The Next Challenge in M&A Dealmaking: Governing AI

Ashwin Moranganti

00:15 - 27:48

Welcome, everyone, and, thank you for joining us today. My name is Ashwin Muranganti.

I'm the VP of product here at SSNC Intralinks. We have a thirty minutes packed agenda here for you, so let's just, get right into it.

But one thing one more important thing, please note that the audience can use the q and a section to submit the questions for me. That we will address at the end of the session.

Okay? So today, we are sharing results from what we believe is the one on the most comprehensive study ever conducted on AI adoption in m and a deal making. We partnered with Reuters Insights to survey about 400 senior deal professionals, partners, managing directors, c suite executives, and analysts across all private equity firms, investment banks, corporate acquirers, venture capital, and law firms worldwide.

And the headline is AI has crossed the adoption tip point tipping point. But the story doesn't end there.

In fact, that's where it starts getting interesting. Because while nearly every deal team has moved beyond pilots, most firms are now grappling with the harder set of questions.

How do we govern this? How do we trust it? How do we close the gap between what AI can do and what our organizations are actually ready for? These are the questions this report and this webinar is designed to answer. And let's start with, you know, who we talk to.

The report is based on 400, deal professionals, as we said. They were surveyed in this, last, quarter, q one twenty twenty six, by Reuters in partnership with SSENSE Intralinks.

The sample is structured, so 80 responded from each of the five org types, ensuring no single segment dominated the findings. So we and, geographically, we covered US, Canada, Europe, UK, Asia Pacific, and the rest of the world.

And one thing I want to call out before we get into the findings that this isn't a survey of just early adopters or AI enthusiasts. These are actual deal professionals whose day jobs are closing transactions.

And they are telling us unambiguously what AI looks like from inside. Let's just get right into finding one.

Let me start with the statement. The question of whether the deal teams are using AI has been definitely answered.

Look at these numbers. 49% of the dealmakers report AI is fully integrated.

Mark my words. I just said fully integrated across most deal stages.

41% said partially integrated. Meaning, 90% or nine in 10 deal teams have moved beyond the pilot phase.

You know, the notable Goldman Sachs global co head of investment banking, Kim Posnath, He said it plainly in Fortune. Just this January, global enterprises have moved past the pilot phase and into a period of deep structural transformation.

Now that's a very important nuance there. While 49% say AI is fully integrated, only 28% describe it as well integrated within their core deal platforms.

That gap also tells us a lot. Teams are using AI, but often through a patchwork of general purpose tools layered on top of some of the existing workflows, not embedding within them.

That's a distinction that matters enormously for security and governance, which we'll come back to. And there's a striking divergence by our type.

You know? Advisory firms and investment banks, they lead at 69%. Full integration, they say.

69% full integration. PEs follow at about 58%.

Corporate acquirers sit about, like, 29%. Law firms are doing great at 39%.

And that's that's your finding one. Let's move on to finding two.

AI is delivering real and measurable time savings. So is AI actually delivering the dataset? Yes.

Clearly and consistently across every phase of the deal life cycle. In deal sourcing and screening, 87% of the respondents report that time savings of at least 11%.

The largest cohort saying about 20 to 30% through AI automated financial signal extraction and company discovery. In due diligence, where AI's impact is perhaps the most pronounced, 35% of the dealmakers say 21 to 30% of their time.

The financial analysis is a top application at 69%, you know, and followed by naturally, as you can imagine, document review. There's too many documents in the data room to review at about 61%.

Cybersecurity diligence, people have used it for about 53%. Legal and compliance work, about 49%.

Think about that. AI is now deployed across nearly the full spectrum of the diligence work.

And in deal execution, 81% report time savings of at least 10%, with AI handling q and a, red flag detection, transaction coordination. The most notable quote come from John's teacher.

He's a CTO at, Blackstone. AI gives you the ability to see lot more deal flow than you were able to, you know, do before.

Before, you were pretty much bounded by the analyst time, and you had to open up a deal room, read to hundreds of documents, and all of this, it reduced the friction to transact. That's a notable quote.

That phrase, reduces the friction to transact is the right frame. AI isn't replacing judgment.

