The Role of Virtual Data Rooms and AI in Due Diligence
Learn how tech continues to transform the due diligence process.
Welcome to The Dealist, your trusted insider's guide to navigating the complexities of mergers and acquisitions (M&A). This episode of the podcast explores how technology, particularly virtual data rooms (VDRs) and artificial intelligence (AI), has transformed the due diligence process.
We’re joined by two distinguished guests, Noah Waisberg, CEO of Zuva, and Prakash Kanchinadam, head of AI and analytics at SS&C Intralinks. Together, they explore:
- The history and evolution of M&A, and the introduction of virtual data rooms as an essential tool for secure transactions
- Why VDRs are indispensable for managing today's complex deals, with their built-in security features and ability to handle vast volumes of data
- The role of AI in enhancing due diligence for dealmakers — from automating labor-intensive processes like document classification and redaction to identifying key contractual obligations
- The future of AI in M&A, its opportunities to reduce manual work and its limitations that highlight the continued importance of human judgment
- What the rise of agentic AI could mean for the future of due diligence
Host: Catherine Ford
Further reading: The Dynamics of Due Diligence
Discover how Intralinks DealCentre™ revolutionizes deal management with cutting-edge AI technology and smart solutions here.
Subscribe: Apple | Spotify | YouTube
Transcript
CATHERNE FORD: Welcome to The Dealist, the insider's guide to M&A and the complexities of navigating diligence. My name is Catherine Ford and today we're going to talk about the role of data rooms and AI. And here joining me in this conversation are Noah Weisberg, CEO of ZUVA and Prakash Kanchinadam, head of AI at Interlinks. Thank you very much, both of you, for joining me today. I'm delighted to have you. And I'd like to jump straight in, Prakash, with a question for you. If you can talk to us about VDRs and the advantages of using VDRs in a due diligence process.
PRAKASH KANCHINADAM: Hi, Catherine. And I know, thanks for having me on. It's recent revelation for me. But when I looked at the history of mergers and acquisitions, it's come to my notice that mergers and acquisitions have been going on for about 100 plus years. It started with a great merger moment as at the turn of the century. when big entities came together to make even bigger entities. And that's when it all started, I guess, but over 100 plus years, mergers have been happening and then more and bigger companies have been emerging as a result of that. Over 30 years ago, then that's when technology came into the forefront to make these negotiations and deal-making happen.
PRAKASH KANCHINADAM: That's because of the virtual data room as a result of technological transformation and making technology being the enabler for these transactions to be conducted in a more secure fashion. and that's where what VDRs are known to be today came to the forefront.
CATHERNE FORD: If we just take a step back there, Prakash, do you actually think it would be possible to do a deal nowadays without the use of a virtual data room? In one of our earlier episodes, we actually spoke to an expert, Scott Moeller, who spoke about his experience of, you know, in his early days of working as a financial advisor, going into an actual data room and sifting through, you know, stacks of paper. If you look at sort of what we're dealing with now also in terms of the volume of information that people sift through when you are in a due diligence process, could you even do that without a data room?
PRAKASH KANCHINADAM: It's almost impossible to even think about how a deal would have been done 50 years ago because of the sheer amount of data which is now being exchanged as part of these transactions. As the deals were done, 50-plus years ago in a smoke-filled back room, exchanging documents in a lot of boxes. That clearly is impossible in today's age and time in age. Today, given transactions are more complex, given that there's more and more data being involved and there's a lot more involved in due diligence. Technology is where it's the differentiator. Looking back how bills were done many years ago.
NOAH WEISBERG: We're making me feel very old here because I have been in diligence in real data rooms when I was a lawyer. I'm definitely not as young as I once was, but I hope it wasn't 50 years ago.
CATHERNE FORD: I'm sure it wasn't 50 years ago, but I mean just bringing you into the conversation there then I mean your experience if you think back to what you did then and what virtual data rooms can offer now I mean do you think that you could do a deal from your perspective without a virtual data room?
