Episode 085: What Does Machine Speed Do to Private Markets?

Hi, this is Pat O'Mara. I'm the chairman of CEO of Inventium Capital Partners, we own Inventium I/O, and we recently launched Inventium Chain. We have interested a number of other entities as well, but we're super excited to be with you, and I'm joined by Richard Walker, who leads our Inventium I/O product. Richard, you want to introduce yourself real quick? Yeah, that thanks. Great to be here with you again after a number of years. Great to be on the team this time, not just a visitor. As I said, I'm leading the Inventium I/O business under Inventium Capital Partners. After a number of decades and managed for consulting, advising large firms on the use of digital technology to transform financial market infrastructure, I've joined Inventium, and I'm super excited about what we're doing. We're going to talk about today, regarding the I/O platform, as well as Inventium Chain. And Richard, we're really talking about digital, intelligent, private markets, and AI has come out with all sorts of capabilities. It can process at speed, it can interrogate at speed, it can do data matching at speed, but it can also deceive at scale. And that's a detection function. If it's not matched up with a mechanism to provide proof or show or verification at scale, why? And that's the connection with AI and blockchain that Inventium has been working at for quite some time. Our focus, which is unstructured document data, where we're managing and administering the lifecycle of the lingua franca of private markets, which is where all the value is in the context of this document where it lives. And this really is what Inventium I/O is, is that right? Yeah, Pat, what we've done with Inventium I/O since the conception of the business has really brought forward and implemented the text that capability to deliver unstructured document data in a way that can be verified for origin, for process, and for state, and used by diligence teams to reprise and value private market assets dynamically. And we've always held that this would lead to systematic trading of private market assets. And it wasn't until what we're seeing in the demo that we're going to discuss with the introduction of a gentick AI that not only do you have the data that you can act on, so the data exists in a form that can be used more efficiently. The efficient use of that data is achieved through AI. Now, if I think about the private market assets overall, Pat, I just want to add that the biggest issue with private market assets is the cost of verification. The collecting the documents, the verifying of the right documents, the complete document said, it's a current document said, going through that, and you've got two trusted parties, a buyer seller in the private markets. And does the buyer trust the seller? And so the buyer spends a lot of time with other parties to prove the veracity of that data around origin, process, and state. And so what we did was we, through the history of Inventium I/O, is we brought that forward in a way that it would be red index, vector, and betted, hashed an anchor on chain. So you would know when you looked at the data, that the provenance, and all of those things that I've mentioned, around origin, process, and state. But that wasn't enough. It wasn't on its own. It wasn't enough. It didn't reduce enough of the friction and the cost of verification, because you still had teams using that data. It's valuable. No question. But value at scale is now achievable through the machine reading, machine activation, using a GTI to be able to do agent to agent negotiation against the buyer seller set of interest with the underlying data sets, which is a big unlock for scale. And the large data set pattern recognition that what AI is able to do, where it's at scale, where you have these agents able to do specific functions, where I have an agent. And we're going to show this where one agent using Anthropic, one is using JTPD, both validating the data back and forth, anchoring on and venom chain, the proof of verification. And what happens is not at a single moment, seeing those areas of chain to state of data, indexing it, vector embedding it, and then in many cases now, we're going to be delivering in the very near future, the ability to have the semantic chunks anchored as well for a registry within the document. But what we're doing is we're administering this again, and again, seriously. So we create memory of the performance of the asset. So we can see aberration, as many data points relate to one another. And what this is doing is it is reducing the long poll and the tent for a real estate transaction or a private equity transaction, private credit transaction, where infrastructure transaction, which is the diligence. I can have my agents running validating data, letting you know it's been validated by my agent. You can go on the other side, have your agent do it. And this is what this video is doing. It's literally giving you the ability to have agents processing data at scale, delivering back initial feedback that used to take associate six months. I don't know how I'm going to take minutes to deliver back and deliver it to the real intelligence to make decisions with and argue over the one thing that matters, bid versus ask, because you've validated all those other external data points. You know what else I love about it, Pat, is on the video we show-- the artifacts that being built