The era of testing the waters with generative AI in M&A is over. According to a study from Deloitte, the technology is no longer mostly a pilot project but a top investment priority for senior corporate and private equity executives. With 86% of organizations already integrating gen AI into their deal workflows, the focus is shifting toward practical applications that sharpen due diligence and accelerate insight generation.
In this email Q&A with The AI Innovator, Erik Dilger, managing director of Deloitte Financial Advisory Services, breaks down the three essential takeaways for finance leaders. He explores how firms are balancing fast adoption with proper governance, why the deal team still owns the ultimate outcome, and whether we will ever see a truly agentic M&A process.
The AI Innovator: What are the top three key takeaways from your generative AI M&A survey that would be important for finance leaders to know?
Erik Dilger: One of the clearest takeaways for finance leaders is that gen AI in M&A has moved beyond experimentation and is now a clear investment priority. In our survey of 1,000 senior corporate and private equity leaders, 86% say they’ve already integrated gen AI into M&A workflows, with 65% doing so in just the past year — highlighting how quickly adoption is accelerating.
Of these leaders, 83% have invested $1 million or more in the technology, specifically for their M&A teams (88% of private equity, 77% of corporate). That momentum continues: Many anticipate increasing their gen AI investments over the coming 12 months either slightly (54% of private equity, 58% of corporate), or significantly (24% of private equity, 28% of corporate).
Just as important, most respondents expect returns within one to three years, with improved risk assessment and cost savings cited as the leading measures of value. For finance leaders, that means gen AI should increasingly be viewed through a capital allocation and ROI lens, not just as a technology pilot.
Second, with gen AI adoption strongest earlier in the deal lifecycle – and in areas where faster insight generation and better analysis (strategy and diligence) can improve decisions before signing, for finance leaders, the implication is that gen AI is becoming a practical tool for sharpening diligence, strengthening assumptions, and improving the speed and quality of deal evaluation.
A third key takeaway is in the areas of governance and guardrails. Corporate and PE leaders cited data security (67%), data quality (65%), model reliability (64%), ethics (62%), and regulatory compliance concerns (61%) as the leading barriers to adoption, underscoring that finance leaders will need to stay closely involved in risk management and operating model decisions.
Based on your findings about where gen AI is being applied in M&A workflows, do you think users are targeting the right use cases? Why or why not?
Broadly, users are targeting the right use cases, especially for where the market is today. Deloitte’s GenAI in M&A Study shows GenAI adoption is strongest in the early, information-heavy parts of the M&A life cycle, including strategy and market assessment (40%), followed by target identification and screening (35%) and due diligence (35%).
These stages involve large volumes of unstructured data, repetitive analysis, and draft generation — areas where gen AI can accelerate insight without taking final judgment away from deal teams. The specific use cases respondents cited also reinforce that point: 48% use gen AI for initial drafts of legal and regulatory materials, and 41% use it to analyze financial and operational data against benchmarks. In other words, users are starting where gen AI can act as a force multiplier rather than a decision-maker.
With that, organizations still have room to push further into higher-value applications across the full deal life cycle. Yet expansion into these higher-value applications is being approached with caution. Leaders are clear that organizations are balancing enthusiasm with concerns around data security, accuracy, explainability, hallucinations, workflow integration, and compliance.
Were there any findings that surprised you?
One thing that stood out was how quickly organizations have moved from experimentation to real deployment. Even though gen AI is often viewed as early-stage, many organizations are already embedding it into core parts of their M&A processes and backing that up with meaningful investment.
What’s equally interesting is that this adoption is happening alongside consistent concerns around data, reliability, and governance. So rather than unchecked enthusiasm, what we’re seeing is more of a disciplined acceleration — organizations moving quickly, but with clear guardrails in place.
We see this trend accelerating even faster with the latest releases and updates to the gen AI models.
If gen AI becomes a co-pilot in M&A decision-making, who ultimately becomes responsible for a bad deal driven by flawed AI insights – the deal team, the model provider, or the board?
The deal team ultimately remains responsible. M&A is still a nuanced, judgment-driven process. Gen AI should be used within secure, well-governed environments with human review, validation and clear accountability policies. In practice, that means the model provider may contribute to risk and the board is responsible for oversight, but management and the deal team still own the decision and its outcome.
How do you see gen AI changing the nature of M&A in the long run? For example, will we see a fully agentic M&A process in the future?
The majority of corporate and PE leaders from our survey indicated they expect gen AI to have a moderate or significant impact on M&A decision-making within two years, and they see future benefits extending well beyond early-stage research into due diligence, integration, execution and overall capability building. There is a future state where gen AI likely plays a much larger role in surfacing insights, drafting materials, stress-testing assumptions and accelerating workflows across deals.
That said, dealmaking very much remains a nuanced, judgment-driven process. We are more likely to see agentic elements within M&A workflows — for example, tools that automate pieces of diligence, analysis, or integration planning — rather than a fully autonomous end-to-end M&A process.
Be First to Comment