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 Expect real-world case studies, tactical frameworks you can actually use, and moments that challenge outdated norms.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.
The scarcest resource we have in the firm is the time of senior professionals. And I think the more time that they can be spending doing more value add work rather than formatting, making charts, spell checking, that kind of thing, that's where we see kind of the biggest wins from AI or from a process efficiency standpoint.
Victoria:Welcome to The Agentic Allocator. We are joined today by Harry Sepulovits from Aksia. Aksia is a leading alternative investment specialist with over 360,000,000,000 in assets under supervision and a partner to many of the largest and most sophisticated allocators globally.
Harry:It was pretty clear early on that we'd be able to leverage those tools to a great extent for work. Once we got comfortable with a lot of the enterprise solutions that were out there, we were pretty much off to the races.
Victoria:In today's conversation, we are moving beyond the theory of AI and talk through concrete stories from across Aksia. How different teams are actually using AI to enhance their research, client solutions, and back and middle office functionality. What has worked and what has not, and what that signal for how allocators can use AI to sharpen decision making and strengthen their organizations.
Harry:It's been hugely helpful both from an initial screening perspective as well as from a process perspective. You can quickly go into their data room, put all the information into the AI tools we're using, and very quickly find out exactly every deal they've done in their history and how it's stacked up across the 100,000 plus deals in our database. So you can quite very quickly get a sense of whether a manager is interesting, whether they've been doing things kind of towards the upper quartile of their peer group, or whether it's less interesting.
Victoria:I hope you enjoy the conversation. Harry, thanks again for taking the time to speak with me today.
Harry:Yeah. My pleasure. Pleasure to be here.
Victoria:Would love to start off by asking you when did AI adoption become a strategic priority for Aksia rather than just an interesting thing to consider?
Harry:Yeah. I think we were pretty quick. I think we've got a lot of people in the firm that are pretty tech savvy and were pretty early adopters of a lot of the free AI tools that were made available to, you know, the general public in 2022, 2023, around then. And so it was pretty quick for us to be able to realize the benefits that we can use for those tools once we got comfortable from a data security standpoint. You can appreciate that we work with, you know, well over a 100 clients that have holdings in 14,000 plus distinct entities, and we are diligencing hundreds, if not thousands of opportunities per year.
Harry:So, it was very important to us to be able to protect the confidentiality of our clients, of the managers that we're underwriting, and once we got comfortable from a data security perspective that we wouldn't be breaching any of our obligations, and I think we were pretty, pretty quick to be off to the races.
Victoria:Incredible. And so how are you leveraging AI internally through your investment due diligence process?
Harry:Yeah. There's been loads of ways. I'd say, you know, one aspect is from an initial screening perspective, and I guess the other is more from how it helps us carrying out the diligence itself. I mean, a screening perspective, being able to leverage our data that we have on private markets more broadly has been hugely effective. So, we've got database internally we call Deal Vault that our clients have access to as well.
Harry:It's got information, I think, on more than 150,000 individual private markets deals that have been done over the past twenty plus years in companies that we're monitoring over time. So when you, for instance, are first looking at a new fund and that you have intelligence in the past, being able to extract all the information from the data room, upload it into our AI tools, and map out every single deal that a manager has done in its entire history across region, industry, sector, and how has that performed relative to all the 150,000 plus deals that we have in our database is really helpful in being able to get to a no pretty quickly. I think once you find something that's interesting from that initial lens, it allows you to progress through the much more time consuming element of your DD process, which maybe some of it you can't leverage AI as much when it comes to on-site visits, reference checking, things like that. But it's very helpful at being able to get to know, I think, more quickly and do an initial view on managers and opportunities very quickly. If I think through co invest, I think it's equally, if not more valuable.
Harry:If you're looking at a specific company, the first thing we think about when we're shown a co invest opportunity is what is the manager's wheelhouse in executing these types of deals? So if it's a European healthcare company, I want to be able to see every single deal this manager has done in European healthcare, how it compares against every deal in our database that's been done in European healthcare over the past twenty years. Is this an area where they've excelled, have great performance in those types of investments? Being able to use AI to come to a determination very quickly to decide whether a deal is worth pursuing, you know, is hugely valuable from a timing perspective, but also makes you very attractive from the manager's perspective. Cause managers really appreciate when they're showing these types of opportunities, whether or not you say yes or no, of course they want you to say yes, but being able to say no quickly is often just as valuable.
Harry:And so I think leveraging it in in the kind of initial screening process has been hugely helpful.
Victoria:Phenomenal. And would love to understand, how are you thinking about leveraging AI to unlock the collective intelligence at Aksia?
