In-Brief:
- AI can broaden due diligence review, but buyers should identify which findings constitute material risks and determine their implications for the transaction.
- Sellers must protect confidential information without withholding information material to the buyer’s decision.
- The digital review record may shape later disputes over what was disclosed, identified or omitted from the diligence report.
AI can make a data room easier to review. Its use also introduces new risks. Buyers still need to distinguish material risks from minor defects. Sellers must balance confidentiality with meaningful disclosure. Both sides should consider what the digital record of the review may later reveal. In M&A transactions, the challenge is not simply to review more documents faster, but to ensure that greater visibility leads to better decisions rather than greater uncertainty.
Assessing materiality in AI-assisted due diligence
A red-flag due diligence report is not an inventory of every drafting defect in the target’s contracts. Its purpose is to identify matters that may affect (among other things) the transaction price, the conditions to completion or the contractual protections required by the buyer. Each finding should therefore be assessed for materiality. While monetary thresholds may inform that assessment, the central question is whether the issue warrants action in the context of the proposed transaction.
AI tools can identify a large number of potential issues quickly. Assessing their significance to the transaction nevertheless requires legal and commercial judgment.
Consider change-of-control clauses. An AI tool may identify consent or termination rights across dozens of contracts. Where two of those contracts are with key customers and account for a substantial proportion of the target’s revenue, they warrant specific consideration, including whether the relevant consents should be obtained as a condition precedent to completion. The remaining contracts may not justify bespoke protection in the transaction documents.
The assessment is not solely commercial. Materiality also requires an understanding of the applicable legal regime.
For example, a restriction on the transfer of shares or a pre-emption right may have different practical consequences depending on whether the target is governed by mainland, DIFC or ADGM law. The question is not simply whether the restriction exists, but whether it could delay completion (and by how long), prevent registration of the transfer or require a tailored solution in the transaction documents.
AI can support a broader review and help lawyers identify issues more efficiently. Responsibility remains with the lawyers to determine which findings are material, what should be reported and how those findings should be addressed through further diligence, pre-completion action or contractual protection.
The Seller’s Perspective: Balancing Confidentiality and Disclosure
Sellers should prepare data rooms on the assumption that every document may be reviewed by an AI tool. AI can cross-check documents rapidly and at scale, making inconsistencies that might previously have gone unnoticed, such as a board resolution that conflicts with the company’s constitutional documents, more readily identifiable. Commercially sensitive information may also be processed through systems outside the seller’s direct control, including third-party AI platforms used by the buyer or its lawyers.
Statutory protection is available under the new Civil Transactions Law (Federal Decree-Law No. 25 of 2025) (“New Civil Code”). Article 123 provides that “Any person who uses or discloses without permission confidential information obtained in connection with negotiations or a contract shall be liable in accordance with the general rules.” This protection does not depend on an NDA having been signed, but should not replace express contractual safeguards. A properly drafted NDA should address AI use specifically, including whether recipients may upload data room documents to AI tools, which tools are permitted, and whether the information may be retained by the tool provider.
The instinctive response to AI-assisted scrutiny may be to redact more. Some information should remain protected, particularly where it is not relevant to the buyer’s assessment, such as employee home addresses and bank account information. However, excessive redaction of material information creates both commercial and legal risks.
The first is commercial. If key contractual terms are redacted, the buyer may be unable to properly assess the agreements. This may lead the buyer to seek broader warranties and specific indemnities to account for the resulting uncertainty.
The second is legal. Article 122 of the New Civil Code requires a party to disclose known information of decisive importance to the other party’s consent where that party is presumed to be unaware of it or has placed its trust in the disclosing party. The obligation applies to both parties and requires due care in providing information relevant to the negotiations and the proposed contract.
Given the proactive nature of the duty, merely uploading a document to the data room may not be sufficient. Sellers should consider whether information of decisive importance requires express disclosure to the buyer, rather than relying on the buyer or its AI tools to identify it.
Maintaining a clear disclosure record is therefore important. Under Article 122(3) of the New Civil Code, the claimant must show that the information was withheld, and the defendant must show that it was disclosed.
Sellers should therefore adopt a proportionate approach to redaction that balances confidentiality and data-protection concerns against the need to provide sufficient information for the buyer to make an informed assessment of the transaction.
The Audit Trail: evidential implications of AI-assisted review
Depending on their functionality and settings, data room platforms with integrated AI review tools may record which documents were opened, by whom and for how long. They may also retain records of issues identified by the AI tool.
These records may become relevant evidence in post-completion disputes.
Consider a buyer bringing a warranty claim, or a claim under Article 122 of the New Civil Code, on the basis that a liability was concealed. The seller may be able to produce records showing that the relevant document was available in the data room, that the buyer’s lawyers reviewed it and the duration of that review. If the buyer’s own AI tool flagged the issue, this may support the seller’s argument that the matter was disclosed or known to the buyer. It may not be conclusive, however. Given the proactive nature of the disclosure duty under Article 122, a record showing that a document was available or opened will not necessarily establish that information of decisive importance was properly disclosed.
The same records may also be relevant to a claim against the buyer’s advisers. If a material risk was omitted from the diligence report, the client may rely on a record showing that its lawyers opened the document containing the relevant information, or that the AI tool flagged the issue, without it subsequently being reported.
Firms should therefore adopt a proportionate approach to documenting the review, including the reasons for excluding potentially material AI-flagged issues from the diligence report.
Conclusion/Recommendation:
The value of AI in due diligence lies not in the volume of findings it produces, but in how those findings inform the transaction. Greater visibility must be supported by careful materiality assessments, appropriate disclosure practices and a clear record of the judgment exercised.
For advice on managing these issues in an M&A transaction, Hadef & Partners can help buyers and sellers develop their due diligence approach, disclosure arrangements and contractual protections.
This article is intended for general informational purposes only and does not constitute legal advice. Readers should seek independent legal counsel in relation to their specific circumstances.