Brian Bell (00:02.223) Hey everyone, welcome back to the Ignite podcast. Today we are delighted to have Rajesh Gopi on the mic. He is the founder and CEO of Zive, an AI native fund operating system that helps GPs, fund admins, and CFOs automate back office work such as compliance, capital calls, and reporting. Before Zive, he co-founded two twelve, an Austin based equity management modeling platform. After earlier stints at IBM, Freescale, Coinbase, Hypori, and others. thanks for coming on, Rajesh.
Rajesh Gopi (00:29.334) Yeah, thanks Brian. thanks for having me.
Brian Bell (00:32.005) Well, I'd love to start with your origin story. What's your background?
Rajesh Gopi (00:35.38) my background is pretty much writing software code. So I've been a software engineer for the last 30 years. got excited about internet back in the mid-90s when I s first saw what was an early version of a laptop connecting to Sony.com. it took about three minutes for the page to appear, but that really told me something magic was happening here. And and then I worked my way into building applications in the internet era. built a lot of protocol stack, worked on device drivers, worked on wireless technology, blockchain, that led to equity management. And there I met with a f a lawyer who was basically helping VC make their shadow books match with what the fund admin was putting out. And that's how I got into the nuances or complexities of fund administration and that led to starting Zyve. So it's been a it's been a journey, fantastic journey.
Brian Bell (01:33.957) What's been the most surprising thing about building Zive? Because you're sort of you know, you you it's not like you spent your whole career in in this space, right? So you're you're an engineer who got exposed to it and decided, hey, I can build something better. you know, what's been your biggest surprise as you've, you know, been working on this project for a number of years now?
Rajesh Gopi (01:45.281) Okay.
Rajesh Gopi (01:57.049) So I think I see everything as like a math problem. Like everything has got a lot of variables, and it's really trying to understand how to best optimize those variables for the optimal solution. So that's how I I see problems in general. with fund admin in particular, the traditional thinking was okay, you need a lot of fund accountants or controllers. they use something like a spreadsheet to run the operations. and and they get like dashboarding to see how everything would perform, whether what the state of the fund is, but they always had a fund control in the middle of everything, right? So now the the disruption here was can we have AI do a lot of those tasks and be in the middle of the entire workflow operations and then and then take the controllers away from doing day-to-day manual tasks. And get them to be part of the final judgment call, making sure they are seeing all the numbers accurately and and whether they can stand by the numbers that are being produced by the system. So it is really having these fund admins think differently out of the box that okay, they don't have to be in the middle of the entire workflow, but they can be at the at the beginning, end and oversee what happens through the entire workflow has happened. So that switch is where we kind of started working on and I think we are beginning to deliver we have more than six billion dollars of assets on the platform so we kind of prove that okay we are able to show how that transition can that is correct yes six point two yeah
Brian Bell (03:39.076) Six billion with a B A U on the platform. Pretty, pretty incredible. And if I recall, there's like two tracks. There's the self administered track, meaning like, hey, I can just use your platform and kind of run it myself. and there's the full service. walk us through kind of why you have this, you know, two levels of service model.
Rajesh Gopi (03:59.779) Yeah. So so we we started with emerging managers. So for us, emerging managers are like sub seventy million AUM. and we picked them because of some very specific pain points that have, especially in terms of the cost of doing fund administration. So a twenty million dollar fund and a two billion dollar fund is no different. It's the exact same requirements, it's just that the the AUM is yeah.
Brian Bell (04:22.201) another zero is what I like to say because I'm trying to raise a much larger fund than my previous ones. which is public so I can talk about it. But yeah, I'm just like it's just another zero.
Rajesh Gopi (04:27.178) Exactly. Right. It is, yeah, it it doesn't make the complexity any different. It's still the same complexity, it's just the magnitude is tightly different, but they still have to do everything that eight billion or ten billion fund would do. So so for us, we got a market with early stage pre-seeds, seed type emerging managers. two reasons for that. one is their needs were Pretty constrained, right? So they they wanted to onboard LPs, call capital, record expenses, take management fees, produce the quarterly statements, annual statements, and get K1s filed. So I think th those are the big seven or eight things that they they needed to happen. So as we were building the tech, we wanted to kind of solve for those and it was very easy to go hit those up, right? So that's really how we got the market. and then we had our controllers overseeing the the data that is coming out. We thought it was accurate, but then there's always this the the the thinking behind your head saying, Hey, how do I know if it's really accurate? So that really required us to go take it to other segments where we had firms with controllers in-house that basically double clicks into all the numbers coming out of the system. So that made us switch to taking on enterprise clients towards the middle of last year. so there the customers were CFOs, and what they really wanted was a software that can take it in-house. But fund administration was a small piece of the pie, right? So they look in terms of what are my deal flow, how much cash I need, what is the ca do I mean cash reserves? They have banking needs, portfolio management, fund admin then becomes a piece of that. Then they have manco GP. all kinds of entities. They have waterfall distributions, taxes, K1s. So it's an entire shebang of needs that they have. but the most importantly they are pretty anal about where did the penny disappear. Like they want proper footing. So working with those type of customers made us become more accurate, get all the traceability. We are able to drill down the numbers, we are do audit trail
Rajesh Gopi (06:42.98) replay history. So those are all the key pieces that we started working on, which automatically leveraged the emerging managers. So they got the benefit of really what an enterprise firms would do. Right. That's that's really how we kind of mapped it. As we got to these two segments then we started getting inbounds from existing fund admins because they already have relationship with different GPs solving very important problems so they wanted to leverage our solution so we started licensing our software to some of fund admins So it became a very interesting mix of different types of personas using the platform. And from our standpoint, we are a technology player. We deliver solutions and we getting proof points from different types of segments that tell us that our our our system is good to go for scale, which we're targeting sometime early next year.
