How Agents Cut Hour-Long Data Checks to 5 Minutes - Amol Kapoor & Clifford Ressel, Nori Agentic
https://www.youtube.com/watch?v=kIlVNGZZr8o
AI Engineer · 19 min · 226 sentences · sentence-level transcript from YouTube's caption track; each line begins at the time shown

[00:01] [music]
[00:12] Hi everybody.
[00:13] It's really really great to be here on stage at AIE New York.
[00:16] This is kind of awesome.
[00:17] Um today we're going to talk a little bit about the Agentech SecMaster.
[00:21] But before I dive into things like introductions, I want to talk about something that's been on my mind recently, especially as I was preparing for this talk, which is finance has gotten kind of weird, right?
[00:31] Like when I was growing up, finance mostly meant public markets, you know, trading floors, ringing bells, things like that.
[00:39] And that's not really what happens anymore.
[00:43] There has been a huge influx of capital investment into private assets.
[00:47] And some of these numbers are kind of staggering to me.
[00:50] Apparently last year, there were $240 billion of secondaries volume in 2025, which is up 50% from 2024.
[00:58] I'm sure in 2026 it's even more than that.
[01:00] And people are now forecasting that we're going to do $32 trillion of alternative assets under management by 2030, which is only 3 years away.
[01:08] Now, why is that happening?
[01:10] Why is this the big story in the finance world?
[01:12] Well, some of this is because AI seems to increasingly be the only growth area in the broader economy.
[01:19] So more and more people are holding on to and investing into all of the things that feed the AI world.
[01:25] That's land, energy, data centers, GPU clusters.
[01:29] And of course secondaries, right?
[01:31] Everybody is holding on to OpenAI and Anthropic secondaries.
[01:35] They're everywhere.
[01:35] All of these private AI companies, everybody's trying to get a hold of them.
[01:39] But it's not just an AI story.
[01:41] Some of this is also a regulation story and a story about seeking alpha.
[01:46] High growth companies are finding less and less reasons to go public.
[01:50] They would rather just raise a series F. More control for the founders, less paperwork to do with the government.
[01:57] And if there are these massive well-funded firms that are taking over and squeezing the juice out of public markets, well, you're just going to start looking for edge somewhere else.
[02:07] As with all things, I think Matt Levine said it best.
[02:10] He actually wrote this a year ago.
[02:12] Private investment firms have become huge, and the legal distinctions between public and private markets have become way less important.
[02:20] If an increasing share of the economy is in private assets, investors aren't able to just index.
[02:26] You can't get the market return by buying every public stock if half of the market is actually in private companies.
[02:33] So, you end up with this cycle.
[02:34] There's more investment in private assets, which means more people have to invest in private assets just to make sure that they're getting the returns from the wider market.
[02:42] Now, one thing that's really frustrating is dealing with private assets.
[02:46] It's actually really, really hard.
[02:48] How do you manage private assets?
[02:50] How do you keep track of things like how many shares you own or what your strike price is?
[02:53] I mean, I'm a startup founder, and I find that frustrating just for my own little bit of stock.
[02:58] And if you invest in a plot of land next to a hydro plant, things get really weird.
[03:02] Where is all of the information about how much you paid and what your rights are?
[03:07] Well, unfortunately, it's in PDFs.
[03:09] It's in a box folder or in an Outlook email.
[03:13] It's in a Zoom call transcript.
[03:14] Sometimes, it's on a napkin with a hastily scribbled signature from a bar overlooking Columbus Circle.
[03:21] Welcome to the nightmare world that is managing private asset data, something that more and more people in this room and in New York in general have been dealing with over the last couple of years.
[03:30] My name is Amol Kapoor.
[03:32] I am the CEO of a company called Noriagentech.
[03:34] We build cloud agents, and today we're going to talk about how you can use cloud agents and AI in general to make it easier to manage, record, and audit a company's investments.
[03:43] In particular, a company's security master database.
[03:46] Now, my background is actually in AI.
[03:48] I've been an AI engineer for about 12 years.
[03:51] Uh four and a half of those years were at Google.
[03:53] So, funny enough, I am not the best person to actually talk about what a sec master is.
[03:57] But, I'm very lucky to be joined by my CTO and co-founder Cliff, who actually has spent some time in back office finance here to talk to us a little bit about what a sec master is before we dive into how AI can solve those problems.
[04:19] Great.
