Episode Transcript
[00:00:00] Speaker A: What may be risky to one investor is not risky to another. And that lies one of the reason for the bespoke nature of this. And then different jurisdictions too. You would see a European investor emphasizing certain areas more versus the U.S. investors.
[00:00:15] Speaker B: Hey, and welcome to the Momentum Podcast, the show where fintech meets AI. I'm your host, Stan Altschuler, investor, founder and advisor to some of the world's top hedge funds, asset managers and data companies.
Each week we'll dive into practical AI strategies to help you cut through the noise, beat data overload, build smarter workflows and turn your ideas into results.
We'll explore everything from AI tools for your daily workflows to building data driven growth engines for your organizations. Here you'll discover what's next in finance and how to stay ahead of the curve with AI.
Let's dive in.
Hey everybody and welcome to the Momentum Podcast. Today we have a great guest. Monel Amin is here with us. Monel, welcome to the show.
[00:01:05] Speaker A: Thank you, Stan, for having me on the show.
[00:01:08] Speaker B: Well, it's so great to have you. Why don't you give us a quick rundown, tell us a little bit about yourself, about your journey and how you came to start Diligence Vault.
[00:01:16] Speaker A: Yeah, so my journey about starting Diligence Vault started, I would say, over a decade ago.
So before starting Diligence Vault, I used to work at Citi for several years. But the defining moment that got me to start Diligence Vault was towards the end of my career at Citi, which was in the asset and wealth management where I had a very interesting role where I had risk oversight for Citi as an asset manager and then risk oversight for Citi as an allocator. So there was a great place to sit and see the whole process of how an allocator allocates to asset manager and how an allocation alloc, asset manager manages money for allocators. And what is that process that happens when the two sides meet and diligence on each other? So that was the deep process through which I learned about this industry and led me to start Diligence World.
[00:02:09] Speaker B: And you probably came at this from a very technical perspective, right? Because before Citi used to be a software engineer. So you're a very technical person. Every single time I speak with you, I'm like in awe of just your structured thinking and your ability to analyze things. So you probably, probably came at this problem from like a fundamental perspective, which makes it really interesting. Can you talk, you know, to us a little bit about that problem?
What was it back then that made it so, so difficult, you know, from maybe to the world where folks are not familiar with allocators, allocators, like very large institutional investors, they allocate tons of money to these managers. But what about the process that was that, that lit that light bulb for you and you said, aha, there's a problem here, There's a problem here that needs its own solution.
[00:02:59] Speaker A: Yeah. So I would say let's think about an allocator. Whether it's an endowment, foundation or a bank, where I was, or a pension plan, they're all allocating millions and billions of dollars to fund managers. Whether it's a mutual fund, hedge fund, private market strategies.
And that process of allocation involves a lot of diligence. Right. And I think what I was seeing there is that we used to do a lot of diligence on these managers, pre investment and then from an ongoing monitoring perspective. But as an allocator, you're collecting a lot of information and that's coming to you in 40, 50, 60 page documents and then 20 of those documents.
And they're all coming to in different formats regardless.
[00:03:43] Speaker B: And then for different strategies, offering memorandums, performance exposure sheets, legal documents, all that stuff.
[00:03:54] Speaker A: Correct.
You're sitting there processing the documents. In theory you have a lot of information, but in practice, it's not really a lot of reusable information. We're talking about years before AI became forefront, where it helps you process a lot of this information.
Similarly, on the asset manager side, they are responding to these diligence requests for all of their investors. Right. Prospective investors, even before they become investors, and then ongoing monitoring, once they become an investor again, it is all coming to them in different formats as well. So they are also wrangling with the same problem of responding to bespoke data requests, diligence requests, different lens, different formats, when in reality both sides are exchanging 80% of similar information, but just formatted, worded differently over different time periods.
[00:04:44] Speaker B: And this is because maybe an endowment or a pension fund might ask the same thing differently, like me, here's an RFP that's structured differently, but let's say a fund of funds or a public pension or something else, they kind of want the same thing, but they just ask about it differently. And the manager now has to double, triple, quadruple all the work.
