Investology: Investment Management Intelligence
00:10:18 · Andrea Carnelli Dompé, Andrea Carnelli Dompé, George Aliferis
[00:10:18] S1: and make it available on a software that also generates all the downstream [00:10:22] S0: analytics. All of that we can support. Some people will open the menu and say, actually, I just need you here. Yeah, we can support that end-to-end. [00:10:29] S2: And in practical terms, how do you insert yourself in this structure, in this process of communication between GPs and LPs? [00:10:39] S0: Well, the source of information is GP portals. [00:10:44] S1: And GP portals tend to run on a few systems that are quite common across the industry. [00:10:50] S0: But there's loads of exceptions. [00:10:53] S1: And those systems are not necessarily known for being very tech-friendly or open to [00:10:58] S0: third-party systems. So unfortunately, it's a wide range of techniques from having simulated logins to automated [00:11:10] S1: attachment extractions in the case of the notifications or sending the documents as [00:11:14] S0: attachments. [00:11:15] S1: There's loads of different strategies that are very specific to the GP portions. [00:11:20] S2: And how do I implement this as an LP? [00:11:22] S1: What is the first step in order to move from my traditional approach to your AI [00:11:28] S2: agent-driven approach? [00:11:30] S0: Typically, when somebody uses us for the end-to-end portfolio monitoring approach, [00:11:35] S1: it would all start by authorizing us to receive documents from GPs on their behalf. [00:11:42] S0: So that's step number one. [00:11:43] S1: Step number two would be tell us what you have so we can map out your portfolio.
00:11:43 · Andrea Carnelli Dompé, Andrea Carnelli Dompé
[00:11:43] S1: Step number two would be tell us what you have so we can map out your portfolio. [00:11:47] S0: Step number three is [00:11:49] S1: explain to us what kind of accounting monitoring reporting laws you have so that we can design the system in a very scalable way that goes and ticks all those boxes so you know we understand what you need based on that we work out what kind of raw data we need to extract and normalize and on top of that we work out also what kind of derived analytics you need so we really start with the clients and end goals And is this a high-touch sort of consulting process, or do you have a way to automate it? [00:12:27] S0: It sounds like a consultancy-like, but actually it's not. [00:12:29] S1: So we designed the system to be incredibly general and adaptable to different use cases. [00:12:35] S0: We are able to fine-tune it very well, [00:12:38] S1: but we are always using the same generalizable framework that we just roll out [00:12:42] S0: between new cases, which is one of our strengths. So it feels like it's just done for you, but it's not. [00:12:50] S1: And how are your clients able to measure the increase in efficiency and improvement? [00:12:57] S0: I think we have three main areas of impact. The first one is we reduce the amount of labor required for these tedious day-to-day tasks, [00:13:05] S1: which typically means that the people who used to do that are freed up to do
00:13:05 · Andrea Carnelli Dompé, Andrea Carnelli Dompé
[00:13:05] S1: which typically means that the people who used to do that are freed up to do [00:13:09] S0: something that is more value additive and value generated. So that's benefit number one. And that's in the order of, you know, for say a portfolio of 50, 200 funds, it's in the order of hundreds of hours per quarter. So that's a lot of the time. Secondly, we reduce the risks of missed data or incorrectly extracted data, and that, depending on the downstream implications, can save a lot of money and effort. And thirdly, more broadly, we make data more available, more quickly, more accurately, in general. So typically, clients not only automate what they have already, but are able through us to extract more data more promptly, and this is unlocking loads of new use cases. Classic example is look through data. [00:13:57] S1: Some LPs are so sort of stretched to simply process the minimum they're necessary things such as processing calls and working out valuations that they kind of don't have the time and bandwidth to extract look through information from their [00:14:13] S0: investments. But that's key to understand how your exposures are building up, where your performance is coming from, whether the GP is doing exactly what he said he would do. So there's loads of value that can be generated there. And typically when people come to us, since we make that information easily available, that's an extra edge that they gain to us.
