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 Expect real-world case studies, tactical frameworks you can actually use, and moments that challenge outdated norms.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.
You are leaving money on the table by not implementing AI models as they are right now. All sorts of questions have yet to be answered. Cybersecurity and accuracy. But if you are waiting for those questions, then you are waiting for the industry to pass you by before you have the answer.
Victoria:Welcome to The Agentic Allocator. Today we're joined by Alex Harstrick, managing partner and co founder of J2 Ventures. J2 is one of the most distinctive early stage funds in The US, focused uniquely on deep technology for the US government.
Alex:We wanna go home every night and we sleep well when we tell ourselves that we were right today. We don't sleep well when we tell ourselves we were wrong. And you actually need AI from a broader perspective to come and look at all of this data and be like, there is an inherent bias in the way that you are looking at the world. It's that contextual element. It's your feedback plus what you're getting from the algorithm that will help you make better decisions.
Victoria:As a firm, they aren't just investing in AI. They're actually living it. Building it into the fabric of how they source, evaluate, and support companies.
Alex:I've been impressed at how frequently venture capitalists have put very little work into creating sympathy with the counterparty that they're trying to get a hold of. We're not selling stock. We're connecting with human beings that are building their life's work. The more you can create that connection, the stronger it is. This will help you get around that.
Victoria:Today, we're digging into how j two is actually using AI internally. How they're advising their portfolio companies to implement it, and what Alex is telling LPs who are under pressure to figure out AI but don't necessarily know where to start. I hope you enjoy the conversation. Alex, thanks so much for joining me today.
Alex:Thanks for having me.
Victoria:So at j two, you have a unique term for AI which you call advanced computing. What does that signal for how you think about technology and AI internally?
Alex:You know, Victoria, I'm surprised how unique that term is because I think when we look at the lexicon related to AI, it's almost lost all meaning. Right? Like, what even is AI? If you were to say AI versus machine learning, I think that delineation would go over a lot of people's heads. And especially when you think about technology companies that define themselves as being AI based oriented or any kind of depth of engineering within AI, you'll be really, really hard pressed to find any validity behind that.
Alex:And so when we look at what are the actual foundational aspects of how you're creating what people perceive as being AI, that's really an advanced computing methodology that I think, because the public has just not caught up to all of the different types of technologies that feed into this, they're unaware. And so it would be myopic to call a lot of what we're excited about purely AI. And moreover as an early stage VC, you'd miss a lot if all you're doing is looking at AI in the broad definition.
Victoria:Makes perfect sense. And so how are you leveraging advanced computing, a subset of which is AI, within your internal firm operations and from an investment perspective?
Alex:Well, separate from the operations, mean that's our how we're focused from a sourcing perspective. But internally speaking, what I find exciting about AI is there's really no aspect of how we set up the firm that can't be positively disrupted with AI. VC is a relatively niche industry, despite how we all perceive ourselves and the various podcasts that introduce us to make us feel perhaps more significant than it really is in the broad aspect of asset management. But because it's so niche, there are very niche service providers that are incredibly expensive. And so their pricing for everything from a CRM to a research tool to I mean, insert whatever it is, like small aspect, becomes increasingly and precipitously more expensive the more users you have because they know once they have you, they can ratchet up the pricing.
Alex:The problem is the pricing goes up, but the quality doesn't. And AI, in the last year, has disrupted that entire industry. I'll give you an example. I would think you'd be hard pressed to find one person that likes their CRM. Right?
Alex:Externally speaking, I mean, I've talked to bulge bracket providers. I've talked to, like, small VC funds. The best and most sticky tool that anyone has ever used is, like, Google Sheets that are shared, and they look like an Excel document. There's, like, no technology in there. And so a lot of people have made a big industry of selling more and more of these larger CRM platforms, but they're really not that much better.
