AI for M&A lawyers: faster first-pass review of deal documents in Word

Mergers and acquisitions work is document-heavy and deadline-driven: an NDA to open the process, a letter of intent, a purchase agreement, disclosure schedules, transition services, employment and restrictive covenant agreements, and a pile of assigned customer and vendor contracts. AI will not run a deal, but it can take a large share of the first-pass reading off your desk. This guide is honest about where LexDraft fits.

Plain answer first: LexDraft does not have an M&A, merger agreement or asset purchase agreement playbook yet. It has four built-in playbooks: NDA, SaaS MSA, DPA and Services. For a purchase agreement you run a general clause-by-clause review in Word with the governing-law state selected. For the NDA, the data processing terms and the services or transition agreements around a deal, the playbooks apply directly.

Where AI helps on a deal

• The confidentiality agreement. Deal NDAs carry points ordinary NDAs do not: non-solicitation of employees, standstills in public deals, no-contact provisions, and clean-team rules for competitor information. Run the NDA playbook and then check those deal-specific terms yourself.
• Diligence on target contracts. Assignment and change-of-control clauses decide whether the target's key contracts survive the deal. A review of each material contract for anti-assignment, consent, termination-on-change-of-control and exclusivity language is exactly the repetitive reading AI speeds up. LexDraft reviews one document at a time in Word; it is not a bulk data-room extraction tool.
• The purchase agreement. A general review flags one-sided indemnities, missing caps or baskets, vague definitions and inconsistent cross-references. The defined-term and cross-reference quick checks are particularly useful on a 60-page agreement with a long definitions article.
• Ancillary agreements. Transition services agreements read like services agreements, so the Services playbook fits. Where personal data moves with the business, the DPA playbook helps on the processing terms.

Deal points worth a second look

• Structure. In an asset deal the buyer generally takes only the liabilities it agrees to assume. Courts recognise exceptions, typically express or implied assumption, de facto merger, mere continuation and fraud, so the excluded liabilities definition matters. See our asset purchase agreement review guide.
• Representations and indemnity. Survival periods, the basket (deductible or tipping), the cap, special indemnities for known issues, and whether representations and warranties insurance is the main recourse.
• Sandbagging. Whether the buyer can claim for a breach it knew about before closing. The default rule varies by state, so say it expressly either way.
• Materiality scrape and the definition of Material Adverse Effect.
• Purchase price mechanics. Working capital adjustments, escrows and holdbacks, and earn-out definitions and covenants.
• Tax allocation. In an asset acquisition, the parties allocate the price among asset classes under Internal Revenue Code section 1060 and report it on IRS Form 8594.
• Employees. Under the federal WARN Act (29 U.S.C. 2101(b)(1)), in a sale of a business the seller is responsible for any required notice up to the effective date of the sale and the buyer after it.
• Antitrust. Deals above the Hart-Scott-Rodino thresholds need a premerger filing and waiting period. The thresholds are adjusted every year, so check the current figure.

A workflow in Word with LexDraft

1. Open the purchase agreement or target contract in Word and open LexDraft in the side panel.
2. Select the governing-law state and run a general review. There is no dedicated playbook for this contract type, so LexDraft reviews the document clause by clause, flags risky and missing standard terms, and grades each finding Critical, High, Medium or Low.
3. Run the defined-term and cross-reference quick checks. They are rule-based, run without AI and finish almost instantly. Long agreements with schedules are where they earn their keep.
4. Accept the word-level redlines you agree with as native Word tracked changes, reject the rest and add your own.
5. Work through your own checklist (like the one above) for the points a general review will not know about your deal.

What LexDraft does not do

LexDraft has no M&A or purchase agreement playbook, does not draft merger or purchase agreements, does not extract data across a data room, build disclosure schedules, run HSR analysis or give legal or tax advice. It reviews the Word document in front of you. AI review can miss issues or be wrong, so read every marked-up document yourself before it goes to the other side.

Using AI ethically

ABA Formal Opinion 512 (July 2024) and many state bars say the same four things: understand the tool, protect client confidences, check every output, and bill fairly for AI-assisted work. Our state-by-state guide to AI ethics for lawyers covers 23 states and D.C.

Plans

Start on the Free plan: 5,000 generated words a month, no time limit and no credit card. Professional is $99 a month or $990 a year for 100,000 words a month. Business is $199 a month or $1,990 a year with unlimited words. LexDraft runs in Word 365 (Windows and Mac), Word 2021 and Word on the web, and it does not use customer documents to train models.

This guide is general information, not legal advice. LexDraft does not give legal advice; qualified counsel must review every draft. Spotted something out of date? Email support@lexdraft.ai and we will fix it.