Brian Bell (07:35.405) Yeah. You know what it kind of reminds me of is some almost like what Mercury did in business banking. Right. they kind of came along and they just built a better stack that was easier to use and, you know, very user friendly. would you kind of describe Zive as that for fundadin?
Rajesh Gopi (07:40.366) Uh-huh. Yep.
Rajesh Gopi (07:50.439) we are working towards that, right? because at the end of the day, it's not just the financial, the data is just one piece of it. Then you have all the contracts, your LPAs, your side letters. you may have again sub docs with different nuances. So the question really is okay, can AI process all of these data and contracts and start producing the right output for the GPs? So the ease of use is really with you as a GP interacting with the AI and trying to get all the information. So it's a very common type of questions that is being asked is how much management fee can I capture in Q4 of this year? And what is my budget look like? How much cash is available for me to invest this year? and and the LPA would say that okay, if your if your LP comes less than 500K, you're gonna call 100% upfront, but not everything is available for you to invest, right? So you need to kind of meet the the prorado shares. So those are the kind of things that we we are kind of fine-tune our agents to basically go solve. So it's a very complex problem to to basically address and we are trying to simplify it to your mercury analogy like okay how do we make it simple so it is easily understood by both GPs and LPs. So so we are trying to get there. It's not there 100%, but we are we are heading in the right direction.
Brian Bell (09:10.533) So you describe yourself as an agentic fund operating system, 200 fund to where AI agents, maybe there's more now, seven hundred and fifty deterministic compliance checks running on a native ledger. walk me through what that actually means for a GP who kind of logs in and and and starts going through their workflows.
Rajesh Gopi (09:29.846) Yeah. So the at the end of the day, it's it's not about whether AI is accurate, right? So AI that's not the question that you're asking. Is AI accurate? let's say in a traditional font admin, it's a font controller doing the work using some spreadsheet database system. but something goes wrong, you're able to go ask that individual, okay, tell me what happened here. and okay, did you fix it? So that that easy part happens, right? But when you talk about AI. If AI gets something wrong, then the question comes up saying, hey, what else did it get wrong? Like that that becomes a bigger challenge to to address GP concerns. So so for us it's all about can we audit the the data that the the system is producing? it is deterministic, meaning can we ensure that the same number happens every time that that you replay the the entire financials? are you able to justify what happened with with with with the numbers that came out? so those are some of the qualitative responses that we want to kind of solve for with with respect to what a GP would look at look at. so we knew the problem going into starting Zybe, we can't just depend on AI to do all the work. the problem when we started Zive two years back was a context window. The context window is an idea within AI that says, okay, if the con is the window within which it the AI remembers the I the the the notion that it's gonna work on as a next step. But when the window is small, it doesn't know what it just recently worked on. It starts hallucinating, it starts making stuff up. So we had to address that at time zero. So we took two very important steps to make that happen. one is we created like a deterministic rule engine, right? So that is basically looking at all the ASC standards, like ASC 820, 718, 343, whatever. So we went through creating a if, if, then else checks, right? So a bunch of rule engine. And then we created our own rule engine, but now it is more than a thousand rules that we kind of built in. So that is a very first step.
Rajesh Gopi (11:48.131) The second very important thing was we don't use spreadsheets in Zybe. So we built our own ledger operating system. So meaning we compress the entire fund administration to the unit unit economics, which is the journal entry. So when anything hits the journal entry, then we immediately run it through our thousand-plus checks. We look to see whether any mistake was made when the journal got recorded. And we have different personas now, right? So we have a GP persona. LP persona, a fund controller persona, we have a CFO persona, an analyst persona. So these are all different skill agents that are kind of monitoring the system 24-7. So anything that hits the journal, runs through the check, different persona looks at it and flags, hey, I don't like this number, this doesn't seem right. So that we get to that point. now that we know the journal is correct, that basically lets us automate the entire financials. P caps, capital account statement, everything gonna flow out of that. That's really how we kind of solving for the GP problem. So at the end of the day, it's it's not just about the accuracy, but can we repeatedly ensure that it is accurate? Can we replay the history? So those are the big pain points that we go and hit up.