It's removing the bottlenecks that slow judgment down. Fewer than 86% report savings exceeding 50%.

You know? That tells us a lot that there is significant headroom for efficiency gains as the integration, deepens. Let's get going.

Let's go to finding three. Trust is earned.

Confidence is not uniform, but it's not uniform. Here's a finding that might surprise you.

Dealmakers broadly trust AI outputs. A majority rate they trust at 70 out of 100.

I mean, that's pretty high. Or higher across every work stream we measured.

Surprise, surprise. Valuation and financial modeling earns the highest trust at 62%, followed by post merger integration at 58 and due diligence at 56.

But more importantly, four and five dealmakers are comfortable with AI executing multistep workflows autonomously. I mean, that's that's news.

I mean, that's the big highlight. The trust in AI is really increased.

That's a remarkable signal where the market is heading. We'll come back to it again when we talk about where the deal professionals see where they're in 2030.

But here's the paradox buried in that data, what we found. When asked where AI aids human judgment the least, respondents pointed to deal execution and due diligence.

The very phases where AI is most heavily deployed, that's not a contradiction, though. It reflects the fact that these phases involve the most complex contextually and legally consequential decisions in any transactions.

AI can surface these patterns and flag anomalies, But the step from insight to action, particularly in complex high stake moments, demands human judgment. And increasingly, that means not just a human in the loop, but human in the lead.

I like that. The implication for firms, the next way of efficiency gains won't come from by adding more AI tools to individual task.

They will come from integrating these tools into connected and continuous workflows. The agent capabilities that everybody is raving about, you know, the financial industry is about to embrace.

Let's, go to the next one. The next 80% and this is, I have to I wanna pause here.

I want you guys to look at the slides. 80% of the firms have already experienced AI security incidents.

I mean, that's that's a lot. Four and five, that's not a fringe risk.

This is operational reality affecting the overwhelming majority of the dealmakers today. The most prevalent incident type is access control lapse.

AI agents given permission over too much sensitive information are making unwanted changes to the digital assets. That's reported by nearly half of the respondents at 48%.

Right after that, the hallucinated AI outputs, which we thought, you know, would be a major concern, is at about 40%. Here's the irony, right, that makes this whole finding particularly sharp.

95% of the same firms operate you know, all of them have at least one formal AI governance framework. I mean, these are you know, if you talk about in the industry, this industry has the most, one of the most highest governances.

So the problem isn't an absent of the policy. Right? The gap is between policy and practice.

And that's widening as AI adoption is accelerating faster than the governance structure can mature. You know? I think the most notable quote, is from the Bank of America CTIO.

It's Hari Gopalakrishnan. He put it this way.

I think the concept of guardrails and responsible AI is a big part of what we do. We have a process by which we look at 16 different dimensions at risk of risk.

And and that here's the structural issue that creates much of that risk. Right? When deal teams move documents in and out of their data rooms to plug into these external AI platforms, they're creating data exposure with every transfer.

A VDR whereas with which is m AI embedded natively, where the data never leaves the secure environment is architecturally different in ways that matter enormously at scale. So let's move to the next one, finding five.

57% of the dealmakers report senior level resistance to AI. That has increased in the last twelve months.

You might think that as AI proves its value, the resistance would decline. The opposite is happening, and it's not irrational.

The top drivers of senior resistance are actually, as you can imagine, accuracy concerns, explainability, where you get that information, the fiduciary risk, and the client perception. Right? These are important.

These are in technophobes. They are accountable leaders.

And increasingly, they are being asked to put their name behind the analysis that they don't fully understand. Consider this seniority split in the data.

67% of the c suite and partner led respondents consider AI fully integrated, but only 29% of the associate and the analyst say the same. The senior leaders may be significantly overestimating the AI maturity in the firms.

That's, you know, that's what I take out of it. There's a notable quote from the chief operators, operating officer of Apollo, Byron.

Our gate is not money. The gate is the organizational's ability to handle all of the change.

I I think that's a very profound statement. The gate is not money.

The gate is the organization's ability to handle all of this change. We are investing in as much in people and operating model behind the digital infrastructure as we are in the AI itself.