[00:05:06] Speaker A: And I think there's a reason people ask things differently also. It's not just because they love their formats, their risk tolerances are different. Right. So if you think about an asset consultant, they are doing the diligence on behalf of their end clients. So their risk tolerance is going to be more on the conservative side because they are recommending these products to their clients versus if it's a family office, they're probably far more aggressive as an allocator what we've seen, or if it's an insurance company, their risk tolerance is different because these are regulated entities and they have capital charges for certain types of investments and different frameworks as well. So as a result, what may be risky to one investor is not risky to another. And that lies one of the reason for the bespoke nature of this. And then different jurisdictions too. You would see a European investor emphasizing certain areas more versus the US investor. The classic example being, if you think about one of the biggest supporter of climate related diligence sustainability has come from Europe versus us has been leading more with diversity and social aspects of esg. So even if an asset manager receives diligence requests from both sides, in theory there would be very different things and very different emphasis.
[00:06:19] Speaker B: So let's dig into the problem a little more. Is it like a physical problem of how do we securely transfer all this information and it gets to the right eyes or is it more of kind of like an information overload problem? How do we as a manager answer all these RFPs and diligence requests and how. And as an investor or as institutional investor, how do I process and make sense out of all this data? Or is it a bit of both? Can you just talk to us and give us some specifics if you can?
[00:06:48] Speaker A: Yeah. So I think pre. Pre diligence vault a lot of people would exchange this or email. So it's as pure as email. Right. But the problem set has three components really. One is what you said is information overload. So in theory you have a lot of information, limited amount of information, things you can do with it. Second is the operational inefficiency is because we are exchanging a lot of these information in different formats and there's a lot of manual effort that goes in and a lot of redundant effort that goes in. And third is the intelligence debt. You have this information but you don't have intelligence that you can actually use from this.
So I think there are many different problem sets that I was seeing. And the reality is this, right? As an asset manager we would respond to hundreds of allocators or LP requests. As an allocator we were interacting with hundreds and thousands, nearly up to thousand across some of our portfolios managers. But the process of the diligence was still bespoke so imagine one on one, it's so inefficient. You scale it to that level of hundreds and hundreds, that inefficiency just scales up.
[00:07:56] Speaker B: Right. So just take to take me through it to drill even a little more. Obviously emails are extremely inefficient, but technology has come a long way now. You've got data rooms, you've got, you know, players like Snowflake, I mean, you know, even stuff like Dropbox and forgive me if I don't know if people have used that anymore or not. I don't know. But there's. People have moved beyond email and is there still a problem if you can establish like a data room that's fairly secure or a data lake that's got all types of different information where.
Why not just do that?
[00:08:30] Speaker A: Yeah, I mean, data rooms exist and lots of people use that. Right. So I think one of the biggest ones in the industry that we see is Intralinks and then followed by Efraim products and Sunguard products. So they are there, but there are 20 or 30 of them that we see being used in the industry. But data rooms typically tend to hold documents.
They don't understand relationships between data points, questions, answers, documents, downstream analysis. So data exchange is one part of the problem, but what do you do with that? Data is the second part of the problem. And diligence is not about documents. It's about understanding having a process, managing the process. All of that can't happen just in a data room. Data room can be a form of dissemination of information for managers. But the process that happens afterwards and the data that people review, the audit trail that you need, the decision backups that you need, is not something that's in a data room.
[00:09:27] Speaker B: Right, that's great.
Take us one step further. What is the technology that you use to help your clients and to help these institutions actually get to that level beyond the data room? How do they get to the point of like, okay, now this is a structured process. Now I understand these relationships.
What did you actually have to build to get together them to that point?
[00:09:53] Speaker A: So I think there are two things we have to build. One is the technology and I'll talk about that. But second is network of managers. Yes.
[00:10:01] Speaker B: Yeah.
[00:10:02] Speaker A: Because if you have to solve this problem, it cannot be an internal use only software. It has to really help the managers as well. Because why would managers use an allocator's technology if it's not benefiting them? Right.
