For decades, much of the dealmaker’s edge came from the ability to process more information, move faster through diligence, build better target lists and keep track of hundreds of details across a transaction.

AI is beginning to erode that advantage — and replace it with a different one.

A new global study from Datasite, based on a survey of 1,000 senior dealmakers, suggests AI is moving from an experimental tool to part of the basic infrastructure of M&A. 62% of dealmakers now say relying on human-only decision-making in complex deals is no longer defensible, while 71% believe firms that ignore AI today will struggle to compete over the next five years.

The implication is bigger than faster workflows. As machines take on more of the research, screening and analysis that once consumed deal teams, the skills that differentiate the best investors may be shifting toward the things AI still struggles to replicate: judgment, accountability, negotiation and the ability to understand the people behind a transaction.

Adoption is already visible across virtually the entire deal lifecycle, but it is far from evenly distributed.

Due diligence is the clearest example. 49% of dealmakers say AI is either fully embedded or used regularly during diligence, compared with 44% in sourcing and screening and 39% in deal marketing. At the other end of the lifecycle, just 16% say the same about closing.

That pattern is revealing. AI penetration is highest where deals involve large amounts of information that need to be searched, compared, classified and analyzed. In fact, 96% of respondents are at least exploring AI in sourcing and screening, while only 4% say they do not use it at all in due diligence.

Closing looks very different. Nearly one-third of dealmakers still do not use AI at this stage, reflecting the greater importance of interaction, empathy and negotiation as a transaction approaches the finish line.

In other words, AI appears to be compressing the information-processing layer of M&A faster than the relationship layer.

The more interesting question may no longer be whether AI participates in a deal, but how much decision authority it should have.

Dealmakers are already comfortable giving AI a meaningful role in earlier, analytical decisions. When identifying potential red flags during diligence, only 15% believe the task should remain entirely human. Another 40% favor an AI-first recommendation that is then approved by humans, while 44% prefer a human decision informed by AI.

Target identification follows a similar pattern.

But as decisions become more consequential, authority shifts back toward the dealmaker. 45% say deciding whether to proceed to signing should remain an entirely human decision, and another 33% favor an AI-first recommendation subject to human approval.

That creates an emerging division of labor: AI increasingly finds, organizes and challenges the evidence; humans remain responsible for conviction.

The distinction matters because investment decisions are not simply exercises in finding the statistically optimal answer. Understanding why a founder is selling, whether a management team can execute a transformation, how counterparties may behave under pressure, or when a risk is worth accepting requires context that may never appear cleanly in a data room.

AI can support that judgment. It cannot own the accountability that comes with it.

There is also surprisingly little geographic disagreement over where this is heading.

74% of dealmakers in the Americas and 72% in EMEA believe those who ignore AI will struggle to compete over the next five years. Even in APAC, where the figure is lower, two-thirds agree.

That level of alignment suggests the competitive question is shifting. Simply “using AI” may soon say very little about a firm's sophistication. The differentiator will increasingly be where AI is embedded, what proprietary workflows surround it and how effectively humans use its output to make better decisions.

And as adoption becomes baseline, trust becomes part of the infrastructure.

Dealmakers are not responding to AI's limitations by keeping it out of the workflow. Instead, they are building controls around it.

58% apply human review or validation to AI-generated outputs, making it the most common step organizations are taking to build trust. Purpose-built AI tools follow at 54%, while 52% restrict AI to secure environments. Formal governance is further down the list at 44%.

The hierarchy is important. For now, the dominant answer to AI risk isn't full automation or even governance alone. It's human validation.

That may ultimately define the new deal team. AI provides scale, speed and an ability to interrogate information that no individual dealmaker can match. The human provides context, challenges the output and ultimately decides when the evidence is strong enough to act.

Datasite's research describes AI as increasingly becoming the baseline rather than an emerging advantage. If that transition continues, the competitive advantage won't come from having access to AI. Nearly everyone will.

It will come from redesigning the deal team around it.