Alex:I mean, might have some cool integration that's, you know, charges you a ton of money for that integration, it really doesn't deliver the value. AI, you can vibe code this kind of stuff over a weekend and make it bespoke for your firm, for your workflows. That is incredibly valuable because I think at the end of the day, especially in early stage VC, you are, if nothing else, a talent scout. You're trying to find things really quickly, and you don't have time to come and sit back and then interrupt your workflow and say, let me remove sort of the romance from this interaction and like type in like a bunch of notes that I have about the person that I just met. I mean that interruption makes the whole thing feel a lot less organic, which means the rapport that you're building with that founder gets lost.
Alex:And and AI really does help not disintermediate that process.
Victoria:It's a phenomenal observation in terms of AI's role in different systems and the tech stack of different organizations both from a GP and LP perspective. You said, okay, you can vibe code something over a weekend.
Alex:Well, someone could. Maybe not me.
Victoria:I was gonna ask you, right, like from a practical perspective and as a technology founder, I could, for example, go and buy a system or a tool to help me with SEO or I could build it myself, but it would take some time. So how do you implement that actually on a day to day perspective? Do you have someone on your team that is actually building out these core capabilities?
Alex:That's a great question. It has to come from necessity. At the end of the day, people are only going to do something in their workflow if it leads to better deal flow, if it leads to better outcomes. And so I would ask less of how and more of why. And so you say, are you losing deals because you are spending this amount of time?
Alex:Are you losing deals because you are forgetting a founder's contact information? You're forgetting small details about that founder. I had dinner with Sebastian Nalby last night, and he was talking about some of the best founders that he interviewed in the research for his book, The Power Law, were people that could recall details about their customers, about their venture capitalists, about various stakeholders that were incredibly minute. That is really, really hard to do, especially when you think about what are the backgrounds of the people that you're trying to build rapport with, whether you're a venture capitalist or a founder. And so I would encourage a lot of VCs to say, look at the amount of time that you're spending on implementing things manually through traditional SaaS products.
Alex:The amount of time would surprise you. And let's just, like, break it down to its component elements. Let's say it takes five minutes to write in proper notes, proper everything after each individual meeting. A good VC will probably take five to 10 meetings a day, so let's call it thirty minutes over the course of a single day. Over the course of the week, that's a hundred and fifty minutes.
Alex:You've lost two and a half hours. Over the course of the month, that's ten hours. Right. That's a lot. Ten hours over the course of one month in a period where you really actually don't have an equal distribution of time to source means that over the course of the year, you've lost a 120.
Alex:It's five days of potentially what you could have been doing much more organically, finding better deals, getting in the way of things. And you can only excuse yourself on that so much before eventually you start to look and say, man, these great opportunities pass me by, and I really have no other reason. Now separate from the opportunities, think about all the other demands in your time. The reason we do this is because we like our jobs, we have a, you know, a family life, a robust personal life that we also want to maintain. That's five days that you've lost of those kinds of opportunities.
Alex:Do whatever you want with that time. But as soon as you start to think about it in that context, like, it gets really, really bad really, really quickly. And I think it only gets worse as you go up the capital stack because the demands on the amount of things that you need to monitor for each individual deal actually go up. So that five to ten minutes becomes thirty minutes, becomes potentially an hour, and if you are then in an, you know, IC meeting and having to discuss your rationale for something and you aren't prepared, you don't have those notes taken, then you'll even begin to lose the opportunities that you thought you had previous to AI.
Victoria:Makes perfect sense. And the ROI you've defined in such a clear way in terms of freeing up time using AI from an efficiency and productivity perspective to allow you to focus on the value add for for your organization and for the firm.
Alex:More than even just time. It's opportunity. And once you start losing opportunities, you start to die because people shouldn't allocate to you just because you really want them to. They should allocate to you because you're the best, and if you're the best, you're the fastest and most efficient.
Victoria:Love it. And so, speaking of allocation, many LPs are obviously putting pressure and and are interested in understanding how GPs are leveraging AI internally. Have your LPs come to you and asked for your advice on what to do with AI?
