Agentic AI Is Transforming M&A. Are Lawyers Ready?

The lede

Sixty-two percent of dealmakers say human-only decision-making is no longer defensible in complex transactions, and 71% believe firms that ignore AI today will struggle to compete within five years, according to The New Deal Team, a survey of 1,000 senior dealmakers across 27 countries published by data-room provider Datasite in partnership with the Financial Times' research studio, FT Longitude, in July 2026. The survey — conducted across corporate development, private equity, law and accounting professionals, each with decision-making responsibility on at least three deals in the prior 24 months — found AI now embedded across the entire deal lifecycle: 96% of respondents are using or exploring AI for sourcing and screening targets, and half report regularly embedding AI directly into due diligence, the stage delivering the highest reported return on investment.

The legal profession's own account of what has changed is more specific than "AI helps with due diligence." Law firm Montague Law, describing the shift in a 2026 client analysis, put it bluntly: "The M&A process has quietly crossed a line." Agentic tools in live transactions now review full contract populations rather than samples, draft first-pass disclosure schedules, chase outstanding diligence requests, and reconcile working-capital schedules against the general ledger overnight — work that traditionally consumed junior associates' billable hours for weeks. "For decades, private-company diligence has been an exercise in sampling under time pressure," the firm noted. "Agentic review changes the denominator."

The main report: what agentic diligence actually replaced

The shift is best understood as a change in what gets reviewed, not just how fast. Traditional M&A diligence has long relied on sampling — pulling a company's twenty largest customer contracts, reading leases above a materiality threshold, spot-checking the rest — because reviewing every document in a mid-sized company's contract population was not economically feasible within a deal's typical 30- to 60-day timeline. Agentic tools remove that constraint. Platforms built specifically for legal work now let deal teams upload and bulk-analyze thousands of agreements simultaneously, running pre-built or custom diligence protocols across an entire document set rather than a curated subset, according to legal AI platform Harvey's own description of how the technology is deployed in live transactions. The practical effect, Harvey's analysis argues, is that lawyers spend less time on extraction and more time on "the judgment calls that actually shape deal outcomes" — flagging inconsistencies against the original term sheet, identifying deviations from precedent, and producing structured issues lists for partners to negotiate from.

That shift has produced a genuine market for specialized tooling. A 2026 industry comparison of AI due-diligence vendors found that a typical middle-market private equity fund running deals in the $500 million to $2 billion range now operates a full stack of tools across the deal lifecycle — a customer-relationship platform for sourcing, a virtual data room, a legal-extraction tool for contract review, a separate financial-diligence tool, and post-close contract-management software — at a combined annual subscription cost between $150,000 and $400,000, a figure the same analysis notes sits below the savings typically generated on a single $300,000 traditional diligence engagement. Consolidation is already reshaping that vendor landscape: Workday's acquisition of contract-AI company Evisort, and Litera's ownership of M&A-focused review platform Kira, both signal that standalone due-diligence AI tools are being absorbed into broader enterprise platforms rather than remaining a fragmented specialist market.

Where dealmakers themselves draw the line

The Datasite survey's most consequential finding is not about adoption but about where dealmakers refuse to hand off control. Forty-five percent said the decision to sign a deal should always remain a human responsibility — not a majority, but a plurality large enough to signal that even enthusiastic AI adopters treat deal execution as categorically different from deal analysis. Respondents identified negotiation and relationship management, strategic judgment and prioritization, assessing trust and credibility in a counterparty, and accountability for high-stakes decisions as the attributes hardest for AI to replicate — a list that reads less like residual work AI hasn't reached yet and more like a deliberate account of what dealmaking actually is beyond document processing.

At the same time, the survey found real erosion in the assumption that human judgment is categorically superior: 43% of respondents said AI is already making better deal decisions than humans in some scenarios, and 24% credited AI with helping them complete a deal they otherwise would have missed. Sixty-six percent said AI helps de-risk transactions across the deal lifecycle generally. Trust, rather than capability, is what the survey identifies as the actual bottleneck: accuracy (71%) and security (70%) ranked as dealmakers' two most important requirements for AI tools, and 58% said they rely specifically on human review and validation to build confidence in AI-generated outputs — meaning the verification step legal ethics rules require is also, independently, what dealmakers themselves say they need to trust the technology at all.

