Subscribe
Share
Share
Embed
The "manual era" of capital allocation is in its final chapter. The firms still relying on manual data extraction and analysis aren’t failing overnight, but they are falling behind one week at a time. While most of the industry continues to "white-knuckle" through 200-page documents and legacy databases, and manual Excel extraction, a new breed of Agentic Allocators is quietly rewriting the rules. They aren’t just using AI to summarize emails; they are leveraging AI-augmented workflows that intelligently automate parts of their investment and operational processes that were previously impossible to automate.
Hosted by Victoria Sienczewski, CEO and Founder of AuumAI, The Agentic Allocator is the "behind-closed-doors" look at how the world's most sophisticated Limited Partners (LPs), allocators and General Partners (GPs) are actually deploying AI, and the hard-won lessons from those building the systems.
This isn't a series about high-level theory or technical gibberish. Each conversation features industry leaders, forward-thinking LPs, GPs and experts who are rewriting the rules of capital allocation through agentic AI. Expect real-world case studies, tactical frameworks you can actually use, and moments that challenge outdated norms. You'll come away with a clearer understanding of the critical questions every allocator must ask - about data privacy, team adoption, integration, and governance - before investing in any AI solution. If you're tired of the "black box" and ready to evolve your investment office for what comes next, you're in the right place.
Just having the data and putting a tool on the data might not always do the trick. It's also making sure that you know how the various documents are connected to each other. In the environment we're in, with fund documents, where you have a fund agreement, you have site letter subscription agreements, DDQs, information memorandum. You have all sorts of information. It helps a lot making sure that you are able to connect all those various documents, guiding AI also making sure how those documents connect.
Victoria Sienczewski:Welcome to The Agentic Allocator. Today, we're joined by Vilmar Feenstra, partner at Loyens and Loof, where he is part of the investment management practice in Amsterdam. Vilmar has spent eighteen years at the intersection of funds law and market innovation, advising GPs and LPs alike on everything from fund formation, carried interest to co investments, and GP stakes across Benelux.
Vilmar Feenstra:A lot of parties like LPs or GPs might have concerns about the use of AI. And in the beginning, you saw a lot of parties who basically started off saying, you cannot use it, and we wanna prevent you from using it because we just don't know what will happen with the information. Over time, everybody is now acknowledging that that's not a realistic approach, where you need to use it, where you need to make sure that AI is being used in an efficient way.
Victoria Sienczewski:Today's conversation explores how AI is reshaping fund formation, LPGP documentation, and the broader alternatives ecosystem, and what that means for the lawyers, GPs, and allocators navigating this shift.
Vilmar Feenstra:I think it's important to always be aware that the question might not be answered correctly, but also if the answer is correct from an AI perspective, it might not be correct from a legal perspective.
Victoria Sienczewski:I hope you enjoy the conversation. Vilmar, thank you so much for taking the time to speak with me today.
Vilmar Feenstra:Thank you so much for having me. It's it's a pleasure.
Victoria Sienczewski:Before we dive into our topic today, which is AI, wanted to ask you a little bit of the backstory. So you advise LPs and GPs, but you've also actually been in the legal kind of tech space in many cases for a number of different years. You actually created a unique system internally over a decade ago to process subscriptions as well. And so we'd love to ask you to take me back to the beginning. When did AI first appear on your radar?
Vilmar Feenstra:Yeah. So it's quite interesting times, I I have to say, because it's... Indeed, we have been developing all sorts of tools and and and things over time because we're always sort of trying to sort of improve our our service level and and by using all sorts of means. And since technology has developed over time, we always try to use those at the utmost and indeed, most of those things start off with some form of frustration or things where we feel that it's inefficient, either because we have internal discussions or we see that it takes too much time or we get some feedback from clients or investors. And indeed, some time ago, more than a decade already, I think, one of the things we also had was the onboarding of investors, which at that time, and sometimes we still encounter this, it's...
Vilmar Feenstra:It was still in the in the old fashioned way where people were just filling out word documents, subscription forms, either actually had to print it. And it took a lot of time and ultimately also we saw a lot of mistakes being made and people were quite frustrated. So at that time, we were already looking at solutions in the markets, but we didn't found any. So that's where we decided to sort of build something for ourselves. And that's sort of the start of our subscription tool, which, well, has been developed over time, and that also was for us a clear message that we had to sort of accommodate all sorts of ways to improve efficiencies by using technology.
Vilmar Feenstra:And, well, since technology has developed, you saw a lot of digitalization over time. We saw quite some automation, obviously, in recent years, and now I think AI is sort of the next phase we're all living in. And also there, we try to make use of it in most efficient way, and it's quite exciting, I have to say.
Victoria Sienczewski:And it is an exciting time. It's also a challenging time, right, for LPs and GPs, and they're tasked with not only running their existing organizations, but now saying, how the heck do I actually adopt AI? And there's many obviously risks associated with it. And so we'd love to ask you from a legal perspective when you think about LPs and GPs and AI adoption, what are some of the risks they need to consider?