Brian Bell (12:59.641) Yeah. So you've been administering over six billion of AUM, effectively zero marketing. help us understand how you did that, right? Was it just word of mouth? you know, just how'd you grow that quickly with with effectively no marketing?
Rajesh Gopi (13:14.039) Yeah.
Rajesh Gopi (13:17.974) Yeah, it it is all mainly word of month. So it's basically talking to GPs. I get into like my LinkedIn post, like kind of talked about stuff on LinkedIn. we spent a lot of time with emerging managers and their ecosystem, talking to getting referrals. at the end of the day, it is trust is very important. so it really depends on who's introducing me to another GP. So that's really how we kind of worked our way through. and and we didn't want to spend a lot of because we didn't even know what we were building. Like we weren't sure whether we are a fun admin platform, we are a fun admin services, we we didn't know what we were. So so it didn't make any sense to go prematurely market ourselves as something that we probably wouldn't wouldn't be. but over the last three years, like we are now very sure about what we are. We are really a fun admin platform. So this is a system with some services component available to emerging managers, but otherwise, when we talk to enterprises, they say, hey, give us a software, we already have all the people, we know exactly what we want to do, give us a software so we deliver software to them. Same with fund admins. So they want software and we give it to them. but just for the emerging side, we use them to provide some services so we don't have a lot of overhead with respect to what is needed, and the needs are much simpler. and once we start proving ourselves out, then we build a marketplace of l law firms, marketplace of CPA firms, audit firms, and they start referencing it. They know our pricing is very transparent. It's not based on AUM, it's not based on number of LPs, number of investments, which is typically what happens in other firms where they have people, headcount, and and and spreadsheets to deal with. we are pretty transparent in terms of what we need. So that also leads to a lot of word of mouth and we start picking customers that way. So again, we're still not thirty percent.
Brian Bell (15:12.517) Well, yeah, because if you're if you're managing tens of billions of dollars, right, and you're getting quotes from some of the other leading fund admins, they're they're hitting you up at a huge price point, right? And you're like, look, we have all the the s we have all the accountants and and services people in house that can do this. We're we're managing fifty billion dollars, right? As our, you know, pension fund or whatever it is. And we just need a a software to help us.
Rajesh Gopi (15:38.646) Exactly. Yeah. I think I think their economics is pretty interesting. So for them it is roughly one fund controller for every 10 funds, and they can stretch it to maybe 20 funds. but now they they're now thinking, okay, how do I have one fund controller manage 100 funds? so and really the equation is okay, how do I have one fund controller manage instead of 200 million AUM, can it can they do two billion AUM? So The unit economics is changing for them. and there's also like a big gap in terms of very skilled fund controllers. it's not very easy to find them and they want to hold on to the right fund controllers, give them leverage so they can scale rapidly, do more interesting things. usually what they say is that the fund administration ends when the book is closed. but for us, we think that's when it begins, right? So once the books are closed. Now the fund admin or the fund controller is basically now looking to see, hey, how do I handle liquidity? How do I work with the auditor and tax? how do I answer these LP questions about these scenarios? how do I help the GP with budgeting, compliance need? Those are all critical problems that an AI system cannot solve. And we want to kind of power those controllers to get to those levels of complexity rather than doing mundane tasks like calling capital, recording distributions, journal entries, those are things that AI can take care of because that that's the easy part. So that's really how we we look at this space. and and we've been pretty successful both in the venture and private equity space and that's really where we are kind of focused on.
Brian Bell (17:19.299) Yeah, you what it strikes me if you if you think about the history of I don't know fin financial technology, fintech. you know, the average bank in nineteen eighty had a certain AUM and a certain number of people that it took to, you know, run run a bank, right? And that I I I forget those numbers, but it like something like fifty billion and a thousand people. And now the average bank ac actual AUM's gone up. I don't know about how much, you know, two X, three X, whatever. But the people have gone down by 90 something percent. And so that same bank takes takes like nine people to run that same AUM. and it it kind of strikes me that Zive is sort of like this harbinger of of the graduation of venture capital and and private equities maybe already there to to to a more sophisticated financial product rather than this little cottage industry with like little fund admins.
Rajesh Gopi (17:55.107) Yeah. Yep.
Brian Bell (18:15.622) it's it's becoming this much more professionalized area of capital.