NOAH WEISBERG: Sure. Yes. You could do a deal. You can do anything. Anything is possible. You send the lawyers there. They'll read the stuff. You send the accountants to... I went to St. Louis. Places like that. You go. You stay. Pittsburgh. You go to the room. They bring you the documents. You read the documents. you type out some notes on your laptop that you brought or something, but maybe you didn't even have a laptop, then you can do it. Like it's totally possible to do it. It's just the thing is it's going to take you a lot more time, which will cost a lot more money.
NOAH WEISBERG: Or what it'll mean is you'll just get to look at less stuff, right? Like you'll have a fixed amount of time and you'll just be able to look at less. And so the actual like really interesting way to turn this to the subject of our talk, which is AI, is just that AI takes that one step further. In the sense that if I had a virtual data room with no download permissions, or I had to print everything off, I could do that work. I know I have done that work in a physical data room in Pittsburgh.
NOAH WEISBERG: It works. You can do it. You could still do it today. It just means you're going to look at less. It's the same thing as if you use a data room without using AI tools. You can do it. You could totally do it. In fact, a lot of stuff happens that way, but it just means you're going to have less ability to see what's in there, and so that's going to have a cost. There'll be a financial cost because you'll be going slower, so they'll be more expensive, or it'll cost the same, but you'll review less stuff.
NOAH WEISBERG: You'll get less risk mitigation and less knowledge for the buyer of what they're getting. It's possible though, totally possible in both cases.
PRAKASH KANCHINADAM: Yeah, one thing to add to that Catherine is, um, influence has been in business for about 30 plus years, right? That they were the pioneers in, uh, launching, um, VDRs. And one of the things, uh, you know, to consider his security and privacy of these deals, right? Yes. The complexity, uh, involved with the data, the number of documents to be reviewed. If you were to look at something like a VDR, it actually is this built-in security mechanisms. And maybe you can tell me that bankers are very paranoid about what they're working on and they don't want anything to be let out when they're actually working on a deal.
PRAKASH KANCHINADAM: And security and privacy is definitely part of that equation.
CATHERNE FORD: That's really nice. My next question was going to be what are the advantages of using a VDR and not just isolated to the ability to actually process that much more information. But I think the security element is obviously really crucial there as well. Is there anything else that sort of springs to your mind when you think about the advantages that a VDR can offer?
PRAKASH KANCHINADAM: Yes, it is the community which comes together when you're doing these transactions. Obviously, you're not getting together at the back of the room. This is a more virtual environment where experts from all over the world may be involved at various stages of the deal. Then this actually provides that virtual data setting where you can bring in various number of experts and subject matter experts as the deal progresses. It's just a sheer volume of the data can be handled in a more secure and
CATHERNE FORD: great. Thank you very much. Noah, I'd like to bring you in. And as part of this conversation around how technology is shaping due diligence, spend some time talking about your favorite subject, which is AI. What tasks on a very basic level, Noah, can AI fulfill in a data room?
NOAH WEISBERG: So my favorite subject is probably my kids or something like that. I think the core thing that AI can help you do is sort of sort, so if you're thinking about setting up a virtual data room, AI can help you sort the information into categories like into folders. it could maybe help you rename documents perhaps, might even be able to help you redact those documents, perhaps do other setup things for you in terms of setting up the virtual data room. The second thing it can do is it can help people process the information that's in those documents.
NOAH WEISBERG: So there can be vast amounts of information in a virtual data room, and what the AI can do is it can help you find some of the stuff you really care about where you need to be paying attention. So for example, one thing we've got lots of experience with is helping find important parts of contracts in a virtual data room, right? Like one thing that's super critical if you're buying a company is knowing if they have any contracts that have exclusivity or non-compete or non-solicitation or most favorite customer obligations in them. and it can be really hard to find those like those can be buried they may not be in one of the sort of top 10 most material contracts and what the AI can do is it can review certainly the top 10 contracts but also like the thousand other contracts that are in there and help you process that massive volume to
NOAH WEISBERG: find like those restrictive covenants or things like change of control and assignment so it can help with that and finally it can help you deal with what comes after help you, like some of this wouldn't even be the most sophisticated technology, but one thing you need to do as a lawyer after you've found these contracts is like you actually have to send notice or get consent from counterparties and forms of AI can help you generate those documents. They obviously can't sign off for the other party on them. but they can help you put that stuff together.