Harry:Yeah. That's, I think, one of the most exciting things about these tools, that it's removed any kind of barriers that existed between different teams, between different asset classes within the firm, and it's also just made the twenty plus year institutional memory of the firm available at everybody's fingertips. So, for instance, if a client asks me, it's been certainly topical over the past years, how should we think about FX hedging within our hedge fund portfolio or private equity portfolio? Sure, I've got a lens through, you know, the different clients I've worked with in Europe, you know, that have thought about FX hedging, how they have different ways of implementing. But now I've got the institutional memory of how every client has done this, whether it's a Japanese insurance company, an Australian superannuation fund, a European pension, even US institutions when they're investing in non US denominated opportunities, to be able to have all that at my fingertips and be able to quickly revert to a client with very specific use cases and how different people have implemented FX hedging, think has been really powerful.
Harry:The other way it's been really interesting is, like I said, taking down barriers between teams. So if I think even within a single team, if you're a hedge fund analyst and you're underwriting a manager and you want to look at some of their, you know, best long ideas, you can put that ticker into our AI tools and it'll quickly tell you every single manager in our universe that has been both long and short that company in the last five years. So now when you speak to that manager, you can bring up all those bear cases and see, have they thought through these? Do they have thoughtful responses for, you know, why they're comfortable with those risks with respect to that investment? So I think that's been really interesting.
Harry:If I look at a specific co investment opportunity, and I want to see how every other private credit manager in our universe has approached that specific company, who's in the capital structure right now, you know, what are the terms in which they've invested? Going across to the equity side, what are the different private equity managers that we have in our system that have invested in that company? What is their thesis? Is there a different story on the credit side than the equity side? And we've always tried not to have barriers within the firm between the different teams, but being able to have all of that at your fingertips now to just type in the name of a company and see the full wealth of information we have across asset classes, across teams, has been hugely valuable.
Harry:I mean, one other area that just comes to mind as well is on benchmarking terms. So, we've got a library of how many tens of thousands of PPMs and LPAs that we've reviewed, And every time we're reviewing one of these documents, we think the value add is not so much summarizing what the key terms are, but identifying what's off market, what's non standard, and being able to do that across our entire database instead of relying on individual people to see a term and think, oh, that's something unusual, but to quickly be able to identify what's off market, what's different, what we haven't seen, and what have been trends over time in certain terms have been a really fascinating use case for AI.
Victoria:So cool. And I see a world actually in which, we're not there necessarily yet, but where you go into a meeting with a manager, for example, and your AI analyst or assistant is able to flag real time for you based on your historical data, based on all of the information that you have on other managers their positions, let's say inconsistencies in what they've said or interesting nuances to dig in deeper.
Harry:Yeah. I think nothing frustrates managers more than when you have access to historical data that runs counter to how they like to position themselves. I say that's pretty common. I think that that happens all the time where our clients like us to go alongside them in some of their diligence meetings, and, you know, a manager will tell them about certain things they do or their historical track record or that they only operate within certain segments of the market, and being able to have hard data that you can query in the meeting. I mean, the tool I referenced, the deal vault tool that has all that data, that's interactive and clients can access that themselves through their portal.
Harry:So, if you're in a meeting with a manager, you can already pull up all their historical deals and say, oh, I'm looking at so and so deal that you've done that's maybe on a, you know, enterprise value that's a bit higher than the range that you said you focus on, you know, looks like it hasn't done so well. Can you tell me about that specific deal? I mean, that's less of an AI tool specifically, but I think leveraging AI to, you know, be able to parse interesting insights from our data is kind of helps it take it to the next level.
Victoria:Absolutely. The data is the foundation, the first step, and then you can leverage it.
Harry:Exactly, yeah.
Victoria:Put AI on top. Super interesting. And curious to hear a little bit more about some of the efficiencies you found AI driving internally so far.
Harry:Yeah, there's been loads, think what's interesting is that these efficiencies extend across all the different teams that we operate. I think some of the lower hanging fruit are probably within the middle and back office part of our business. I mentioned that we have client investments in 14,000 different holdings across a variety of managers and funds, and those funds, you can imagine, generate you know, tens of thousands of transactions on an annual basis, whether it's, you know, capital account statements, investor calls, valuations, etcetera. So, having AI be able to pull a lot of that information from statements to reduce errors and reduce the amount of human time you need doing those tasks has been definitely a huge time saving. Another area is, you know, if you think about how a diligence process naturally progresses, and originally you'll meet a manager, you'll think of funds interesting, you'll spend some initial time doing the work, you know, you'll have, perhaps you'll have a client that's interested.