Rajesh Gopi (18:22.815) 100%, man. So so so for us, the it's the way we look at this is we want the controllers to be in the middle of of oversight, right? Oversight is the most important piece. So is is is everything being done properly? I think that's really where we want the controller to step back and not go in the weeds. most of the controllers today they're going the weeds of solving a problem. And they they're not able to step back and look at okay, what else is going on here? Look for pattern matching went one fund to another fund. So those are all the areas that they like to spend time on and then they don't have the ki the bandwidth. I think that's really where we are getting all the interesting traction. So so today our CTO was telling one of our customers, in three months, your controllers are probably gonna do five percent of the task that they're doing today. once they get Zy fully onboarded. because it's for them, it's not just about recording entries and making sure expenses are being tracked properly and valuations are done. it's more to do with reports. So they have to create reports maybe every Friday. and they produce these reports by looking at different systems. So they need to look at all the investments that happened that week, any markup that happened that week, all the expenses that happened that week. So they're kind of pulling all the information, spending time, putting together reports. We completely eliminated that that entire four hours that spent on Friday morning putting that together. So it's more of okay, now it's already there, look at it. Like what what are the other interesting things you guys should be doing? what are the problems that they need to think about? if a fund says, Okay, I want to go set up another billion dollar fund, the the finance team no longer have to think in terms of okay, let me, I need to go hire. two controllers and and and this software that needs to be turned on. So we're kinda eliminating a lot of those pieces for them. So it's just becoming more operationally efficient. So that's that's really we are beginning to see it. Again, like I said, we are still early stages, but the the feedback has been pretty positive so far.
Brian Bell (20:34.617) Yeah.
Brian Bell (20:40.153) Yeah, to to provide some numbers, I asked AI the the the US bank thing. So nineteen eighty, one and a half trillion. Today twenty five trillion. So a huge increase in assets for the average bank. the employees have roughly stayed the same at one point three seven million to one point three five. So we've increased the assets. And the assets per employee have gone from a million per employee to almost nineteen million per employee. So that's that's just an example of technology and software creating leverage, right?
Rajesh Gopi (20:49.945) Yep. Yeah.
Brian Bell (21:09.231) And I think Zive is kind of coming along right at the right time in venture capital, where it's growing into this very sophisticated asset class. you know, 200 billion, I think, deployed last year. So a lot of money flowing into venture. There's more startups being created than ever, right? So the, you know, it's becoming more more complex to to manage. And so I think you're gonna see this, especially with tools like Zive in the next, you know, five, five, ten years, that you know, the the AUM per controller was that that metric. that you had. You know, instead of two hundred million, it might be two billion.
Rajesh Gopi (21:41.111) Exactly. Yeah. I think I think that's that is where we already see that happen, right? with with the firms that we're working with. and they don't have to start throwing new controllers into the mix. the the existing ones are able to easily scale with number of funds that is being brought on.
Brian Bell (21:58.426) You know what found really interesting? because I built a lot of AI, you know, and I've had to create rules engines and stuff. walk us through how that the the the rules engine works, right? Because if you if you get stuff wrong, you cannot just rely on the context window to do all the calculations inside the model. You have to basically say, hey, if if this then that. And it like here's the math. And so how does that work? Is it just a functional tool call that the AI goes, hey, I've gotten to this point on the ledger. walk through simulation, whatever it is, and I need to actually call this tool with this number and get this back. And then I should run another check and say, is that the number I expected to get back? If not, kind of bubble it up.
Rajesh Gopi (22:37.399) Yeah. So a very simple example is let's say you started your fund on April fifteenth of that year. and you may say that okay, your your management fee is two point five percent that year for four years, and then you cut it down to one point seven five or whatever. So it's very common for fund managers, especially emerging fund managers to set that up. And that's a very common place where we see mistakes being made by traditional fund admins. So in our case, we need to calculate the 2.5% until April 15th. And then from April 15th, 16th onwards, it's 1.75, whatever that number is, right? So when we accrue that particular expense for for Q2, it's immediately it immediately checks it produces that number of whatever's the management fee for that quarter. and then we run it through our Our blue check rule engine to see, okay, does that number match with what the AI was supposed to do? So it's really like another agent that is overseeing what the work was done by the previous one, right? And then if if it matches, then it's good. If it's not, then we immediately go figure out, okay, where the mistake is. So for us, nothing is perfect, right? So there's always corner cases when things can go wrong. So we never release anything that an AI produces without our controllers overseeing that. That that calculations. So it's only after like maybe like thousands of runs of doing the exact same thing that we feel confident that okay, we've hit so many corner cases now that we can have a judgment call made before we release. But otherwise, we were actually spending time physically calculating those those that math. and that's how we build the level of confidence. And now that we build the confidence level, then we then we let it ride. So it's the same thing, like when you record when you record an investment, for example, right? So the interesting thing when we start looking at enterprise customers, they may have like 20 active funds, okay, that they're deploying off of. And let's say they have an allocation of $15 million into portfolio company. So they have 15 million coming from seven funds. and it's like a one million here, three million there. and and they have a single stock purchase agreement document, so they just drop into our system.