Alex:You know, surprisingly, it's bifurcated. And I would say the larger and more sophisticated, historically speaking, the LP, frequently I'm seeing, the more afraid they are of AI tools. On the family office side, I think a lot of people have these kinds of questions, and family offices are incredibly robust these days, and they manage a ton of money. So you'll get a lot of questions about proprietary sourcing models, which usually speaks to some kind of AI algorithm that helps you with sourcing. From the historical kind of multi asset manager, really, really, really big LPs, the questions are actually mostly around cybersecurity and how AI created vulnerability towards other aspects of how you're managing the fund.
Alex:That was interesting to me because I think it's kind of the wrong question. You should really only be asking that if you are simultaneously asking how you are implementing AI over the course of your workflow. Then the cybersecurity aspect is really, really fair. But it's not being asked nearly as much, especially when you consider that a lot of managers got started because they claimed they had a proprietary sourcing methodology that was largely algorithmically driven. Those people really have no defensive mode in a post AI world.
Alex:Now their capability is everybody's capability.
Victoria:And so, from an LP perspective, you spoke about the bifurcation. So the LPs that are interested in AI, have they come to you to ask your advice on what they should do internally as well on how to adopt it?
Alex:Maybe for better or worse, my advice tends to be pretty publicly published. So maybe they already know how I'm feeling about a lot of these things, but the yeah, they have. They but most of the time, I think I've been impressed at how robust an individual's understanding is of the available AI tools, So it's really just an organizational sort of bureaucratic motion that they're trying to get through. And so I think those LPs would say, can you share a success story with me so I can convince the remainder of my stakeholders within the bureaucracy of the value of implementing these tools. That is oh, I've gotten a lot of questions on that.
Victoria:It's fascinating, and we're experiencing that as well where you may have buy in and eagerness from the highest level of an LP or Allocator organization where they're saying, we want to be AI native or AI forwards. And you have younger career professionals who are also very eager who may have a little bit more fluency from a technology perspective. And oftentimes it's that middle management layer where there's fear of what if I share this data that's under a confidentiality provision with a GP? Many different kind of steps to cultural adoption Yeah. That the industry and organizations have to go through.
Alex:Right. Well, and you always get this feedback where someone would say, what if it's wrong? Well, people are wrong all the time. Right? So I think it it should not replace your judgment.
Alex:Right? Things feel and look kind of right, but, like, the augmentation. I mean, when I was getting started in my career, I was a consultant. Most of my job was taking massive amounts of data and making it look fancy so I could put it into Excel so it would be easy to query. And I would go weeks doing this kind of stuff.
Alex:And, you know, the other day, completely open source information that I was using but I was trying to categorize in a certain way, put it into one of these AI models, and it was done in thirty seconds. And I was like, oh my god. I've, like, lost so much time. Like, and and you I guess you could make the argument that, like, maybe you atrophy on some of the other parts that made you really good at this kind of data analysis. I as someone who is very close to that part of my life, I call BS.
Alex:Like, it's not atrophying because you never wanted to do it in the first place. You never needed to do it in the first place. I just couldn't afford to outsource it. So I again, you are leaving money on the table by not implementing AI models as they are right now. All sorts of questions have yet to be answered, whether it be cybersecurity and accuracy.
Alex:But if you are waiting for those questions, then you are waiting for the industry to pass you by before you have the the answer. You have to keep up with it.
Victoria:Like with any technology, there's gonna be an adoption curve, but I fully agree with you that folks that are thinking about it today, starting to implement those specific use cases on an efficiency productivity perspective, and then moving towards the value add as you mentioned, right, is the the core way to get ahead. Yeah. And implement it.
Alex:And dip your toe in. It doesn't have to be all in. You decide to like I when people say that how are you leveraging AI? I mean, it it could be anything. Google something on AI mode, right?
