From AI Experimentation to the New Deal Team

For much of the past two years, AI adoption in M&A has been defined by experimentation. Individual dealmakers have tested tools for research, document review, summarization, modeling support and investment thesis development, finding new ways to accelerate work that once consumed hours.

But experimentation is only the first stage.

As AI becomes more embedded in dealmaking, the bigger question is shifting from how individual professionals use AI to how firms build it into the way deal teams operate. Datasite’s recent research into AI-driven M&A points to that transition: 62% of dealmakers recognize AI’s potential to improve the deal process, with expectations moving toward much deeper integration of AI into workflows and strategy by the end of the decade.

The next competitive advantage may therefore come less from having access to AI and more from turning it into an institutional capability.

From Individual Productivity to Shared Workflows

The first wave of AI adoption has largely been bottom-up. An associate discovers a better way to summarize a CIM. A principal uses AI to pressure-test an investment thesis. Another team builds prompts to accelerate market research or prepare for a negotiation.

Each use case can create productivity gains, but those gains remain fragmented if they stay with the individual.

The next step is to capture what works and make it repeatable across the organization. In its recent discussion on AI-driven M&A, Datasite describes the emergence of “prompt-first dealmaking,” where AI becomes integrated into workflows spanning sourcing, diligence, analysis, presentations and negotiation preparation.

That represents an important shift. A well-designed AI workflow can begin to encode not only a task, but also how a firm approaches that task: the questions it asks, the risks it looks for and the information it prioritizes.

Over time, the opportunity becomes bigger than saving hours. Firms can begin transforming individual expertise into shared infrastructure.

A sourcing workflow, for example, could help teams screen a broader universe of companies against a consistent investment thesis. During diligence, AI could surface patterns across thousands of documents while allowing professionals to spend more time investigating the issues that matter. Ahead of investment committee meetings, teams could use standardized workflows to challenge assumptions or identify gaps in the thesis.

The dealmaker remains responsible for the decision. But the information reaching that decision-maker can become faster, broader and more structured.

AI Is Already Delivering Value Across the Deal Cycle

Dealmakers report better insights and streamlined processes as the most common benefits of GenAI adoption across M&A.

McKinsey’s research suggests those benefits are already materializing. Across the M&A lifecycle, 57% of GenAI users report more or improved deal insights and 56% report simplified or streamlined processes. Another 40% report faster deal cycles — averaging roughly 30% to 50% faster — while 39% cite reduced costs or headcount requirements.

The question is increasingly not whether AI can create value in M&A, but how firms can make those gains repeatable across teams and transactions.

Scaling AI Changes the Trust Equation

Moving from individual experimentation to shared workflows introduces a different set of challenges.

When one professional uses AI for a discrete task, the risk can often be managed through individual review. When AI becomes embedded across deals, teams and workflows, firms need to answer much harder questions.

What information can be shared with a model? Where does that data go? Which tools are approved? How should outputs be validated? Who has access? And who remains accountable when AI-generated information influences an investment decision?

Datasite’s discussion of AI-driven M&A repeatedly returns to security and accuracy as foundations for adoption. In dealmaking, where teams routinely handle highly confidential information and investment decisions can involve hundreds of millions of dollars, an answer that merely sounds right is not enough. It needs to be right — and dealmakers need a way to verify it.

AI Adoption Is Scaling Faster Than M&A Workflows

Manual validation, fragmented tools and workflow integration remain major barriers to scaling AI across M&A.

The gap is already visible. Deloitte found that 59% of respondents say outputs from their current M&A tools require significant manual validation, while 57% say their tools are not well integrated. Another 42% say switching between tools slows workflows, while data security, cross-team collaboration and trust in AI-generated outputs remain additional challenges.

Perhaps most tellingly, 29% say their current tools are difficult to scale beyond specific use cases. That is precisely the gap firms need to close as AI moves from experimentation to infrastructure.

Governance, in that context, becomes less of a compliance exercise and more of a prerequisite for scale. Clear rules around approved tools, data access, validation and human oversight can give teams greater confidence to use AI in higher-value workflows.

The firms that establish stronger guardrails may ultimately be able to move faster, not slower.

Governance creates trust. Trust enables adoption. And adoption is what turns isolated productivity gains into institutional capability.

The New Deal Team

As AI takes on more of the baseline work involved in gathering, organizing and synthesizing information, the role of the dealmaker also begins to shift.

The result is not necessarily a smaller deal team. It is a more leveraged one.

Junior professionals can spend less time on repetitive information processing and more time developing analytical judgment. Senior investors can get to key questions earlier. And firms can make institutional knowledge more accessible instead of leaving it scattered across old presentations, inboxes and the memories of individual partners.

At the same time, the areas where humans create the most differentiated value become more important: judgment, interpretation, relationships and negotiation.

That distinction matters because access to powerful AI models is unlikely to remain scarce. The same underlying technology will increasingly be available to firms across the market. Simply having AI will not be a sustainable competitive advantage.

What will be harder to replicate is the system built around it: proprietary information, institutional knowledge, repeatable workflows, governance and the judgment of experienced investors.

That is where the next phase of AI-driven dealmaking may be decided.

The first chapter was about giving dealmakers new tools. The next is about redesigning how those tools, people and information work together.

The firms that pull ahead may not be those that experiment with AI the most, but those that are best at turning that experimentation into a repeatable, trusted and institutional way of working.