00:14:13 · Andrea Carnelli Dompé, George Aliferis, Andrea Carnelli Dompé
[00:14:13] S0: investments. But that's key to understand how your exposures are building up, where your performance is coming from, whether the GP is doing exactly what he said he would do. So there's loads of value that can be generated there. And typically when people come to us, since we make that information easily available, that's an extra edge that they gain to us. [00:14:32] S2: So the conversation got a bit fixated on the process and the pipes. Ultimately, it's about spending more time on the things that matter the most, making better investment decision, having clarity and so on. [00:14:45] S1: And I would like to understand more the technology that you're using in order to deliver that. [00:14:50] S0: First of all, I mean, there's lots of buzz around AI, of course. [00:14:53] S1: I like to always remind people why we think it's transformational for private [00:14:57] S0: markets beyond the buzzword. [00:15:00] S1: And the problem that AI solves specifically for private markets is exactly the problem of data being so siloed on different sources and being so inconsistent [00:15:10] S0: across formats. [00:15:11] S1: And AI is particularly well suited at dealing with this messiness of sources and formats. [00:15:17] S0: So that's the kind of real big reason why we're investing in us in that direction. When it comes to agents specifically, It all boils down to the definition. [00:15:30] S1: So we kind of think of agents as being software systems that are smart enough to think for themselves and to make actions with some degree of independence and
00:15:30 · Andrea Carnelli Dompé, Andrea Carnelli Dompé
[00:15:30] S1: So we kind of think of agents as being software systems that are smart enough to think for themselves and to make actions with some degree of independence and [00:15:40] S0: autonomy. [00:15:41] S1: And the key words here in this definition are autonomy and taking actions, [00:15:45] S0: making decisions, taking actions. Those are the two key ingredients. [00:15:49] S1: What makes them different from AI workflows or AI co-pilots or the first wave of AI that we saw? [00:15:56] S0: we saw in the first years when everything changed in 2020. So how do we design them then in practice? So I think we follow three principles. One is we hyper-train these agents to the private markets use case. [00:16:11] S1: I think horizontal agents that kind of do everything across different use cases and sectors are not as powerful as hyper-training the system to a very specific use [00:16:23] S0: case, vertical use case, vertical language, logic. So specifically in private markets, they need to understand the jargon. They need to understand the implications. [00:16:32] S1: They need to understand that 95% cold means that you can take the commitment, [00:16:37] S0: multiply by 95% and extract the value of cold. So there's a lot of jargon that they need to understand. [00:16:43] S1: And a lot of rules that they need to understand to be able to look at private [00:16:47] S0: markets data in an effective way. [00:16:49] S1: Just like somebody needs a long training at university and in the industry to
00:16:49 · Andrea Carnelli Dompé, Andrea Carnelli Dompé
[00:16:49] S1: Just like somebody needs a long training at university and in the industry to [00:16:55] S0: actually appreciate what's inside this document. Same thing for agents. So that's one hyper-trained to private markets use case. Secondly, we like to be LLM agnostic. There's a lot of innovation happening in the market. There's a new release every week. [00:17:10] S1: Not all those models are improvements versus the previous versions, first of all. [00:17:14] S0: Not all of those models are good at handling all tasks. [00:17:18] S1: So we have built a system that basically identifies at any point in time the best available LLM for the best use case and kind of integrates them all in a crazy [00:17:28] S0: framework. So that's kind of principle number two. [00:17:32] S1: And principle number three is for mission critical tasks like finance in general, [00:17:37] S0: not just private markets, it's very important to have a very high level of accuracy. [00:17:42] S1: And the way we sort of cope with that is we make sure that we support fully the [00:17:48] S0: human in the loop. for final decisions on critical action items. [00:17:53] S1: So we have thought about the system in a way that it becomes incredibly easy to [00:17:57] S0: audit what the AI has done, why, [00:18:00] S1: and how exactly it has made implications from documents so that it becomes very [00:18:05] S0: easy for a person to click, yeah, okay, I agree with this implication. [00:18:09] S1: And that is a massive component into the productivity gain because I think a system