Alex:If you don't trust Anthropic or OpenAI, then use Gemini. Right? And and and see how much more efficient something is when the answer is summarized. I mean, even on my iPhone, I'll get a text message or a and and the the whole chain will be summarized, which is great because some people will text you, like, 20 times in a row, and I'm like, I can't keep up with this right now. And so, like, that natural language processing is at some level AI, and so bring it into smaller aspects to start to get comfortable with the machines that will eventually be your true copilot in life because it's gonna happen.
Alex:Now, this doesn't apply for everybody because there's always gonna be an exception that proves the rule, but that's how I would emphasize it. Those are the exceptions that prove the rule. Mean, Warren Buffett, like, doesn't read emails. Right? Like, that seems to work out for that guy.
Alex:For most others, I do think you need this information workflow kind of tied into yourself. Right? And for every Warren Buffett, there's some even further anachronistic entity that would be like, yeah, but he does use a pen. You know? And Quill is where it's at, and I really think that's where humanity topped out.
Alex:So my point is there will always be somebody who has something very idiosyncratic in their process, but you're not that somebody, more than likely. And so figure out how to bring it into your life because it will make your life better.
Victoria:And how it can amplify your organization's edge. Absolutely.
Alex:100%.
Victoria:You mentioned something very interesting, right, about applying AI to help you improve your sourcing. And so there's many different aspects, right, that go into sourcing. Obviously, finding your specific universe of investable companies, starting to actually screen them, go through the process. Yeah. And curious, when you think about AI in the future, when you have LPs, GPs, and also underlying companies that are, let's say, AI native.
Alex:Yeah.
Victoria:How do you think that ecosystem will evolve?
Alex:It's a great question. First of all, let's walk through all the parts of managing a firm that don't make money. There's a lot of them. There's a lot of them. There's a lot of reporting, whether it be two companies, two LPs, that takes time.
Alex:There's a lot of every week at j two, we have our Monday morning meeting. Everybody gets together. We're all sharing our notes, kind of showing our homework, if you will, because our philosophy is that everybody should have carry in the fund, so we divide carry among everybody, and as a very healthy byproduct of that, everybody sort of proves why they deserve to have their carry coming, hey. This is what we did during the week. These are the meetings I had.
Alex:These are the and and that can be a really long and arduous process. And I even think about the fact that for that Monday morning meeting, which actually begins Sunday, when we actually report everything and everybody needs to read everyone else's reporting, that process takes a really long time to do nothing more than summarize what I did this week and the deals that I'm most excited about this week. So exactly back to that point is how you source is also a function of the amount of time that you have to source. And so if you are not optimizing your time when you're not doing something explicitly sourcing, that is implicitly getting in the way of your sourcing cause you're spending less time on it. So that entire workflow is something that could be replaced.
Alex:From a pure sourcing perspective, you would be hard pressed to find a fund, an inventor, that doesn't say something in the order of this type of sentence: I am looking for so and so who has been at this firm for this period of time that is probably about to leave and probably about to start a company. I'm like, oh my god. You're the first person to be really excited about the CTO from high flying startup that's been there for four years and is at the end of their vesting schedule and wants to start a new company? That's amazing. That must must be the first person to ever say that.
Alex:Now, the question of speed though and how you get there, because your network is finite. You can only know so many people. And then moreover, when you get there, what are you even gonna say? And a lot of VCs don't think about that kind of thing. More often than not, what they'll say is, I'm from this fund.
Alex:I'm sure you've heard of me. And all the founder has to say is, no, I haven't. And it's like, oh my god, well that's 80% of the lines I had prepared. Right? So, even better would be, how can I, first, find that person using AI, which is functionally the equivalent of whatever, like, LinkedIn navigator hiring function that you that LinkedIn would have, but now is available more ubiquitously?
Alex:And two, when you get there, what are the things that somebody says to this person that usually resonate? And so, for example, if someone were to try to get your attention, Victoria, they would know your podcast. They would know some of the hotter takes on the episodes. They would know that you went to Stanford. They would know that you went to Harvard.