Harry:The natural progression takes you through investment due diligence, operational diligence, reviewing legal documentation, you're doing a background check, reviewing financial statements. There's all these steps along the way where you can theoretically identify something that would have killed the whole process earlier on. And, you know, an example of that would be, you know, something that comes up in a background check or a company check where you've got a private equity manager that backed some company twenty years ago that was maybe doing something that your client had wants nothing to do with. Before, there was nothing you could do to truncate that process that would have spared you tens of hours of people's time going through this process when you had that bit of information upfront that could have sunk the deal originally. Now, the very first thing we can do if we're underwriting a private equity firm, we can get information on every company that this manager has ever invested with, identify any potential headline risks, flag it to the client upfront, and that could potentially save you loads of time that you would have spent across investment team, operational DD, legal DD, risk, just by doing that very simple check upfront, which in the past would have been something that you're saving if you're paying someone for a background check.
Harry:You're usually going to backdate that process towards the end to avoid incurring costs until you know that, you know, investment is closer to the finish line.
Victoria:So fascinating. And so in terms of the ROI and how you measure it for AI adoption, you mentioned a couple of different elements already in terms of specific tasks that the team was doing from a data reconciliation or number crunching perspective. You mentioned resource and team time allocation, moving the due diligence or some of the risks upfront to identify them quicker. Are there any other elements that you think about when you look at the ROI of your AI initiative internally?
Harry:Yeah, I mean, way we think about it is definitely not about, you know, how can we reduce head count because we're becoming more efficient. It's very much how can we make sure that people are spending their time in the most value add way possible. So, can measure that in terms of how much time of a senior person's time does it take to complete a due diligence report, and how is that time allocated. If you think through kind of specific KPIs that we might be tracking, I mean, just simple, you know, how many input mistakes are you making on our record keeping process? Any sort of input errors should really plummet if you're having AI screen a lot of this data and having just a final level human review.
Harry:This may sound silly, but one interesting KPI is the length of a due diligence report. We have the sense that there will be a natural tendency for these reports to get longer and longer and longer as you can have AI just, you know, summarize information, regurgitate manager materials, just things that are really not going to be value additive to clients. We want our reports to get shorter. People don't have time to read through seventy, eighty page reports. We want our reports to be focused on, you know, data driven insights based on our proprietary data that we have, judgment from our senior people that have decades of experience investing in funds, context for where managers fit in the broader market, I think that's where we're going to see, start to see differentiation, is people that are overusing the AI tools to generate their diligence reports, I think you're going to end up seeing kind of less value add, a lot of fluff in some of the content that they're putting out, and something we're very mindful of.
Victoria:Yeah. I love that. You kind of joke that in some cases you need AI to write it and AI to summarize.
Harry:Yeah, yeah. I think if we're producing a diligence report and a client's first instinct is to then run it through their own internal AI tools, I think we've do it's like rolling. Yeah.
Victoria:And how have you thought about implementing AI internally through a team perspective in terms of how you've organized yourself?
Harry:Sure. I mean, we've been very thoughtful about what we've centralized versus not centralized in the process. So from a centralization standpoint, obviously, selection of vendors, centralized compliance, data privacy, that sort of thing, I think was important to be centralized to be efficient. Other than that, we've really tried to move to more distributed champions across the different teams within the firm from an implementation standpoint. We found that, you know, the technology is progressing too quickly and the use cases across the different teams is so different that having kind of the central working group is more of an impediment to implementation.
Harry:That's something that will really encourage people to use it. So, you know, we've, like I said, we've had champions across the different teams that can be leaders within their teams in terms of thinking through different ways that we can leverage the tools we have at our disposal to make the different teams more efficient and effective with their time. So I think that's been pretty successful. And one other element is just incorporating it into, you know, end of year review processes. If you make it explicit that says when you come into your end of year review, you should be able to articulate how you've been able to leverage AI tools to make yourself more efficient and to deliver a better end product to clients.
Harry:I think that's been a great way to encourage people to really, you know, take AI seriously. It's not, you know, an option for people to use these tools or not. I think they're here to stay, and I think we are really encouraging people on the team to use them, and it becomes part of their evaluation.
Victoria:So interesting. Have you seen a difference in terms of the seniority of folks in in usage, or has it been pretty consistent across the full organization?
Harry:I guess when I when I think through who the different champions are across the different asset classes at the firm, I think it probably has tended to be biased towards, call it, more mid level or even some junior people across the different teams, and I think that's presented a great opportunity for a lot of those people to, you know, from a very early point in their careers, take more of a leadership role in a project that's, you know, incredibly important and strategic for the firm. So I think we often get asked, you know, have you stopped hiring at the junior levels, or what's the value out of junior people? Actually, found that a lot of the best insights and use cases we've had for actually implementing AI across the firm have come from, you know, younger people within the firm that probably were quicker to adapt in their personal lives using a lot of these AI tools, and have been at the front in terms of thinking through, you know, how can we use these tools, how can we manipulate them in a way that's going to be most effective for organization, really thinking outside the box.
Harry:So I think it's, you know, everyone at the firm from, you know, the most junior to our CEO realizes the importance of using these tools to make us more effective, but I think a lot of the younger people in the firm have been at the forefront of of really finding the best use cases.