Rajesh Gopi (24:57.107) And and we our AI automatically parses that document, understands which fund is invested what, it gets into the individual fund instances, record those expenses, and now we are running checks against is everything tallying out. So it is that kind of sophistication that we built in into our system. and then we have a controller that oversees the work, makes sure it's good, and we see that repeatedly and and everything is good, then we kind of let it go, right? but it's a process. that's why we feel that we are not ready to scale until we have enough runs under under under our belt. And we think we are getting there probably by end of the year.
Brian Bell (25:34.373) So leg legacy fund admin has been dominated by the same handful of players for a long time. what's a belief in the industry that's popular that you think is wrong?
Rajesh Gopi (25:44.904) I think the I think I think the the typical one is okay, you need the fund controllers and they need to be in the middle of the entire action. So that's if you if you look at this, it used to be things used to be done on paper that went into some sort of a dashboard. then there was spreadsheet that went into database. so th th those are minor improvements that were being made, but at the end of the day. Was a controller that was in the middle of the entire action. So they had to oversee all the work. The way we see this fundamentally change is that AI will do the work, right? we don't need controllers doing the work, we just need the controllers sitting there to oversee the for the oversight, for the for the judgment part of it, and then the accountability. So at the end of the day, like if if the number produced is something that the the controller can vouch for. So that's the that is the disruption that we say that we're putting in place rather than have the controller be a requirement to be part of the entire workflow so it's a it's a change in mindset so when I speak with large fund admins that is where I'm spending a lot of time okay you don't need the controller to do this action we'll take care of it a good example is One of the largest fund advents today, they manage more than four trillion dollars of assets. they have about 300, 300 people just focused on onboarding new funds. they have another 200 plus that only do bank reconciliations. there's another 200 plus that only do expense tracking. So that's really how they kind of spread their workflow. And we are what we have to pitch them in, and we've done pilots with them saying, Hey, we don't need all that. Like This these agents can do all this work. and once everything is done, we'll produce a report where we will tell you what are the challenges that issues that we have addressed and then have your controller come and oversee and fix those issues. So that that's really the disruption that we are basically forcing and making them see the get comfortable with the with the idea that hey, you know what, we really don't need 300 people doing bank reconciliation. So it's just a learning, but that switch is happening pretty soon.
Brian Bell (28:07.333) Yeah, and you think about the the just the enormous labor cost there, right? On a four trillion dollar co four trillion of AUM and it's thousand or so people just doing repetitive stuff that agents can do.
Rajesh Gopi (28:12.441) Exactly.
Rajesh Gopi (28:20.589) Yeah. And and even when I look at some of the numbers, the margin is super laser thin. Like they're operating between maybe around four to five percent margin. now the the pitch that I have for them is increase your software spend from one percent to ten percent, but increase your margin from three percent to twenty-five percent, right? So so they're already seeing the value proposition. So there's warming up to that. So the the the
Brian Bell (28:25.401) Right.
Rajesh Gopi (28:48.409) The economics of this is gonna be fundamentally very different very soon. I expect within the next two years there's gonna be a light bulb moment and it's gonna be very different, how how these admins operate.
Brian Bell (29:00.109) Yeah, I think you're definitely creating price pressure in the industry, you know. with with my current fund admin, which is with which isn't you, I I kind of floated like, hey, I can go there and like I can save money. And, you know, they've had to keep their prices lower because of that. Right. So you're cre you're creating a price pressure in the industry, which which is great for us GPs, right? And LPs. you know, because if every everything's 10x more efficient and just as good.
Rajesh Gopi (29:16.315) Okay.
Rajesh Gopi (29:21.837) Yeah.
Brian Bell (29:28.161) that means, you know, these fund admin costs don't have to scale with AUM like they did before. But, you know, speaking speaking of that, how do you scale your your pricing? Cause obviously managing a billion dollar fund, you know, it does cost more an agent cost and tokens and AI and software than a ten million dollar fund. How do you think about that pricing scaling as a platform?