Alex:They would know that you played polo. Those are all the kind of things that it's like, you come into the room, and if you know that kind of stuff, that's an advantage. I've been impressed at how frequently venture capitalists have put very little work into creating that sympathy with the counterparty that they're trying to get ahold of. We're not selling stock. We're connecting with human beings that are building their life's work.
Alex:The more you can create that connection, the stronger it is. And sorry, a lot of VCs can be kind of obtuse. This will help you get around that. This will help you leverage the parts of your personality that maybe you're sensitive are not effective, and will help you become better at connecting with another human being, ironically.
Victoria:Yeah. Which is the part that AI cannot do. And we shouldn't let do, right, in the sense of that the focus on the human relationship and expanding your network is extremely important. Yeah. Would love to also look at the other side of the ecosystem, Alex.
Victoria:So from an LPGP perspective, similar dynamics exist. Yeah. And so I have a vision of a future that may be in five years, it may be in ten years, where you have LPs that are AI native and that are able to actually start to truly identify a broader universe of managers. Right? Even the most sophisticated LP, LP organization, multi billion dollars investment team will review max 200, 300, 400 funds and and firms a year.
Victoria:And so if they're able to actually see a universe of the 56,000 alternative investment managers of which VC is a subset Yeah. They're able to make a better judgment call to actually see what is out there. Curious, like, from your perspective, do you think that'll change how LPs and GPs interact?
Alex:It's a great question. Never been an LP.
Victoria:But you're on the receiving end?
Alex:On the receiving end. Yeah. It's you know, where it will make a difference, emerging manager programs, I think, will always be some element of grunt work because you're gonna have to find something that people don't see. There's the easy button that a lot of people press, which was you are a GP from high flying fund, and now you're starting a new fund, so emerging manager button hit, bam. Oh god, they're really an emerging manager.
Alex:They're just emerging on their own, but they've been in this game for a long time. Everybody is has the exact same processing power on how to evaluate those types of opportunities. I do think that there's an algorithmic aspect that can help you with the people that are coming from nontraditional backgrounds in starting these things. By the way, I think this has kind of fallen out of vogue recently, but a lot of the a lot of the perpetual issues that we have with the industry, whether they be around a bias against women, minorities, etcetera, are things that AI would actually perpetuate. Right?
Alex:Because those are all going to be backward looking heuristics. So if you look and say, what do the best managers look like? Well, most managers were white guys. So you shouldn't rely on whatever the information was historically to tell you everything you need to know about the industry, because you would miss out on some of the best venture capitalists of our time if you were operating under that methodology. There's also like a lot of elitism on the schools and all this kind of stuff, right?
Alex:So that's imperfect. However, where AI will be very helpful, I think, moving forward is how are you evaluating your manager on an ongoing basis? Say something like, that I think is more dangerous than it should be. Your manager is there to make you money, right? You want to invest in a manager cause they are the best at their asset class, and they are going to produce the returns commensurate with the risk in their asset class.
Alex:Very frequently, I'll talk to LPs and say, well, they have a really great annual meeting, or they get all these really cool guests, or this guy knows so and so, and you're like, sure. But that hasn't actually come into the returns, and in this case, ironically, AI does the opposite for you. AI can evaluate the way that the manager is performing against 56,000 other alternatives and tell you, I know that their uncle is so and so, but this guy actually sucks at his job. Or, I know that their uncle isn't so and so, but this gal is actually really good at her job. And then mechanisms that you may not know.
Alex:For example, it would be really easy to confuse a lot of people and saying, I'm in all of these high flying deals. And someone would say, yeah, but you're in the pre IPO rounds for all of those, and that company has had a fall from grace. I don't even know what the performance is. I I I got a a slick sheet from someone who was trying to raise money previously. They had they had listed a deal at a mark to market that I knew was inaccurate.