Governance remains the least mature layer of adoption. Nearly a quarter of respondents, 24%, said poor use of AI could jeopardize major deals over the next five years, and 27% said they are not using AI for board reporting at all — the furthest point in the deal lifecycle from AI adoption, according to Datasite's own framing of the finding, because board-level recommendations carry the highest accountability stakes of any output in the process.

Legal background: the professional-responsibility rules haven't changed, but their stakes have

Lawyers deploying agentic AI in live transactions operate under the same professional-conduct framework covering any other AI-assisted legal work — there is no separate, M&A-specific set of ethical rules. The American Bar Association's Formal Opinion 512, issued in July 2024, requires that lawyers using generative AI retain full competence, supervision and verification obligations regardless of which tool produced a given output. State-level guidance has begun to make the M&A-specific application explicit: Florida Bar Ethics Opinion 24-1, cited by Montague Law's analysis of agentic diligence in Florida deal practice, applies the same competence and supervision duties directly to agentic tools operating inside a live data room.

What has changed is not the rule but the exposure a violation of it now carries. When diligence meant sampling twenty contracts, an unverified AI summary touched a small, bounded slice of a transaction's risk. When diligence means an AI agent has reviewed a company's entire contract population and generated the first draft of every disclosure schedule, an unverified error can propagate across the full breadth of what the buyer is relying on to price and structure the deal — the same verification duty, applied to a dramatically larger surface area.

Practical implications

For law firms serving M&A clients, the competitive question has shifted from whether to adopt AI-assisted diligence to whether firms can demonstrate credible AI governance to increasingly AI-literate clients. Legal-services M&A activity itself reflects this: according to Arrowpoint Advisory's Q1 2026 Legal Services Market Update, acquirers evaluating legal-services targets have moved from broad optimism about AI-driven efficiency toward more defensive, technically focused diligence specifically scrutinizing a target firm's revenue vulnerability to AI-driven business-model displacement — meaning law firms themselves are now being diligenced on their AI exposure when they become acquisition targets.

For corporate acquirers more broadly, Accenture's 2026 analysis of agentic AI in M&A argues that most acquirers still treat a target's data architecture and AI readiness as an integration afterthought behind financial diligence, while the leading acquirers assess AI readiness and data maturity during diligence itself — treating AI infrastructure as a genuine value driver rather than a technical footnote to be resolved after signing.

For in-house counsel and deal principals, the survey data suggests the operative skill is shifting rather than shrinking: as AI absorbs document-level extraction and first-draft generation, the negotiation, credibility-assessment and accountability functions the Datasite survey identified as AI-resistant become, by process of elimination, where a lawyer's remaining time and judgment concentrate.

What's next

Whether the 45% of dealmakers insisting a human must always make the final signing decision holds steady as AI's demonstrated accuracy improves, or erodes the way sampling-based diligence eroded once full-population review became feasible, is likely to be the central question the next edition of Datasite's survey addresses. In the nearer term, the governance gap the July 2026 findings identified — nearly a quarter of dealmakers worried poor AI use could jeopardize a major deal, more than a quarter not yet trusting AI with board reporting at all — suggests the technology has outpaced the institutional frameworks meant to govern it, a mismatch that professional-responsibility regulators, board risk committees and law firm management are all still in the early stages of resolving.

Sources and references

• AI Is Becoming Indispensable in Dealmaking, But Trust and Governance Will Determine the Winners — Datasite, via GlobeNewswire
• Most Dealmakers Say Human-Only Decision-Making Is No Longer Defensible in Complex Deals — Datasite, via GlobeNewswire
• The Board Decision Gap: AI Transforms the Deal, But Boards Still Govern the Outcome — Datasite, via GlobeNewswire
• Agentic AI in M&A — Accenture
• When AI Agents Run the Deal: What Agentic M&A Execution Changes — and What Still Requires a Human — Montague Law
• The Guide to AI-Powered Due Diligence for M&A Professionals — Harvey
• AI Due Diligence Tools for M&A in 2026: 10-Vendor Comparison + Use Cases — CT Acquisitions
• M&A in the legal services market tilts towards firms with credible AI strategies — Consultancy.uk / Arrowpoint Advisory

This article is intended as independent journalism and analysis. It does not constitute legal or financial advice.