Rajesh Gopi (29:52.334) So so we looked at our pricing is very transparent, right? So so for emerging managers, we say, hey, it's 12k per year during your investment period. That's your active investment period, you're deploying capital, calling capital. So we have our controllers overseeing the work. So that kind of covers for for that. But once you're done deploying capital and now you're in the harvest mode, we cut your cost to 6k per year, right? And the way we looked at this is what is a number that gets us 90% plus margin? And and that's how we arrived at it. we we could have easily charged more, but for us 90% gross margin, 6K gives us 90% gross margin, 12K gives us 90% gross margin during those busy, busy times. but then once we start getting into into enterprise customers, our ACV there starts at 300K, 400k. So we are we are basically getting a lot more cap a lot more cash there. And the reason for that is it's not just about fund accounting, right? We are solving much bigger problems for them. They have different systems that that connect into the standard metrics, Chronograph, Cobalt, for portfolio management. They have different banking needs. distribution is a real painful process for them. they need to spin up different SPVs, understand what is a gap to tax conversion. So there's a lot of problems that we are solving for them. So it's a turnkey product that we built that deals with valuations, portfolio management, metric capturing. legal documents, so all that is being handled for them. and and and that's where we're able to capture bigger ACV with those customers. So so for us it's pretty simple. And for the fund admins that we sell okay it's a it's a standard base price for the platform and then for every fund they bring in we just make it very interesting for them to add more and more funds. So So we see ourselves as a technology player. We want to get maintain our 90% plus gross margin. And and this way we're able to make stuff happen. for the emerging managers, it's it's a services component, right? So they really need human capital be their back office. so there is a there is that that is the only place where place where we have services piece. but again, that's very effective. Like let's say you want to call capital, you just let us know. Okay, I want to call a million dollars tomorrow.
Rajesh Gopi (32:11.635) somebody just types it into our system and then the agent goes and creates all the documents within 30 seconds your cap call is done and it's ready for you to approve. So again, they're not spending a lot time working with an individual GP so that's how we can have optimized for scale and margin.
Brian Bell (32:29.379) Amazing. So a lot of AI native software is just old software with the chatbot on the front. What's the tell that separates real agentic infrastructure from just that bolt on?
Rajesh Gopi (32:41.229) Yeah, I think I think the for us it's all if you turn off the the chat bot, will is any AI being used? So that's a very simple question, right? so the traditional fund admins, what they've done is they've slapped an LLM or they use Cloud connect through MCP to the system of record. And it's really not AI. There's no AI there, right? So for us, it is really foundationally Is there other agents doing all the work along the workflow? That's really how we look at it. So for us, onboarding is an agent. we don't have an onboard specialist. We don't need that person. you want to call capital, it's an agent. you want to reconcile bank transactions, an agent. expense tracking is an agent, reporting is an agent. So we have more than 200 agents very specialized to do all kinds of activities on the system. We are still a small team. We just we have 14 people right now. That's it. 14 of us, including development team. A product person, customer success, a 14 person team doing six point two billion dollars of assets. and I like to say we are a 400 person company with AI agents doing the rest of the work, right? So that's really how we think in terms of in terms of us in terms of what we do. So to answer your question, a traditional fund accountant, fund admin platform, you turn off AI and nothing happens. The company still operates because they have fund controllers doing stuff on spreadsheets. but for us. You turn off agent AI on our platform, the company dies. So that's really the formal difference for us.
Brian Bell (34:12.259) Yeah. And in case anybody is listening and thinking, hey, that that bank stat that you that you named earlier, you know, one million to nineteen million is all, you know, nominal terms. I want the real terms. It's actually four X with inflation. But still, you know, with technology, we're four X more efficient in in l in the last 40 years, in the financial, at least the banking sector. paint for us a future. What are you excited about over the next five to ten years on Zive?
Rajesh Gopi (34:41.075) I think it's how do we solve for different verticals, right? So today we've already hit up VC and private equity. Now if we already have inbounds from hedge funds and real estate funds, but we've looked at their problem statements. It's slightly more bespoke. there's more complexities there. we kind of said, okay, let's go target that in 2028 and later. we just want to do a good job with VC and the private private equity space.
Brian Bell (35:07.545) Right. Yeah. You got to build a whole new r rules engine for all the depreciation, schedules of real estate, all that. Yeah, it gets gets complex in a different way. Yeah.
Rajesh Gopi (35:10.595) Exactly. Yeah.
Rajesh Gopi (35:16.047) Yeah, it's it's a total different beast. so anyway, so that's really where we think we're just gonna evolve into and even within the PE space, that is PE buyout, that is private edit private credit, just regular traditional PE firms. so and then there is this notion of okay, secondaries, and we're already building for a complete solution, right? At the end of the day, it is not just about fund accounting. it's all about okay, Can we close the books in five days? We are already getting to a point where within five days of end of quarter, the books are completed, right? So a traditional fund accountant takes 45 days to 60 days. That is their typical range. Now, how quickly can we close these things? Can we identify issues with with with the accounting, or is there something that the GP is overlooking? So we want to start putting intelligence layer, surface them to the top. rather than them focusing on what is the minutiae behind actual every number. So we want to basically give them confidence that the number is good, it's auditable, you can trace it, but now let's make it more smarter, intelligent. And we just want to help emerging managers like yourself, for example. Okay, how do you spin up a $100 million fund or $200 million fund? Assuming that is something that you want to do. how do you get to kind of punch a boy of weight and start scaling really rapidly and take some of the best practices from from bigger firms. And and I know you and I kind of chatted recently about okay, you make all these investments early on, but then you lose visibility into that portfolio company as they raise more capital and becomes large.
