Review 4.3: cross-document consistency in transaction suites
Contracts rarely exist in isolation. Review 4.3 introduces suite-level memory, allowing you to review a statement of work directly against its master agreement.
Contracts are rarely drafted as single, self-contained documents. Commercial reality demands modularity. A master services agreement (MSA) establishes the legal framework, while a sequence of statements of work (SOWs) dictates the commercial deliverables. A share purchase agreement (SPA) contains the warranties, while a disclosure letter qualifies them.
Until now, contract-review software has treated these documents as isolated text files. If you loaded an SOW into Review, the system would parse it against your firm’s playbook. It would flag missing liability caps or non-standard governing law clauses. But it could not tell you that the SOW was actively contradicting the master agreement sitting in the partner’s inbox.
Review 4.3 ships today. It introduces suite-level context to the Word add-in. You can now define a document hierarchy, allowing our engine to read a subordinate document through the lens of its governing parent.
The architecture of a transaction
When you open a draft in Word, the Review sidebar now includes a ‘Governing Context’ panel. If you are reviewing a schedule, an SOW, or an order form, you can attach the parent agreement directly from your local machine or via our two-way DMS sync.
At launch, Review 4.3 natively recognises and maps the dependencies for three common transaction structures:
- Master agreements and subordinate statements of work or order forms.
- Share purchase agreements and disclosure letters.
- Framework agreements and data processing addenda (DPAs).
Once a parent document is attached, ZAAN-7B Counsel performs a background pass of the parent text. It does not run a full playbook review on the parent. Instead, it extracts a structured graph of the transaction’s architecture. It maps the defined terms, the order of precedence clause, the liability framework, and any explicit prohibitions on what subordinate documents are permitted to amend.
This graph is then held in local memory. When you initiate the review of the subordinate document, every clause is evaluated not just against your firm’s playbook, but against the specific constraints of the parent agreement.
Order of precedence and the rogue schedule
The most frequent application of this feature is the rogue SOW. In many organisations, master agreements are negotiated by external counsel or senior in-house lawyers, while SOWs are drafted by commercial or procurement teams. This creates a predictable pattern of structural conflict.
A commercial manager might copy an old SOW template that includes its own limitation of liability clause, or one that attempts to shorten the payment terms from sixty days to thirty.
If the master agreement contains a strict order of precedence clause—stating that the MSA prevails in the event of any conflict, and that an SOW may only amend the MSA if it explicitly cites the clause being amended—a standard SOW limitation of liability is entirely ineffective. Worse, it creates ambiguity that invites a dispute.
When Review 4.3 encounters this, it flags the issue as a High severity deviation. The inline redline proposes deleting the rogue liability clause in the SOW. The accompanying comment does not cite your general playbook; instead, it cites the specific section of the attached MSA. The comment will read: “Deleted to align with Section 11.2 of the Master Agreement, which prohibits limitation of liability amendments in subordinate documents unless explicitly stated.”
Expanding to the disclosure letter
The mechanics of suite review apply equally to M&A transactions. When reviewing a disclosure letter against an SPA, the critical test is whether the disclosures meet the standard required to effectively qualify the warranties.
If the SPA states that disclosures must be “fair, specific, and accurately cross-referenced to the relevant warranty,” a general disclosure stating “we disclose all matters contained in the data room” is legally insufficient.
When you attach the SPA as the governing context, Review 4.3 calibrates its threshold for the disclosure letter. If it reads a general, sweeping disclosure, it flags it as a Note severity issue. It alerts the reviewing associate that the drafting fails to meet the specific standard of disclosure agreed in the SPA’s interpretation clause, prompting counsel to request precise data-room references from the sellers.
Context windows and the 500-millisecond budget
Building suite-level review was not a product-design challenge; it was an engineering constraint. We maintain a strict 500-millisecond budget for clause-level review latency. Lawyers will not wait thirty seconds for a sidebar to populate.
Modern large language models have expansive context windows, but they suffer from the ‘lost in the middle’ phenomenon. If you paste a 120-page MSA and a 40-page SOW into a single prompt, the model’s attention mechanism degrades. It will reliably recall the first and last few pages, but it will hallucinate or ignore definitions buried on page 68.
We do not feed the parent document into the prompt window. Instead, ZAAN-7B Counsel compiles the parent agreement into a deterministic reference table during the initial attachment phase. When the model reviews a clause in the SOW, it queries this table just as it queries your firm’s playbook. This separation of extraction and evaluation keeps our latency stable, regardless of how long the master agreement is.
Review 4.3 is rolling out to all existing tenants this week. There is no configuration required, and zero retention remains strictly enforced across all attached documents.
A contract is only as strong as the documents attached to it.