NOAH WEISBERG: So I think there's a lot of manual tedious work that goes into setting up a virtual data room as well as helping people process through the virtual data room that AI can help you with. And I think it's kind of exciting because it can accelerate some pretty miserable parts of the process. I think it can also help you answer questions about what's in the virtual data room too. And I think that helps accelerate a pretty miserable process, but also helps you spot stuff that you just might not be able to if you weren't using AI.
PRAKASH KANCHINADAM: Yeah, I can jump in a bit on that. The intro links have started on that journey about six or seven years ago into using AI into the BDR process. One of the things is to eliminate mistakes, like when it comes to organizing content in the VBR, what Noah alluded to is what we call the classification of documents, which is one of the early cases of AI we have actually gotten to. So when a seller organizes content for the buyer, You know, the seller obviously wants to put the right documents in the right place. And if you throw in a few thousands of documents, AI can take care of, you know, classifying those documents into different categories and then organize them into folders.
PRAKASH KANCHINADAM: And then the other one we actually worked on is using AI to redact private information. There are cases when the seller wants to redact sensitive information out of these documents and AI does a pretty good job of doing that in a very efficient way.
CATHERNE FORD: Great. Thank you to both of you for all that really, really interesting information that you've shared with us. We're going to take a quick break now and then we're going to come back with a look into the future and what AI may be able to do for us in the future.
CATHERNE FORD: Well, we're back and we want to talk about the future and what AI might be able to do for us in the future. Noah, let me come to you first of all. Can you tell me a little where you see AI going? What role it can play in the future?
NOAH WEISBERG: So I think there's a lot of opportunity for AI, again, to save people's labor and to help them find things that weren't there before. So on the labor-saving side, I've been an M&A lawyer, but I've actually sold a business as well. And it's a pretty miserable process. Even with a virtual data room, setting up the virtual data room is really painful. Helping respond to the questions from bidders is really, really, really painful. And then preparing disclosure schedules and notices and the like also can be really, really, really high effort. And I think there's real opportunity today for AI to help take that labor and reduce it down.
NOAH WEISBERG: So that's on the sell side. On the buy side, I think there's lots of exciting opportunity to either reduce the amount of time spent doing diligence or to keep the amount of time spent the same, but it really, increase the scope of what people are looking for. I think that's the flip on the buy side of what AI can do. There's lots of things we could say like AI could look for maybe fraud or look for big risks that might be an issue. I don't know if Russian connections are going to be a problem anymore or not. Like are there connections to Russia or are there state taxes that are unpaid or things like that like I think you can do that but I think it all comes down to two overarching categories which is number one reducing labor and number two increasing the scope of what you're able to look at.
CATHERNE FORD: can I bring you in and maybe have you talk a little bit to the question of where the limitations of it are? Because obviously, I mean, you know, it's wonderful to think that all the tedious stuff can be taken away and it doesn't need to be done by a human anymore. But there are obviously limitations to what AI can do as well.
PRAKASH KANCHINADAM: Yeah, just to touch on a little bit about what AI can do on what Noah was saying earlier, it's about productivity gain. So how much productivity gain can we bring in using AI in the laborious intelligence phase, for example, in the end of the day cycle? The way I look at it is when we started with AI, just even a few years ago, it's AI assistance. AI was an assistant. Now, if you look at it even the last couple of years, it's been a collaborator. Now you could be, it's another person maybe helping us in the deal making.
PRAKASH KANCHINADAM: But the way I look at it, how it's going is when you look at agent, AI, and things of that nature, that's where some of the limitations may come in. What if it makes a mistake? Who is to check on those things when AI goes wrong? So all of those are potentially can be happening if the data behind these models are erroneous. If one of those steps in those agent take flows is somewhat mistakenly going in the wrong direction. So who is to check for those things? So those are some of the risks involved when it comes to AI.
PRAKASH KANCHINADAM: But today when we introduce AI in our products, it's always the human in the loop. So I think that is something which is very important to maintain.