Brian Bell (36:56.335) Yeah, 'cause we do a lot of small checks and so we l we typically lose information rights in the A, if not the B.
Rajesh Gopi (37:01.529) Correct, correct. So that is another big problem, right? We already recognize that now how do we kind of make that information available to you? And the way we do that is having more customers in the in the larger segment because typically eight billion firm may have like more than seven hundred active portfolio companies. And there's a good chance that your investment and your pre seed stage is now being managed by an A sixteen or a Sequoia, right? So you if if we start getting those data, then we can actually triangle it and start surfacing some information as well.
Brian Bell (37:30.261) like hey you you have this you have this marked at a lower price per share. Did you realize that they raised the series C? No I didn't because they have like a whole basically yeah there's there's two kinds of cap tables. There's like the early stage cap table and then kind of the later stage cap tables. And they're they're they look and behave completely different.
Rajesh Gopi (37:44.27) Yeah, okay.
Rajesh Gopi (37:47.754) Exactly. Yeah. Exactly.
Brian Bell (37:49.094) I've realized this. how do you handle a couple technical questions, you know, because I I manage three funds, so I've seen a lot. how do you handle custom capital calls? So you sometimes you have an LP join a little late in the fundraising process, and it's like, I, you know, and we're already 60, 70, 80% called on this fund. So we'll get you caught up with a custom capital call. Is that something you guys have built in?
Rajesh Gopi (38:05.219) Yep. Yep.
Rajesh Gopi (38:14.799) absolutely. Yeah. I think at the end of the day, like the your rule is governed by the LPA, right? and any side letters that you have. So every time we call capital, there's always rebalancing. You have to is mandatory. So so let's say you have somebody that came in after your first close, then the LPA typically says that treat them as if they came on time zero. So we have to rebalance their investment. So that is automatically calculated for you. there are some LPs that come in after your final close, right? In which case there would be some rule that say, okay, they pay a joining fee or prime plus five percent. So we automatically calculate all those. So the AI does all those things. It already knows how to look for. It automatically creates all the statements and then our controller reviews, make sure it's good, then it lets you know, hey, this is what we are finding, good to go. And it's very quick, like within within a minute, the whole thing is already done for you.
Brian Bell (39:06.905) Yeah. What about reallocation? So sometimes, you know, emerging managers will have, you know, fifty, sixty small LPs, you know, a few of them have some sort of financial difficulty. They can't make their capital call. Keep the fund whole, we have to transfer their interest to a new LP.
Rajesh Gopi (39:21.123) Yep. Yeah, that's very common. reallocation, or sometimes you have defaulting LPs where you need to kick them out. so and then your LP again may have some clauses where it says okay, if somebody else is picking up their ownership interest, or are you basically going to turn that to zero? and then again there'll be some rule that says okay, 50% of the capital contributed will be returned back to the LP. those are all different nuances that will happen within your within your LPA. So for us, the way we look at this is there is a standard operating procedure, right? with for every fund that you have that is governed by the LPA and any side letter. So those are basically ingested into our system. Our AI reads those and it completely knows what it needs to do. So in you in our in our case, let's say you have that situation, you can actually type into our chat bot to say, hey, I have this LP that I need to distribute. How should I think about it? And and our AI looks at all the documents, it'll tell you this is if you're using section six three nine, this is how you do it, and six three seven, this is how you do it. So you give you all the controls.
Brian Bell (40:24.931) Yeah. But watch out because this side letter gives this LP first right of refusal and da da da da da da da da. It's like, yeah, you can adjust all that, create rules around it and and guide the GP.
Rajesh Gopi (40:31.704) Exactly. That's correct. The thing is you want information very quickly, right? Because you you as a GP, you're just focused on finding LPs and investing in good companies. That is your primary focus. You don't want to spend time doing back office work. and that's when you come in and say, okay, you you ask your question and the AI will know exactly what you're doing, gives you the answer.
Brian Bell (40:47.098) Right.
Brian Bell (40:53.711) Well let's let's do some wrap up questions. what's something you believed with total conviction five years ago that you'd now argue against?