Vilmar Feenstra:Well, I think one of the most discussed risk, I guess, especially if you look at from a legal perspective, is obviously that AI doesn't always give you the correct answer. There's there's this whole discussion on hallucination where sometimes AI actually comes up with fake court cases and stuff like that. So that's that's the obvious thing you need to take into account, but I think it's it's slightly more complex in a way where AI also gives answers which in itself might be perceived and might be correct, but if you look at it from a legal perspective, it might not be the actual complete answer. So, that in itself is a risk if you consider that as sort of the final outcome of the question you have and especially if AI also gives you like recommendations or further input or further sort of steps you need to take and you don't sort of use it as a starting point, but use it as the final outcome of your question, I think that might be a risk. So I think it's important to sort of always be aware that the question might be...
Vilmar Feenstra:Might not be answered correctly, but also if the answer is correct from a... From an AI perspective, it might not be correct from a legal perspective, and and I think that's that's something people need to take into account.
Victoria Sienczewski:Very important. And it's also interesting from perspective. Right? You need the human in the loop. That's critical.
Victoria Sienczewski:And that'll continue, I think, for many, many years into the future.
Vilmar Feenstra:I think so as well.
Victoria Sienczewski:Yeah. But wanted to also touch on confidentiality. So this is something I hear from LPs, for example, all the time. We have a huge amount of data, huge amount of confidential documents. Yep.
Victoria Sienczewski:How do we even think about starting to ingest those into AI systems that we use?
Vilmar Feenstra:Well, I think that's that's a very good point and obviously something we encounter more and more where everybody's struggling with it. And I think it all starts with sort of having a good governance system in place. So just buying or having a tool in itself is sort of the starting point, but that's sort of not not where what you're aiming at. Basically, need to make sure first you have a a proper governance system in place where you basically agree on what you can use, who can use it, who needs to sign off. Obviously also which data you can use and which you cannot use.
Vilmar Feenstra:But by thinking on the governance in itself internally that might help you first of all to have an internal discussion about it, but also making sure that you have an actual decision on what you do with the data you have and how you sort of treat it from that perspective, and then based on the governance you have in place, you can accommodate it by using the correct tools in order to to use AI efficiently.
Victoria Sienczewski:Putting in place the guardrail around the systems that you use, extremely important.
Vilmar Feenstra:It is, and it's something that people tend to forget cause everybody is very much focusing obviously on finding or having the right tool in its place, but having a tool in itself isn't isn't the trick we notice. So, we spend a lot of time to sort of making sure the governance is in place, so everybody within at least our firm is able to experiment because I also think that's important that basically you set the guardrails, as you said, and within the sort of guardrails, everybody is free to experiment and actually find out what makes makes most sense.
Victoria Sienczewski:Wonderful. And would love to dive in. The legal industry is a very interesting parallel for LPs and GPs and and the LPGP ecosystem in the sense of large amount of confidential, highly important documents. Yeah. And so legal firms have adopted tools like Harvey or Legora or Microsoft Legal or Claude for Legal.
Victoria Sienczewski:You mentioned some aspects around governance and putting the governance in place. What are other steps you took internally when thinking about adopting AI tools?
Vilmar Feenstra:Well, apart from having this governance, which is I would say the first step, making sure that is in place. I think another thing which people seem to focus on and we as well is obviously making sure that your data is organized in an efficient and the right way, which obviously takes a lot of time, making sure that the data you want to use is being used in the best way. And I think in addition to that, what people tend to forget is just having the data and putting a tool on the data might not always do the trick. It's also making sure that you know how the various documents are basically connected to each other and especially in the environment we're in with fund documents where you have a fund agreement, have site letter subscription agreements, you have DDQs, information memorandum, have all sorts of information containing all sorts of various data points. It helps a lot making sure that you are able to connect all those various documents, guiding AI also making sure how those documents connect, and that will also help giving you a better outcome, I would say.
Victoria Sienczewski:The context...
Vilmar Feenstra:Is very important. Yeah.
Victoria Sienczewski:Yeah. Yeah. Garbage in, garbage out. So if you don't give it that complete view or structure it in a way that AI and LLMs can actually leverage, you won't be able to get the complete picture.
Vilmar Feenstra:And especially in the fund environment where there are so many documents where it's important to make sure that that indeed the context is clear.
Victoria Sienczewski:Do you think that fund documents are going to have to change in an AI world when you have LPs and GPs both... On both sides using AI to assess, let's say, the the other side?
Vilmar Feenstra:Well, implicitly, obviously, fund documents already provide for some provisions like confidentiality and all sorts of things. I I don't think nowadays you see already a lot of explicit provisions, but I would expect that that will be there more and more, especially as you said, a lot of parties like LPs or GPs might have concerns about the use of AI, and I think in the beginning you saw a lot of parties who basically started off saying, you cannot use it and we wanna prevent you from using it because we just don't know what will happen with with the information. I think over time everybody is now acknowledging that that's not a realistic approach where you need to use it, where you need to make sure that AI is being used in an efficient way. But that's obviously also this sort of triggers the discussion on how you can use it and and who can use it, so I would expect that over time, and that might be quite quite recent or or quite quickly already, I would I would expect that the documentation will contain more explicit prefer... References to AI and and the use of AI.