And that's two way, people and infrastructure. Because that's what turns capability into impact.

The solution isn't really, as we talk, converting skeptics into evangelists. It's building structured AI literacy at the leadership level so the people who sign off on AI driven outputs understand how these outputs are being produced.

And that's a governance investment, not a technology investment. I go to one of the, finals, finding six.

The twenty thirty deal maker, you know, let's look ahead. Because what the survey reveals about the future of the deal making is both exciting and sobering.

Okay? When asked which AI capability would be the most disruptive by 2030, 43% of the respondents pointed to automated quality of earnings. You know, that's that's profound.

I mean, that's the it's two points. Will it take that long? And, but, you know, people are picking up the most complicated stuff, and that's it.

That's fully automated. You know, that's what they're talking about.

It's not just, you know, faster and things like that. Fully automated.

Comprehensive full population analysis rather than just sampling. You know? Think about that, what it means.

A capability that currently requires teams of analysts we can working for weeks, automated, and run across every document in the data room. That's not incremental improvement.

That's step change, how deals get done. Following closely, as you can imagine, fully autonomous digital agents at 41%.

So just click and say do financial due diligence, do HR diligence, like that. AI generated SIMs.

You know? Not sure why it's only at 39%. I would have assumed it's gonna be higher, but that's where it is.

And predictive deal outcomes modeling at 39%. So but these are automated.

You know? I I stand corrected. This is pretty important.

They're less human intervention, more automated. That's what they are expecting in 2030.

So, that another important in order for this to happen, how does the team composition in these deal teams changes? Right? That's an important one. 64% expect ML and AI specialists embedded directly in deal teams by 2030.

Okay? So the deal teams are not just gonna be having, like, finance experts or research experts, but AI and ML specialists embedded in the deal team. Good sign for the, upcoming new graduates too.

53% anticipate a shift toward AI literate generalist, okay, replacing the silent specialist silos. That's people who know how to use AI is gonna be much more important.

Anyway, so let's go to the next slide. This was also an important one.

The market wants specialization, not consolidation. Okay? 49% of the dealmakers prefer multiple specialized AI tools.

Right? They want purpose built for sourcing, purpose built for diligence, for valuation, or a single platform. Only 20% want it to be an all in one system.

28, though, say that they want all of the tools to be well integrated into their core deal platforms. That's a gap.

Teams are running AI alongside these workflows, not inside them. And 54% expect AI to be a standard infrastructure in deal making, not a plug in, not a bolt on infrastructure.

The next battleground won't be which firm has the longest AI feature list. It'll be the trust, security, output reliability, and how seamlessly AI is embedded in the, platforms, where deals actually get, done.

This is this is very profound. And, so that's a good segue into what we are doing at, Intralinks here.

Everything the survey revealed points to a specific set of requirements for the next generation of AI in deal mating. Let me show how exactly deal center AI, a product of SSENSE Intralinks, was designed to address each one of them.

The survey found that 80% of the firms faced AI security incidents, most of them traceable to off platform data movement. Deal center AI is embedded AI is embedded natively within the data room.

Okay? Your sensitive data never leaves the security environment to be processed by an external model. That's not a security feature.

That's an architecture. That's the architecture.

Only 28% of the dealmakers describe their current AI tools as well integrated with their core deal platforms. And that's what deal center's goal is, to cover the full transaction life cycle from pipeline management to deal making to, through diligence, through q and a, and close.

You know, one platform built for the deal, not assembled around it. The market says 49% prefer specialized tools and 54% expect aid AI to be a standard in the deal making software.

Deal center is purpose built for m and a. It's not a general purpose.

We didn't take a general purpose language model reposition for deals. The difference shows in every interaction on how documents are processed and how q and a is managed to how historical deal data is mined for insight.

And, as a survey's forward looking data makes it clear, four in five dealmakers are ready for autonomous workflows, deal center AI. We are building it with full guardrails, bank grade security, full audit trails, and controls beyond the you know, designed for fiduciary grade decision making.

And SSNC as a company has executed, like, 35 trillions in global transactions on SSNC's platforms. Deal center AI brings that legacy of trust into the era of AI.

