditions. If competing versions have plausible approval histories, compare both provisionally and assign counsel to determine the baseline.

Does every execution-stage change require corrective documentation?

No automatic conclusion follows from a difference. Record its effect and approval history. Counsel should determine whether documented acceptance, clarification, or separate corrective documentation is appropriate without assuming an unexplained change defeats the agreement.

How should different signing dates be handled?

Record each signature date separately from the stated effective date. Examine effectiveness language and coverage of earlier disclosures. Do not select a date solely for administrative convenience or assume the last signature date controls.

What if only a scan is available?

Preserve it, use extracted text where helpful, and visually verify relevant passages. Request a clearer or complete copy when necessary. Mark uncertain text unreadable—not absent—and retain material verification gaps as open issues.

Sources

The statutory link is to the official 2024 edition, not a certification of intervening-law coverage through October 13, 2026. The filed agreements illustrate negotiated drafting, not market standards.

This guide is general information about legal workflows, not legal advice for any matter. It was drafted and reviewed with AI assistance. The worked example is fictional. Contract rights and procedure depend on the governing law, the actual wording, and the facts, so check the cited sources and the rules that apply in your jurisdiction before relying on it.

Run this workflow in DecoverAI

Clause-quoted findings, missing-language flags, and tracked changes in your name — with your team deciding every call. Plans from $299 per seat per month, SOC 2 Type II certified.

Book a demo
to learn more.
Book a Demo