CATHERNE FORD: OK, thank you very much. Now, one of the themes that is emerging in this series of podcasts is that the sheer amount of data, the sheer amount of information that is accessible to deal makers is increasing more and more and more. And obviously, AI is going to be incredibly important. when it comes to dealing with that particular aspect of an M&A process. The thing that I'm particularly interested in is where do you draw the line? This is something that I need to know. This is something that I don't need to know. Can AI play a supporting role or is that always going to be a human element to making that decision about what do I need to know and what don't I need to know, particularly in the context of this ever-increasing amount of information that's available in a virtual data room?
CATHERNE FORD: maybe I can come to you first on that one, but no, I'm really keen to hear your perspective on that as well.
PRAKASH KANCHINADAM: Yeah, I think, you know, every transaction is different, but there are certain elements of the transaction which are pretty routine, you know, maybe done by some of the junior bankers. This opportunity is there to automate using AI. I think those are the, you know, the low hanging fruit, if you will. And that's this definitely opportunity to use AI. so that it gives us the productivity savings we're looking for. It's the next step which is somewhat that we're careful with. It's the human decision-making which should not be replaced by AI-based decision-making. That's something I would be watching out for.
CATHERNE FORD: Okay, thank you. Noah, can I have your perspective on this?
NOAH WEISBERG: So, the way that I think about this is it's all about materiality, like how much you care about being right versus being wrong. If you don't care that much, like for example, because you think you just have a little risk, your risk is mitigated through some structure in the deal. the risk. But if there's some bit of the deal where you feel like your downside is limited or there isn't too much risk, maybe you have a lot of trust or something like that in the in the company that's selling itself, then in that case, you can just cut down your diligence anyways, like whether or not AI is involved.
NOAH WEISBERG: And I think where the questions come up is if you think it's a higher risk situation for you because it's a really big number, like I used to as a lawyer do, I did some work for GE and GE was like, really fastidious because they felt like they had a lot to lose, right? Like they were doing massive deals and they had a whole other enormous enterprise that could really be harmed by one bad deal. So they had a different materiality threshold and maybe a PE who was buying a portfolio company that they were going to keep standalone and isolated from the rest of their portfolio.
NOAH WEISBERG: And the way that they thought about materiality was just different. And I think AI is just part of that. Like it's a tool that can help you go faster. as Prakash said, it's not perfect. It definitely makes mistakes, and that can be okay depending on the specific documents you're looking through and the specific context of the deal. There could be situations if I was buying a company for $10,000 or something like that, I might not care. I might be okay to just put all the documents into a tool. chat GPT or Harvey or something like that.
NOAH WEISBERG: But if I really needed to get the results right, I might need to use a different level of effort, right? Like actually put them into a tool where I had a lot more confidence in the accuracy, spend a lot of lawyer or accountant or you know technology consultant time looking at those results and then checking them out like using those on the other side. And so I think it's always been a materiality determination and I think that just continues to be the case with AI being another tool that you can use to cut through some stuff but it kind of it doesn't matter like there were always hundreds of thousands of documents that people could review or thousands or tens of thousands.
NOAH WEISBERG: And in fact, in a lot of M&A deals, people reviewed 200 documents before. And I think even though the pool of total documents they might review might have gone up, it's going to depend on budgets and risk appetite, how much they do now. And AI is just a tool in that.
CATHERNE FORD: I really like the approach that you're taking about treating AI as a tool in a whole arsenal of tools that you use when you're doing your due diligence process. Now I'd like to, we're almost out of time and I'd like to wrap up this conversation by talking about something that whenever people hear AI, there is an automatic almost sort of an automatic response to it, which is, well, is there a scenario where humans will be replaced? There's an element of fear in that. And I'd like to hear from both of you. Do you think there will ever be a situation in which AI can replace humans in an M&A process and specifically in the diligence process?
CATHERNE FORD: Noah, maybe I can ask you that one first.
NOAH WEISBERG: Well, this comes back to that same theme that I was talking about before. If it's a situation where you think the risks are acceptable if you get some stuff wrong, then I can totally see how you'd hand off a lot of the work to an AI. You probably would have a human interpret it, but it would totally not surprise me in that kind of situation where you think the risks are limited to hand a lot of stuff off to an AI. Right now, the state of the world is and makes mistakes. Just definitely, definitely makes mistakes.