Victoria:So important both having the management team buy in and support for AI adoption as well as the kind of junior team implementation.
Harry:For sure. I don't think you can have a conversation with our CEO without him emphasizing the use case of AI. Every time there's an article in the Feet or some other publication about firms using AI or or how to streamline their processes, he's quickly forwarding it to everyone and then making sure that everyone has it in the back of their mind that this is something that's very important to him and and the strategic case for the firm.
Victoria:Fantastic. And we've talked a lot about Aksia and how far along the AI path you are. Curious, where do you go from here? What's your vision for what AI will look like within the organization in two, three, five years?
Harry:Yeah. I mean, I think we're we're increasingly trying to measure things by how has it improved the client experience. So, what can we do with AI that makes our product better for clients? And I mentioned some of the KPIs we're thinking about, but, you know, part of the, one of the benefits of being able to get to know more quickly as part of leveraging a lot of these tools is that the volume of insights we can produce, you know, on an annual basis has gone up exponentially. So a lot of that kind of initial work we do on a manager, I mean, following through from initially meeting a manager to completing full due diligence reports usually is a process that involves multiple meetings, building conviction, making sure there's interest from the clients there.
Harry:But if you can, and a lot of that is very time consuming, but if you can spend some of that time to broaden the funnel because you've got the ability to produce initial insights on managers much more quickly because of these tools, I think that's usually additive to clients. So, one of the things we hope to see is that our volume of insights reports that we're producing, so these are more desktop based work rather than full IDD, ODD, on-site, etcetera, we should be able to produce insights at a much larger volume and more quickly. And doing so will really broaden the aperture for clients in terms of being able to identify things that are interesting for them. We already think we're very well resourced to be able to do that already, but to be able to say, we know every institutional fund in the market, we've done initial insights on all of them, that's a lot of it data driven and comparative to, you know, our broader proprietary dataset, you can use that to fine tune what you want us us to spend more time on, where you want us to progress to, you know, the full due diligence work, where we're doing the really value add work around on-site visits, reference checking, etcetera.
Harry:I think that's one area that we'd really like to see take off as as we increasingly leverage our AI tools.
Victoria:And under your leadership, Aksia is far along the path of AI adoption. For your peers, that may be a little slower to to adopt AI. Do you have any advice on how to get started or what they should do first?
Harry:I think it's overwhelming, right? I think there's so much that you can do with AI, and once you go down a path of one thing, it tends to open a rabbit hole of all these other things that you can be doing. So it can be difficult, I think, to be focused on individual tasks and how you can best leverage the tools that are available. So, I think the best way to start is instead of thinking down with a blank sheet of paper, how can I use AI, it's thinking back through any process you're doing that predates the advent of AI? So, if you've got any sort of recurring IC meeting, recurring DD process, standard questionnaires, just taking a step back before you do any of these recurring tasks and think through on that individual task basis, what are ways I can leverage AI to make this more efficient?
Harry:And the beauty is, you don't even have to think about that yourself. You can then take that specific task, describe it into the AI tools you're using, and say, this is what I'm doing on a monthly basis, a quarterly basis, annual basis. These are the key inputs into that. This is how time is typically spent. Why don't you tell me how you can make this more efficient?
Harry:I think AI tends to be much better at telling us how it can be helpful than us thinking about how it can be helpful.
Victoria:I love that advice, Harry, especially focusing on specific use cases, specific tasks. But what I'm very excited about is how AI and the adoption of AI will actually change how work gets done in the end to end process that LPs and allocators go through, and what that'll look like in three or five years.
Harry:Yeah, absolutely. And again, think that is where, you know, for us, a lot of what the value add we've had historically is going to be, you know, the moat for that is is going to go away given, you know, the ability to, you know, meet managers at scale, taking, you know, taking notes across, you know, manager meetings that you don't necessarily have to dial into. So I think having, you know, more boots on the ground was a huge advantage, and we still think is an advantage, but a lot of that moat is going to go away with AI. But where we think the edge is still going to be in that proprietary data. So, because private markets are such an opaque asset class or set of asset classes, you know, the AI tools are only as helpful as the data you have that can feed into it.
Harry:So, I think our data access and the ability to leverage AI to make what we do more efficient, make us better at what we do, is I think it'd increasingly be our source of edge, as opposed to some of the things, you know, reviewing a PPM or identifying, you know, key terms, things like that. I think a lot of that is going to be streamlined with AI.
Victoria:Harry, incredible insights. Thank you so much for joining me for a great conversation today.
Harry:Thank you so much for having me. It's been a pleasure.
Victoria: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 you're stuck in your own process, go ahead and follow or subscribe wherever you get your podcasts. And if you're curious what Agentic AI might actually look like inside your investment office and how to get there without compromising on security or control, visit auumai.com for demos and resources on AI native LP and Allocator workflows. Until next time.