It's not a feature. It is a foundation.

One segue, the the there is the next phase of AI in deals. I just wanna point out it's kind of a summary of all of these things.

Governance versus practice. You know, that's the most most important.

You can use a lot of AI tools, but governance is more important. 94% of your peers have AI governance frameworks, but 80% experience security incidents.

So the question isn't whether you have a policy or not, if the policy is there. But the gap between what's written and what's practiced is where the risk lies.

Okay? That's number one. Number two, leadership literacy.

Senior leaders are signing off on AI assisted outputs that may be, you know, fully, but they need to understand how to evaluate them. Right? That's not a technology problem.

It's a fiduciary one. What does it take to build genuine AI literacy on top so that people with ultimate accountability can assess? And who who built this twenty thirty team? And that's an important one.

How does that twenty thirty team composition look like? 64% expect ML and AI specialists in their team and generalists with AI skills. So and agentic AI is coming.

So I'm we have seven minutes. I'm gonna open I I tried to rush through the last few ones.

I wanna see if there is any q and a from the audience. Two questions in the q and a section.

I see that. Rodrigo, the main concern we have today with AI, DD, is accuracy.

What do you think is the best way? Now that's a great one. You know? The survey identifies liability fear at the top change management object obstacle.

Right? The honest answer, the legal frameworks are still evolving. What firms are doing now is ensuring audit trails knowing just what the AI output was, but how it was generated.

That's what that gives them the confidence. What data it drew from, the traceability, sourceability, and what human review it received.

That documentation is your production. You know? I think the q and a is increasing.

What size were the deal teams you pulled for in your data? Mid market and bulge. Actually, it's both both mid market and bulge.

Don't know the exact, composition. I'll go to the next one.

How are firms approaching data and context engineering before, adopting these tools? That's that's a you know, it's really the most important one. For AI to be really successful, it's really the data and the context.

So you're connecting to an m and a data in the VDR room, and that's immutable, and that stays there. And, so you should be confident that the data which is being given to the LLMs or the AI is, you know, grounded in truth there.

But then the context that you bring in in addition to that and that is where you need to be really checking in. This is an emerging trend.

So we don't know where it'll go to. I'm not sure exactly I've answered, but it's a great question.

We talked about the AI accuracy. We talked about, what does multistep processes mean in this research? I thought agentic AI is not really in.

Multistep processes is you know, let's say you wanna build a teaser. Right? First, extract the following information from the data room, then bring in this information from S and P five hundred Global, get do this thing and create a, the output in this format.

That's a multistep process into can come up with an, outcome. But, if you'll, you know, take a cloud and you just say build an a teaser, it can do it.

You know? It will do the thinking. It does multi steps, and, it's there.

And that's what but the industry is sharpening it to move from probabilistic to more deterministic. Okay? And that's where the journey is gonna be.

One last one. How long do you predict that it will take to AI tools to be implemented and used as a standard? You know? You know, I don't have a crystal ball, and I shouldn't be trusted even if I said that.

But, the from the data that we are looking, I think it's gonna be much faster than we think. You know? I nobody can you know, the AI outcomes that are being generated, it's amazing.

And so nobody will want to avoid that. So they will rush in, but they'll try to make it more become more diligent.

And that's what I see that the executives will push their teams to become more diligent and not just take the AI output, for its, for whatever it says. So I have three minutes, and I think I've taken a good amount of questions there.

In closing remarks, you know, let me leave you with the central message of this report. The industry has crossed the AI adoption tipping point.

The question is no longer whether to use AI. The question is whether your organization's governance, security architecture, and leadership literacy are keeping pace with the technology.

The firms that will lead in AI to enable deal making will be those that close the gap between technology technological capability and organizational readiness. Deliberately, measurable, and before the competitors do.

Please note that the full report is available at interlinks. com.

I encourage you to read it alongside your leadership team, and you will see also you know, that's a small plug in. You'll see how deal center AI can specifically address the gaps we have discussed today.

Our team is ready to walk you through it. Thank you for all your time and your questions.

I really appreciated, the engagement today. Though I felt it's one-sided, but I'm glad I saw a good amount of q and a, submitted.

Thank you very much.