NOAH WEISBERG: And if you think it doesn't, you probably haven't used it on hard enough stuff or you haven't looked closely enough at the outputs, right? So if you look closely at the outputs, you will see mistakes. And so it just depends how willing you are to deal with those mistakes. Over time, AI has gotten better over the last decade. It's keeping on getting better. And the rate of increase is really good. So I can't say that some point in the future where AIs are doing all sorts of things that they're not going to be able to do a lot more diligence, even to the point of replacing some humans.
NOAH WEISBERG: But right now, I think they're a supplement. I have been building AI for the M&A process since 2011. and our software, the previous software that we kind of built and sold where I had that miserable experience of actually like doing the work to sell the company. In that business, 18 of the top 25 global M&A lead table law firms were our customers, mostly using an unbiased diligence. And I can say that my pretty strong belief is that zero jobs were lost despite the fact that our software made people able to review contracts in literally 50% less time and do it with the same or better work quality, I still don't think that caused anyone jobs.
NOAH WEISBERG: And in fact, I think it enabled some of our clients to even review more documents as opposed to doing less work. And so I think there may be a point in the future where the AI gets so like stunningly amazing that maybe it doesn't need people, but like, As someone who uses the technology all the time and has COMSI PhD teammates, I don't think that point is imminent. And I think as an M&A professional, you got a ways to run before the AI is taking the diligence off your lap.
CATHERNE FORD: I guess that's reassuring for those people out there that were worried. Prakash, can I have some final comments on the question of will humans ever be replaced by AI in M&A and diligence before we wrap things up?
PRAKASH KANCHINADAM: Personally, I hope not. But my answer to that is I agree it will never be fully be automated. I'll give you an example. Earlier on, I think I referred to Redaction as one of the tools we provided using AI, right? We call it AI-assisted redaction. We always built our products with AI as an assistant to the humans in this case. So in the early days of AI-assisted redaction, we fully expected humans in the loop and humans making any in recommending. But fairly quickly, we actually got the feedback from our customers within a few months. We trust AI.
PRAKASH KANCHINADAM: It's doing a great job. We want a fully automated version of that. And then soon after, we give them what we call bulk redaction, which is fully automated redaction. I think what I'm trying to say is it comes to the trust factor of AI. As humans in the M&M deal making, As they use AI in certain areas of automation, as they trust AI, as the accuracy goes up, they will trust it and then we can advance the automation even further. But I don't think it would be entirely automated to any degree. As you might have heard, this agentic AI is in the works in many areas and that would be very much prevalent in the next one or two years and how well it is done with an M&A will actually determine how much automation we can do with AI.
PRAKASH KANCHINADAM: But I don't think that would be 100% at any given time.
NOAH WEISBERG: Well, even what this ties to something I was talking about, which is the materiality of the thing you're looking at, right? So for action, if it's like, you know, a random contract with random, not too sensitive data, and you're giving it to people who have an NDA, who you don't think are that sketchy, you're going to have one standard. If you're giving your customer contracts to your biggest competitor, you're probably not going to be relying too heavily on the highly automated bulk redaction tool, right? Like in that case, maybe I'll take a first crack at it with the bulk redaction.
NOAH WEISBERG: Maybe that'll save you some time. But I would anticipate in that situation, there's going to be a junior banker putting a lot of work into minutely going through those contracts. And then there's going to be a VP at a bank supervising what they did in that type of situation because it really matters if you get it wrong. whereas it's a situation where you can afford to get it wrong, then you can rely on bolted action a lot more and just get it done fast.
PRAKASH KANCHINADAM: And the other side of the coin is we have seen humans making mistakes when AI actually got it right, right? So there's always that, you know, it's in our tests and a lot of times, you know, a customer said, well, you know, this is doing your better job. So, you know, you have to really depends on the use case you're dealing with.
CATHERNE FORD: Yeah. Gentlemen, that's all that we have time for on today's episode of The Dealist. Thank you both so very much for sharing your insights and your expertise. It's been incredibly valuable to all of us. If you found this episode interesting, please subscribe to wherever you get your podcasts from and feel free to share it with your colleagues. With that, I'd like to say thank you very much and join us for the next one. Till then, bye-bye.