So when we started out, we had zero managers who were on the diligence Vault network. So the technology was taking what a process. The allocators had typically of say, sending out a word questionnaire, requesting an Excel data set from the managers, requesting 2030 documentation support, the ones that we talked about earlier, legal docs, offering docs, policies and procedures, pitchbooks and all of that we digitize onto Diligence Vault. So instead of people exchanging information through Word and Excel DDQs and then attaching as a file of documents, Allocators can digitize that on Diligence Vault. So think about it simplistically like a Google form, but for the industry.
And of course, we've taken it further way beyond a Google form or SurveyMonkey. But you digitize it on the platform and then you disseminate. Disseminate, Disseminate, yes, the form on the network. So on Diligence Wallet network, so that managers can populate the responses directly on the platform, attach the documents directly on the platform, and then send it back to the allocators.
[00:11:30] Speaker B: That's excellent.
[00:11:31] Speaker A: And the advantage the managers have is once they've done this, they now have a repository of all of this information on telusiansvault. So when the next allocator comes in, they can reuse it. And that is where the network starts building value.
[00:11:45] Speaker B: Amazing. Interesting. I did not fully appreciate that other piece of the network. With allocators, you know, we, we had a business also that sold into allocators and managers. And with allocators you could kind of convince them to use a software, right? If it really, if they improve their process. With managers, I found it was a little more difficult. You know, they kind of very married to their systems they live in like Bloomberg and whatever that they do all day long. Very, very difficult to get them to change their process in any sort of way. Or at least that's been my experience. So they must be seeing some value in order to do it. Like, hey, this is, this is making our lives easier. It's like saving me time. It's more efficient. Maybe I can search for things.
Can you talk a little bit about that, like on the manager side? Because the way that your model works is your clients. Your clients are actually allocators, right? Your clients are getting most of the value here, but the auxiliary benefit is for the managers. Can you talk a little bit about the managers, getting them over that hurdle, like, hey, yeah, this is worth your time. You guys put down the Bloomberg terminal for a second.
[00:12:51] Speaker A: So two things. Yes, initially the clients were allocators. Now we also have as a manager in the client. So I'LL touch on that a little bit.
Okay, but so the advantage of the managers was one, when they're responding to a 20 page DDQ with just questions in a word document, oftentimes it's not one person on the IR team that has all the answers.
[00:13:15] Speaker B: That's a team.
[00:13:15] Speaker A: They lean on subject matter experts internally, portfolio management team, finance team, legal team, hr, it, legal compliance.
And ultimately when they end up finalizing this document before they want to turn it off to the investor, it's probably version 27 of that document. So you can imagine the, the marathon they have to run internally.
[00:13:37] Speaker B: Yes.
[00:13:38] Speaker A: Managing all of these. Right. You take this process onto a platform like Diligence Vault, Multi user authoring. Right. People can collaborate on the platform anywhere in the world.
[00:13:50] Speaker B: Right. One could be in London, one could be in New York City.
[00:13:53] Speaker A: Correct? Yes. People can log in when they need to, they can work on it. There is no conflicts of version controls issues. So that is another advantage of the managers accrue is just the collaboration aspect of it.
And then the reusability is a second aspect.
Fast forward to today. AI comes along.
[00:14:13] Speaker B: That was my next question. I'm sure it's changed so much. You've been doing this for a decade now. Everything is like getting turned upside down in this AI booth.
[00:14:22] Speaker A: Yes, yes. So I think AI has helped in so many different ways is now people can realize value so quickly. I think pre. We used to have clients do two weeks, three weeks or prospects do two weeks, three weeks trials to see if this works for them. Now they come to that buying decision within a day.
[00:14:42] Speaker B: Wow.
[00:14:42] Speaker A: Right. As we see the difference.
[00:14:44] Speaker B: And is it because they can just use AI to interrogate the data and ask questions of the, of the, of the data very quickly.
[00:14:52] Speaker A: That as well as we've integrated a lot of AI so they can see the output immediately and they can see the efficiency difference immediately.
[00:14:59] Speaker B: Also you're using, you're using the actual tools inside Diligence Vault is what you're.
Oh yeah, talk to me about that. That's cool.
[00:15:06] Speaker A: Yeah. So as an example, we had a GP trying to answer 268 question questionnaire.
Smaller GP on diligence world from one of our clients.
[00:15:20] Speaker B: Right.