Alex:But if you were someone just looking at the deal, you would say, man, that person is crushing it. We should allocate. And you may not even know that there is, like, outright fraud in the way someone is representing the way that this company is perceived, because all you have to go off of is this very esoteric information set that this person produced for you.
Victoria:It's fascinating. Another application related is, for example, the retrospective analysis. So you may be an LP. Yeah. You want a new long short hedge fund manager in Asia.
Alex:Yeah.
Victoria:You do a market mapping. You go to Asia. You meet with 10 firms. You maybe invest with one of them. And what happens with the remaining nine?
Victoria:Absolutely nothing. Right? And AI and the tools actually allow you to track if you're getting fact sheets from that firm, how the ones that you passed on did and actually compare it. Very powerful tools and analysis. Painful in some cases, right, to review Yeah.
Victoria:For some folks, but extremely important to help improve the outcomes of these organizations.
Alex:Well, and acknowledge that it is a very human emotion to not want to be wrong. So more often than not, you would if you were to get that those nine pieces of information, one, you would filter against it. You would be like, I don't need to hear from this manager. I already passed on them. So you wouldn't even know.
Alex:Two, let's say you did receive somebody got through your filter, and you'd say, yeah, they got lucky. They were just in the right place at the right time, but I'm still right. Right? We wanna go home every night, and we we go to sleep, and we sleep well when we tell ourselves that we were right today. We don't sleep well when we tell ourselves we were wrong.
Alex:And you actually need AI from a broader perspective to come and look at all of this data and be like, there is an inherent bias in the way that you are looking at the world. Ironically, I mean, who would have thought that you would need a machine to tell you, like, to get over your bias when our criticism of machines is that they're historically biased? And so it's that contextual element. It's your feedback plus what you're getting from the algorithm that will help you make better decisions, if your intention is to grow. If it's not, then, hey, pass it off, whatever.
Alex:You can, you know, you're doing your thing and I don't know what your comp it looks like. But if your goal is to make more money, then your goal has to be to grow.
Victoria:Yeah. Fascinating. Alex, to wrap things up, would love to ask you a kind of forward looking question. When you look forwards in five, ten years, what do you think the ecosystem between LPs, GPs, and underlying companies looks like with AI adoption?
Alex:From the LP perspective, I do think that you're going to have this broader management really contextualized. That's I guess the word I'm looking for without a better term available. You're going to have a lot more of this, hey, send me the information on this codex, and I am going to put it into the feeder, and that will automatically rank order who I like, who I don't like in a much more blind mechanism that will help LPs make better decisions. I think that's coming. And that'll be everything from mark to market IRRs, like cadence of reporting, DDQs, especially in emerging manager programs, they have, like, these entire DDQ sets that you can then say, hey, are these people actually growing and, like, adopting other workflows outside of is their office cool?
Alex:Like, are they building in better back end functionality that help them make better decisions along the way? I think that's what's coming for LPs. From a a GP to portfolio company perspective, once the relationship is established, I think that part process will also be automated. We ask for the same things every single quarter. I don't understand why this can't be done algorithmically.
Alex:Like, that will come. In five years, it will certainly come, and I cannot wait. Hopefully sooner than five years if anyone's working on this right now. And then I think from a a sourcing perspective, let me begin by saying, you're never going to replace the human in the loop. People want to work, even even outside of the branding of the fund, people want to work with the person.
Alex:They want to meet the guy or gal that's in charge, that is the person that they're giving up a meaningful part of their lives to, which is I think the accurate way of looking at it. But I think that sourcing piece will become, again, much more robust and much stronger in the way that you can find more opportunities and keep track of more opportunities the more that you create that contextual mechanism. And I think if you don't, you're gonna get left behind.
Victoria:Alex, thank you so much for joining today for the conversation.
Alex:Thanks, Victoria.
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 your own process is stuck, go ahead and follow or subscribe wherever you get your podcasts. And if you're curious what Agentic The 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 work flows. Until next time.