Rajesh Gopi (41:03.747) Huh. so I I think it's back to you need people in the middle of things, right? So you need you need it doesn't matter what is it a bank, is it a fun admin, whatever it is, you need a human to be in the mix of all the events that happen within whatever the workflow is. And but now I'm like you don't need that. You need somebody in the beginning, you need somebody overseeing what the work is. You look at a person at the end. AI can take over all the workflow that happens in between. I think that is a fundamental switch that I'm pretty convinced it's gonna be the future.
Brian Bell (41:45.509) What's the best I what's the best advice you've ever gotten but maybe you miss initially ignored?
Rajesh Gopi (41:52.98) so I think we have the the adage really is build a product as an engineer. Like I'm I I think like an engineer, so build the product, get it really good, get it in front of customers, and then iterate over that product. So that's really how initially I was thinking about about building companies. but now I'm at a point where I give them something like a cloud design, or previously there's another project called UX pilot. So I I give a login to my customer and tell, hey, what is your ideal product should look like? go build it for me. So they just prompt engineer the UX experience for them and they tell us, hey, this is what I want to see. And then once they give that to us, then we just build it. So so it that that flip has happened for me already. So now I don't build it, I just give it to my GP or my CFO and tell, hey, tell me what you want.
Brian Bell (42:44.141) That's amazing.
Rajesh Gopi (42:48.643) This is your login, go create it for me, let me go build it.
Brian Bell (42:51.631) Wow. That's amazing. Yeah. I was wondering that the other day because I was using cloud design to redo our pitch deck for for our next fund. And it just did such a good job. Cause I had a like I started, you know, when I started fundraising years ago, I paid somebody to create a PowerPoint. And then I paid somebody else to create a Canva because it was a little better. and now with my cloud subscription, I just uploaded that along with all my fund docs and you know, transcripts and FAQs and I was like And and my, you know, brand assets as well. I was like, go go recreate this deck. And it did a really good job. Like, I'm pretty happy with it. and I was thinking about this as a you know, product manager because I was in product for probably 10 years of my career. how much how much different that should be these days, right? Because if I'm a PM talking to customers and you you just said it, you just give you give the tool to the customer and let them talk to talk to the tool and it'll kind of get it the to the way they like it to look.
Rajesh Gopi (43:24.471) Yeah.
Brian Bell (43:50.071) And with all the features, basically you can talk to it and create a clickable prototype. That's pixel perfect.
Rajesh Gopi (43:55.748) Yeah. And and for us, yeah, from cloud design to cloud code and and working prototype is just one day for us. So and we hook up all the APIs and the very next day, hey, is this what you wanted? And
Brian Bell (44:05.551) Wild. Yeah, Figma's gotta be cooked. Figma's gotta be looking at this like, man, Lot is, you know, killing us, you know, basically. Because that's what Figma was. Like, I take all I take the pixels and kind of, you know, lay them out with HTML and CSS. wow, that's amazing. You you have it all laid out for the engineers now that all all they have to do is code it up. But now you don't even need the Figma tool. You just need your your your voice. You just talk talk to the AI.
Rajesh Gopi (44:14.915) Yeah. Exactly.
Rajesh Gopi (44:33.409) Exactly. Yeah. A big big big win, yeah, exactly.
Brian Bell (44:37.561) Pr pretty wild. what's a skill you think is wildly underrated?
Rajesh Gopi (44:43.241) I think it is right. So when I people keep asking me what my special sauce is, and I think it is recruiting the right team members. that includes not just people working in the company but advisors, investors. that to me is the most critical aspect. How are you attracting the right talent? because I'm still pie in the sky idea, right? So I we don't know where this company is gonna go, any startup is gonna go, but you need people to believe in you and and and basically bat with you or bat for you. I think I think that is a critical skill a lot of people miss. and which is what I've been kind of honing on for the last twenty years.
Brian Bell (45:22.585) Yeah, love that. last question, what do you want your legacy to be?
Rajesh Gopi (45:28.631) Yeah. I think I think it's more of am I am I there for the the person that has a problem? Am I solving the problem for that person? Was I there at the right time? I think I think that's really how I think about it. Even today, like I have folks that I probably interacted with about 10 years back come to me for certain things that I that that that they run into some challenge with. And I kind of respond to them on LinkedIn or wherever. so I I think it's more about okay, am I solving the right problem? am I am I building stuff or delivering stuff at the right time with a sense of urgency? I think that's really how I think about it. And I think I'm I'm getting there.
Brian Bell (46:15.981) No. Well, really enjoyed the conversation. It's great to catch up. where can folks find you and Zive online?
Rajesh Gopi (46:22.907) The best is either LinkedIn, so I'm most active on LinkedIn or send me an email at Rajesh at Zyve.ai.
Brian Bell (46:32.037) Thanks for Josh. It was great to catch up.
Rajesh Gopi (46:33.872) All right, likewise, Brian, thank so much.