[00:15:21] Speaker A: And we gave them access to our AI to try and 80% of that questionnaire was completed once they uploaded the documentation in under 15 minutes.
[00:15:33] Speaker B: The AI already knows a lot of that stuff. It could just fill in what it knows correct.
[00:15:39] Speaker A: From their documents. Right. So it's not making up stuff, it's taking their offering docs, their PPMs, their pitchbooks, their legal docs, their LPAs and pre populating an investor questionnaire, given the size of it, it would have taken that team at least two weeks. Right. And they were able to realize that value so quickly.
[00:15:58] Speaker B: So do you control for hallucinations at all? Are people concerned about that? It's going to make stuff up. Sometimes there's a lack of knowledge. It just fills it in confidently with made up stuff, which is my favorite part of AI.
[00:16:11] Speaker A: So. Absolutely. I mean for us our AI philosophy is always quality before speed is because I think it's very diligence is a high trust process.
[00:16:21] Speaker B: Yes, yes.
[00:16:22] Speaker A: If you end up making a mistake there, people are not going to use it and you create a bigger review burden. If you don't control for quality.
[00:16:30] Speaker B: That's the problem.
Yes.
[00:16:33] Speaker A: So the way we've set up AI is that it gives you obviously source and just the AI reasoning is table stakes for everyone. But we also give confidence indicator. Are we confident response is good or we think there is limited confidence and the reasons for limited confidence. Reasons for limited confidence could be, hey, what you're seeing in a PPM doesn't match your pitchbook.
[00:16:58] Speaker B: Oh, there's two different sources. Yeah.
[00:17:02] Speaker A: So you could verify that or within the same document you are saying two different things about the same fact. So you might want to verify that or we don't have enough information to answer the question. Right. So that helps even the whoever's using AI to be more faster in the whole review process so that they can have a much better outcome. Right. So I think that's. So that's the benefit for the asset managers. It's not just that they have to use this one more portal, but it makes all of their processes and workflows easier. And AI has been such a big deal.
[00:17:37] Speaker B: I mean that's the big one.
That's the unlock at least in my mind for managers. If they can have a model that's smart and knowledgeable about just their own stuff, like it knows their a, it just knows their basic terms.
You know, it can fill in the RFP like to, to a large percent. Right. That is a huge unlock for managers. I mean frankly they would, they would pay for that because that's time and these people are paid, I don't know, a million dollars a year. Right. Like they're paid a lot of money for that time. So that's, that's valuable, valuable stuff. That's great.
[00:18:15] Speaker A: Yeah. And I think now on the asset manager side, I think to answer your question, we do have asset managers as clients. Now we have had clients sign up for a demo and close within the same day.
[00:18:27] Speaker B: That's crazy.
[00:18:27] Speaker A: That is unheard of in our industry.
[00:18:30] Speaker B: Absolutely.
[00:18:30] Speaker A: But that happens now because AI is such a big tailwind and once they see the output, they love it. Right?
[00:18:36] Speaker B: Yeah. So, yeah, I've never. That, that must feel so good. I've never had that. But you know what that tells me is that there is a real pain point that these folks feel. And the AI was, was the unlock here because now they could see, visualize very quickly in their minds how this pain point goes away or completely resolves.
That is awesome.
Is there anything that AI shouldn't do in operational due diligence? Is there anything that, you know, you say, whoa, whoa, hold on. This is actually better left for the judgment of a human being.
How do you think about that at all? Or do you just let the agents go at it?
[00:19:18] Speaker A: I mean, we have put in the boundaries. Right. Of what it can. I think AI is great with. If you're thinking about from. From both perspectives. We also have AI solutions for allocators is when there is a basis for it to create an output.
AI should do the first try, no matter what.
[00:19:41] Speaker B: Give it a try.
[00:19:42] Speaker A: Yep, give it a try. And then obviously we have the guardrails to ensure that the quality is there. And if the quality is not there, we would say, okay, here's the reason why it's not there.
There are areas where you need creative and new thinking, where you don't have basis for input.
Those are the areas still very much human oriented. Right. If you think about just the work of delusions and the workflows.
