June 1, 2026 · 13 min read

Legal AI for M&A Due Diligence: Workflow Lessons from the $32bn Google–Wiz Deal

Legal AIM&A Due DiligenceData Room ReviewAntitrust

When Alphabet agreed to buy the cloud security company Wiz for roughly $32 billion in March 2025, it set in motion one of the largest deals in the company's history — and nearly a year of legal work before the transaction finally closed in March 2026 (1) (2). Deals at that scale are not won or lost on a single brilliant clause. They are won on disciplined diligence across enormous volumes of documents. That is exactly where source-grounded legal AI can change a lawyer's workflow.

Why a mega-deal belongs in a legal AI discussion

The Google–Wiz transaction is a useful lens precisely because it is not exotic. It is a large, well-documented example of an ordinary legal pattern: a buyer commits to a price, then a legal team spends months confirming that the company it is buying looks the way the deal documents assume it does. Public reporting tracked the deal from the initial approach, through a rejected $23 billion offer and a revived $32 billion agreement, to unconditional EU antitrust clearance in February 2026 and completion in March 2026 (1) (2) (3).

Behind that public timeline sits the work that lawyers actually do: reading contracts, scanning a data room, mapping risk, drafting disclosure schedules, answering regulator questions, and keeping a defensible record of every conclusion. None of that is glamorous. All of it is where deals slow down, where costs accumulate, and where a missed document can become a post-closing dispute.

The practical question for any deal team is simple. In a matter with tens of thousands of contracts, policies, financials, IP records, employment files, and regulatory filings, how can lawyers see the risk earlier, more completely, and with the source behind every flag still visible?

The diligence problem at deal scale

Every acquisition eventually becomes a reading problem. A virtual data room for a significant transaction can hold thousands of documents across corporate, commercial, employment, IP, data protection, litigation, tax, and regulatory categories. The deal team has to find the handful of provisions that actually move value or create risk: change-of-control triggers, exclusivity and non-compete clauses, assignment restrictions, unusual indemnities, termination rights, data protection exposure, and open litigation.

The pressure is structural. Diligence runs against a signing or closing clock, the document set keeps growing as the seller uploads more material, and the most important issues are often buried in long agreements that look routine until a single defined term changes everything. The answer is rarely in one document; it is in the pattern across many.

For a cross-border deal, that pressure multiplies. A transaction like Google–Wiz, with operations and regulators on multiple continents, layers antitrust review, foreign investment screening, and sector-specific rules on top of ordinary contract diligence (3) (4). Each regime asks for its own evidence, on its own timeline, from the same underlying record.

What source-grounded legal AI can improve

Legal AI is strongest in diligence when it helps lawyers move from volume to structure. The useful question is not "Can AI write a polished summary?" It is "Can AI help the team find the right provisions, organize them by issue, and show the document behind every flag?" A lawyer-facing platform can support diligence in several concrete ways.

First-pass triage: sort a large data room by document type, contract value, counterparty, and likely risk category so reviewers start with the material that matters.
Clause extraction: surface change-of-control, assignment, exclusivity, termination, indemnity, and limitation-of-liability provisions, each linked back to the exact contract and page.
Risk flagging: highlight unusual terms, missing signatures, expired agreements, and provisions that conflict with the deal's assumptions.
Issue organization: group findings under diligence headings — corporate, commercial, IP, employment, data protection, litigation, regulatory — for the report and the disclosure schedules.
Regulatory workpapers: assemble the documents, market facts, and overlaps that antitrust and foreign-investment submissions need, with sources attached.
Q&A support: let the team ask questions across approved data room materials without flattening the record into a single unsourced answer.
Review handoffs: turn an unbounded document set into review-ready work packages for associates, partners, specialists, and the client.

None of that removes the lawyer. It changes the lawyer's starting point. Instead of spending the first days simply locating the relevant provisions, the team begins with a source-linked map of the data room that it can challenge, correct, and refine.

A better workflow for data room review

In a large deal, a legal AI workflow should begin with discipline, not with drafting. The first deliverable is not a polished diligence report. It is a reliable working view of what the data room actually contains.

Scope the room. Identify document types, date ranges, languages, jurisdictions, contract values, and access restrictions before review begins.
Triage by risk. Use AI-assisted sorting to prioritize high-value contracts, regulated activities, and anything touching change of control or assignment.
Extract key provisions. Pull the clauses that drive value and risk, and link each one to the exact source document and page.
Separate fact from inference. Mark what a contract says, what the seller represents, what the team infers, and what remains an open question for the seller's counsel.
Build the issue list. Organize findings by diligence heading and assign each item to the right reviewer or specialist.
Prepare review-ready outputs. Draft the diligence report, red-flag memo, and disclosure schedules only after the findings have been tested against the source documents.

This is where legal AI becomes valuable to corporate and competition teams. It does not decide whether to do the deal. It helps the team avoid losing a material risk inside the volume of the data room.

The mega-deal lesson: source links matter

When a transaction turns on what a company has actually agreed, owes, or owns, unsourced summaries are dangerous. A clean bullet point can hide the difference between a binding obligation, a draft, a side letter, an expired agreement, and a lawyer's assumption. On a deal where the price is measured in billions, that difference is not academic — it shapes the purchase agreement, the indemnities, the disclosure schedules, and the post-closing risk allocation.

A source-grounded legal AI platform should therefore make the source path visible. If it flags a change-of-control clause, the lawyer should be able to open the contract and read it. If it says two agreements conflict, the lawyer should see both. If it cannot find support for a point, it should say so plainly rather than produce a confident but unverifiable answer.