up dynamically through the processing of AI so that when you challenge something, like these documents don't match, you have the full justification, rationale, and evidence that they don't. Instead of an analyst going through and working to find out that up, I found an error. And then the other party saying, what, air, I don't know. Show me how you got there. And having to do all the work again and rebuild it, it gets built and it gets archived as a natural consequence of going through the process. And it really is an environment, a trustless environment, where you have AI that is going to have an agent that's going to process data at scale, at speed, review it, say, here's what we've seen, here's what we've verified, anchor that on chain, give conclusion, to the other agent, your agent, doing the exact same work against a different LLM, asking different questions and different props from your viewpoint and validating what we're stating or rejecting it. And this really is what's amazing. And there is no way for us to work in an agentic world, agent to agent, intra and intra organizationally without having a mechanism to commute trust in the origin of the data, the process of the data, the state of the data that was gathered and then delivered. Does that make sense? Yeah, totally makes sense. Completely agree. It's so this video that we invite you to watch is showing two agents verifying data on a real estate transaction. I think what Albert's showing you is fantastic. This is going to take you about four or five minutes. And we really appreciate your attention to it and watch our video and come back with feedback. And if you'd like to be a user of Inventie Myo, please reach out to Richard or if you'd like to have feedback or develop a pond inventie Myo, reach out to me. We're excited to expand our ecosystem, expand our user base and see this being deployed in the markets real time. And we're also running a closed beta currently on Inventium Chains. So if you're going to NVNNChains, you know, you can apply for the closed beta there and participate in how your agents can use that test agents on chain, calling from any other agentage communication just like we saw in the video here, to get a proof against a piece of data so that your agent can complete a transaction knowing that data is most accurate. The state is known. The process and origin have been proven and that you can transact. And if we can do this where we're extracting the real value from unstructured document data, where that data lives, it's not leaking, it is secure, it's in the client or your own security envelope. And then you're permissioning people into interact with it, permissioning other parties agents to interact with your data. And as you do that using InventiumIO, you drive the ability to understand the asset that drives price discovery with price discovery, you drive trading and liquidity. And it is the first step which is when divinium kind of lives, which is in surveillance of an asset, price discovery and that ongoing performance of data so you can understand the state of that asset. Yeah, and what last comment on this because this is a long close, but adding the MCP server services to InventiumIO has unlocked a ton of new value for existing clients and future clients because going through the process of reading that data in its location where it lives, not uploading data, not compromising sovereign data for monetization by other parties, but keeping the data where it lives, reading it, indexing, vectoring, betting it, hashing it and then anchoring it, allows agents to then inspect that data in new ways that we haven't historically. So is the data set complete? Are there blackout periods or missing rep roles through certain periods? Do we have resolution on leans that existed, but then disappeared out of the data set? So you can inspect that data set for completeness for anomalies, AI has long been graded anomaly detection, and know that data set's complete. With the hashing of that data down to the semantic trunk level, when a piece of data changes and in the document set. that hash will then automatically get picked up and reanchored on chain that reanchoring on chain can then send an alert to anybody who's monitoring that to go in and not only know that something changed Know exactly what changed what field change where and know whether it was material or not for instance Speaking where they affirm this incident bringing the private credit portfolio onto the platform and To know if something changed in the underlying private credit agreements Do I need to rerun a risk rating if it's a small thing if you don't know exactly what change You're gonna spend money and cycles in time rerunning the risk on that asset when if it's a if it's a small change You'll know exactly what change when and you can act accordingly I think that the notification period. I think the complete document set I think the semantic chunk level but irrigation understanding is a whole new Set of capabilities beyond just adding an MCP server and adding and Viniam chain to the stock It really is and we're excited that we're gonna go through a series of These podcasts we're gonna try and keep them short for you We're gonna attach videos. We're gonna be looking at specific function both on the technology and market function that we're seeking to Accelerate but we're excited for all of you to be with us Richard. Thank you for 20 time with us Thank you [BLANK_AUDIO]