But the biggest thing is also the personal touch. Right. When our allocators go and see managers in their offices, when they do on sites, AI is not doing it. They are looking people in the eye, they're verifying the process. They're having conversations with multiple different team members, validating what person A said to person B. The team dynamics, what's happening within the team dynamics? Is everyone just listening to what the portfolio manager is saying or do people have their own.
[00:20:36] Speaker B: Is there dialogue?
[00:20:36] Speaker A: Yeah.
So that I would say is a very much a human judgment.
[00:20:41] Speaker B: Yes, yes.
[00:20:43] Speaker A: So I think. And then frontier risks are also going to be human judgment. Because the reality is LLMs are going to learn from patterns. If it doesn't have a pattern to look back to, it can be creative, but can you trust that?
[00:20:57] Speaker B: It's tough. Right. So creativity is a feature of the lms. It's really Built in to the way that these models reason.
But in our field, in finance, in your field, due diligence, where accuracy is just so important, that could be a bug, not a feature. So I'm sure you have to be very, very thoughtful on the guardrails and the checks and the audit trail and all that sort of stuff like source matching, things like that.
[00:21:23] Speaker A: Yep, absolutely.
[00:21:25] Speaker B: So this is great. And if you don't mind, step into the hypothetical world a little bit with us and look ahead, where does this, like, how does this evolve, you know, as this technology evolves at a mind bending rate. Right. These new models are coming out.
It used to be monthly, now it's, it's like it's 5.4, 5.5, 5.6. You know, where does this go? Does this, I mean, first of all, does this like eliminate maybe the need for investors to hire human money managers? I don't know. Or, and if not, if it's still going to require a human touch, as you say, Right. How does this change that whole dating process? Right. How does that change the dynamic that happens ongoing between the investor and the money manager?
[00:22:16] Speaker A: Yeah. So I think even before we talk about AI, I think fundamentally, right. If you think about due diligence, investors need due diligence to understand can this manager continue to earn return. For me, they're also looking at risk mitigation and some of those aspects of identifying risk can be done through AI. But then also how do you identify new risks? How do you optimize for the risk? A lot of that would need human judgment. On the flip side, during the diligence process, managers incentives are to position themselves as the best alternative. Right. So if you think about both these parties are coming to the diligence table with very different objectives.
And sure, over time AI can extrapolate for different objectives, but I don't think that's going away anytime soon. I think what AI is really helping today is automating the administrative burden. Getting you ready for your first drafts of your internal presentations, helping you vet out managers that you think would not work for you long term, identifying patterns with your portfolio and with any new managers that you're looking at in real time, identifying risk flags and areas of concern in real time, both from public domain and from the information that you have.
So it's really providing a leverage, a massive leverage to the research teams, the diligence teams. It's not necessarily at a point where it actually replaces their judgment.
[00:23:51] Speaker B: Of course that makes a lot of sense.
So I want to wrap this up this has been eye opening for me and you know I always enjoy our chats and I do really appreciate this. Monel. If you could tell for our listeners we have several institutional investors listening. I'm sure a lot of managers listening.
If they wanted to find you and learn more about Diligence vault and talk to you, how do they do that? Do they reach out to you on LinkedIn? Do they go to the team? Do they go to the website? What's the good channel for them to get in contact with you and just talk a little bit about due diligence with you?
[00:24:28] Speaker A: So I think nowadays the easiest is to go ask any of your LLMs, how do you get in touch with diligence?
[00:24:34] Speaker B: Nice. Good answer.
[00:24:37] Speaker A: Or what's the top diligence platform? And the first thing shows up as diligence falls, you click on it, takes you to our website and they can just fill out a contact us form.
[00:24:47] Speaker B: Beautiful.
[00:24:48] Speaker A: That's it.
[00:24:48] Speaker B: Amazing. Manel, thank you. Thank you so much for being a guest in the Momentum podcast. Really appreciate your insights.
Thank you and everybody, thanks for joining us. We'll see you on the next one.
Thanks for tuning into Momentum. If you found today's episode valuable, don't forget to follow the show and share it with your network. This helps us a lot.
To explore how AI can help your firm, visit Acadia Im.
I'm Stan Alchiller and I'll see you next week with more practical insights to help turn data into growth and decisions into results with the help of AI.