That is not a technical nicety. It is professional necessity. The deal team needs to know whether a finding can be relied on in a representation, whether it belongs in a disclosure letter, whether it survives closing, and whether it must be escalated to the client before signing.

Where legal AI helps the regulatory track

Large deals do not just need contract diligence; they need clearance. The Google–Wiz transaction drew antitrust scrutiny from the outset and only completed after the European Commission cleared it unconditionally in February 2026 (3) (4). Merger control, foreign investment screening, and sector regulators all run in parallel with the commercial deal, and each one is its own document-heavy workflow. Legal AI can assist in operationally meaningful ways.

Filing preparation: gather the corporate, market, and overlap evidence that merger filings require, with each fact linked to its source.
Information requests: map a regulator's questions to the documents and data that answer them, and flag gaps early.
Overlap analysis support: organize the materials that show where the parties' activities meet, so lawyers and economists can assess them.
Multi-jurisdiction tracking: keep filings, deadlines, conditions, and commitments aligned across regimes that ask similar questions in different forms.
Consistency checks: surface places where statements in one filing, contract, or board paper appear to point a different way from another.
Privilege awareness: respect confidentiality and privilege boundaries when the same record feeds commercial, regulatory, and litigation workstreams.

The benefit is not simply speed. It is earlier visibility into what the regulators will ask and whether the record can support the answer.

What AI should not be asked to do

A responsible article about legal AI in deals has to be honest about limits. AI should not decide whether a risk is material, set the indemnity package, sign off a disclosure schedule, or replace specialist judgment on tax, antitrust, or data protection. It cannot weigh a commercial trade-off, read a counterparty's intentions, or tell the client whether the deal is worth doing.

There are practical risks too. A model can over-summarize a nuanced clause. It can miss the significance of a short side letter. It can give too much weight to boilerplate that appears in every contract. It can blur the line between a binding obligation and a non-binding term sheet. It can produce a confident answer when the workflow does not force it back to the document.

That is why reliable diligence workflows are conservative. They treat AI as a way to find, organize, and interrogate the data room, not as an authority that resolves the deal.

How lawyers can apply the lesson now

A $32 billion deal is exceptional in size, but the workflow lesson applies to ordinary corporate work. Mid-market acquisitions, carve-outs, joint ventures, financings, real estate portfolios, and private equity bolt-ons all create the same pattern: too many documents, a fixed timetable, and a real cost to missing the wrong provision. Deal teams can start with modest, matter-level rules.

Every AI-flagged clause should link to the source contract and page.
Every diligence finding should distinguish contract fact, seller representation, lawyer inference, and open question.
Every red-flag item should be reviewed by a lawyer before it shapes the purchase agreement or disclosure schedules.
Every regulatory submission should be checked against the underlying documents, not against a summary of them.
Every search across the data room should respect access permissions, confidentiality, and ethical walls.
Every platform pilot should be tested on one repeatable diligence workflow before broader rollout.

These rules are not anti-AI. They are what make AI useful in a setting where the cost of being confidently wrong is measured in price adjustments, indemnity claims, and post-closing litigation.

How LexVera fits this kind of work

LexVera is built around the idea that legal AI should support reviewable legal work, not isolated answers. In a document-heavy deal, that means helping lawyers move between uploaded data room materials, legal research, prior precedents, notes, and draft outputs while keeping source references visible at every step.

For a corporate or competition team, LexVera can help organize deal documents, prepare source-backed summaries, build issue lists, find the relevant provisions across a large set, support drafting, and connect diligence findings to legal authorities the team can inspect. The platform is not trying to hide the complexity of the transaction. It is meant to make that complexity navigable under deal-timeline pressure.

That matters because diligence is rarely a single prompt. It is a sequence: scope the room, triage the risk, test the sources, draft carefully, review professionally, and update the work as the seller uploads more material. Legal AI earns trust when it supports that sequence rather than short-circuiting it.

Questions to ask before using AI on a deal

Before using any legal AI platform on an acquisition, financing, or regulatory filing, deal lawyers should ask workflow-specific questions.

Can the platform show the contract and page behind each flagged provision?
Can it separate target documents, public law, firm precedents, and user assumptions?
Can access be limited by deal, workstream, role, document set, or ethical wall?
How does it handle confidential data room material and privileged analysis?
What happens when a finding is not supported by the uploaded documents?
Can the team review unusual terms, gaps, and uncertainty as first-class outputs?
Can outputs be corrected as the data room grows and the deal evolves?
Does the workflow preserve enough context for partner supervision and later audit?

The right answers will be practical and evidence-led. Deal teams need systems that help them inspect what the record can actually support, not tools that pretend to know the whole transaction.

Conclusion

The Google–Wiz acquisition is a reminder that even the biggest deals come down to disciplined reading at scale, under a clock, across more than one regulator. Legal AI cannot decide whether an acquisition is wise or set the commercial terms. But used properly, it can help deal teams find the provisions that matter, organize them by issue, and keep the source behind every flag visible.

That is the most useful promise of legal AI in 2026: not autonomous deal judgment, but faster diligence with better source visibility. On the transactions where the price is highest and the timetable is tightest, that can change the quality of the advice that reaches the client before signing.

Legal AI should not close the deal before the data room has been read. Its job is to help lawyers see the documents clearly enough to advise, negotiate, and defend the deal themselves.