Victoria Sienczewski:Super interesting. And you also mentioned the adoption curve for the industry where you have, let's say, early movers that are starting to adopt AI more rapidly from an LP and GP perspective, and then what I would say slower adopters. And I'm curious from your experience working with LPs and GPs, what are you seeing from that perspective?
Vilmar Feenstra:I I would say that those who sort of prohibited it entirely, those are obviously lagging behind. So everybody is now aware of the fact that in some form they need to use it, but just making, as as we discussed, a tool available and then just have everybody use it without having any boundaries or or approach, that in itself also doesn't do the trick, would say. So I think finding the right balance in sort of controlling the way how you use AI and still allowing everybody to experiment with it as much as possible, I think that's that's where we see those who make the the biggest biggest steps in a way. And that's that's the exciting time is what we also see. It's AI is available for everybody within an organization and it's now not so much something where only, like, top level can use it.
Vilmar Feenstra:It's especially those more juniors who perhaps also have some more feeling with it because, let's be honest, some of the more senior management might might might be less less reluctant or might not be able to to use it, but we try to have also everybody being being able to experiment with it, and and that's the interesting thing, but you need to set the boundaries because otherwise, you might might end up wrongly, I would say.
Victoria Sienczewski:Yeah. There's some, you know, quite interesting concepts around shadow AI use in organizations where earlier career professionals are starting to go rogue.
Vilmar Feenstra:And that's also... It's it's important that that that it's clear for everybody which tools you can use, and that you can only use those tools which have been approved and everybody is aware of how you can use it, and then within these boundaries, should sort of let people free, and we have seen quite some interesting developments over the last recent years where we just ask people, just mention those things that are most frustrating for you or sometimes if you have been in this world like myself for more than eighteen years, you're just used to doing things because you have been doing it for so long. Whilst if you have people just starting off freshly out of college, they might have some fresh new ideas when using AI and if you let those people use AI, come up with some new interesting solutions. So, we've been using all sorts of new words on reviewing site letters, obviously on fund reviews. So, that's interesting, but also importantly, keep on trying and that's also something which is not always easy and now you see that the traction is coming where AI is actually giving you what people might expect.
Vilmar Feenstra:But for instance, this fund reviews tools which we have been using now for I think a couple of years, I can tell you the first time we build an agent to do this fund review, the outcome was slightly dissatisfying. So, we could have obviously then decided to stop trying to sort of build an agent which might help us, But you see, we have been sort of building on this for now, I think, a couple of years. And I can tell you now that quite recently, we have been building something which is super useful. So, also keep on going is is something that people should not forget and and let people just feel maybe the first time, maybe the second time, and then the third time you might come up with something super interesting.
Victoria Sienczewski:Yeah. I love that. The continuous improvement mindset, obviously, from an individual perspective, from an organizational perspective. And you have so many different vectors, right, that are progressing within AI. Obviously, you have the underlying LLMs and the models that are improving.
Victoria Sienczewski:Yep. You have context and reasoning engines that are emerging Yep. And lots of others. Right? Individual use of prompting.
Victoria Sienczewski:So very, very exciting.
Vilmar Feenstra:Yeah.
Victoria Sienczewski:And, Vilmar, would love to wrap up by asking you if you were to advise an LP or GP earlier on in their adoption journey, what are some of the key steps they should follow?
Vilmar Feenstra:Well, I think it's crucial to first set this governance, making sure that you have these boundaries in place where you are able to sort of experiment freely, but at least making sure that that's done within a safe safe environment, discussing indeed also the context of where this data might be linked, and obviously just making sure that the data you have is is basically stored in a way that that makes it possible to actually use the data in an efficient way. I think that those are the basic steps, but I think the most important is let people free and let them experiment with it, And some people sometimes are slightly scared about these times where people might find it scary to know what will happen, whether they will still have a job in ten years from now. I think we should all be more excited because it's an exciting time where a lot of the things we are currently doing, which as I said, might sometimes be slightly frustrating, you can basically find something to do the frustrating part and focus on the the interesting parts and and and that's if you perceive it that way, I think we can all sort of find new interesting things and make our our work life way more interesting.
Victoria Sienczewski:It's a wonderful optimistic approach to end the conversation on. Thank you so much for your time, Vilmar.
Vilmar Feenstra:Thank you so much.
Victoria Sienczewski:That's a wrap for this episode of The Agentic Allocator. If today's conversation gave you a clearer vision of where the industry is headed or helped you pinpoint exactly where your own process is stuck, go ahead and follow or subscribe wherever you get your podcasts. Your firm's judgment took years to build. Stop letting it reset every time a team member walks out the door. Visit auumai.com to discover the structured digital intelligence foundation your team needs to leverage Agentic AI to its fullest potential.
Victoria Sienczewski:Until next time.