[04:19] So, I think it's helpful before we before we talk about where this is a hard problem, I think we should get a lay of the land.
[04:24] We can talk about a simpler example with public equities, then we can tackle some of the complexity that comes with private assets.
[04:31] So, every asset manager right now currently runs on a sec master.
[04:35] So, let's define that.
[04:37] Asset managers need to know what exists in their investment universe or in their tradeable universe.
[04:42] This is a critical piece of back office, brings together engineering, operations, and data.
[04:48] You can think of this as just a master database of all securities or instruments that represent the assets they're holding.
[04:55] For the engineers, you can think about this like any master database, it's the source of truth, and usually the starting point for most of the business critical analysis.
[05:03] An investment that looks like risk management, P&L accounting, regulatory compliance, all these functions that are critical for asset managers.
[05:11] So, say two different sources of data disagree about particular values.
[05:16] Ideally, that conflict doesn't get flagged before it reaches the sec master.
[05:22] You can go back to the sources and pinpoint either the divergence in the particular value, or sometimes it's a logic problem and you can examine how each one maps to that particular security.
[05:33] So, when done correctly, upstream problems are fixed before they contaminate all the business critical analysis we just talked about.
[05:40] So, I like to think about this as an organizational primitive for how you handle data to enable all the consumers of data at your company.
[05:46] This layer separates the data producers from the data consumers.
[05:50] So, whether you're working on a complex medallion data lake house that has many different tables, or it's just a single denormalized source of truth, this is the bottleneck where you can enforce the standards and the visibility
[06:02] and the quality checks.
[06:06] So, depending on where you invest, you need to take this simple concept and match it up with how everyone else in the world who invests in this particular asset class,
[06:16] potentially speaking different languages languages, settling in different currencies, and operating in different regulatory regimes.
[06:25] So, for public assets, you benefit from a couple different points of centralization.
[06:29] For the needs of exchange and transactions, you have data vendors who will provide different identifiers.
[06:35] So, these might be like Bloomberg figgies or LSE RICS.
[06:39] Then there's also regulatory needs.
[06:41] And though that point of centralization causes numbering agencies to offer a different set of identifiers, like ISINs or CUSIPs.
[06:48] So, you're able to use this information to map anyone's view of a particular asset into your own view.
[06:53] And that's the only way you can understand your positions, like modeling risk or reporting holdings, joining alternative data, creating financial models.
[07:03] So, let's take a very simple example.
[07:05] You hold some mega-cap publicly traded stock like Google.
[07:08] So, already if you know the tickers, is this GOOG or is this GOOGL?
[07:11] Because there are two different share classes traded on the same exchange.
[07:15] Let's say you need to join in some data that exists only at the level of the legal entity.
[07:20] Well, if you're indexing based off of ISIN, now you can't join that anymore.
[07:24] Or maybe you only trade one share class and you only hold one, but you need to trade it across multiple different exchanges because you have different funds for the needs of multiple
[07:34] multiple funds or multiple investors.
[07:36] So, already this is getting pretty complex.
[07:40] I like to think about this kind of like a daytime library as a comparison for the engineers.
[07:44] It seems very simple when someone solved the problem well.
[07:48] But the moment you're either, you know, working with a new language that hasn't implemented good daytime yet, or you just need to make some modifications and peel back how it works inside,
[07:55] it's all edge cases and complex ontologies.
[07:59] So, this is what that Google example might look like in practice.
[08:01] Here's a denormalized table with many rows for that one equity we want to uh able to now analyze.
[08:08] So, if some data comes in, even if it only has a subset of identifiers or exists at a different granularity than your Sec Master, the code can still map it.
[08:18] So, here's where it gets interesting.
[08:20] If the code if the data coming in doesn't have regular identifiers, how do you map it?
[08:26] It's actually pretty clear how AI helps in the first case, right?
[08:30] It's very easy to reason about this data with structure.
[08:32] The coding agents are increasingly good at writing code and if you have the source of kind of gold standard ontology about the business critical data for your company,
[08:40] [snorts]
[08:41] AI can enable new use cases very easily.
[08:43] But with unstructured, potentially not identified data sources, you have a huge burden of manual verification and manual mapping.
[08:52] So, this is where the happy path breaks down when private assets come into the Sec Master.
[08:58] So, when you're buying the Google stock, you know what you're paying, you know what you're getting, and going into the transaction, everyone has a shared language of different identifiers that they can speak.
[09:08] Let's imagine you're going to lend some money to a company building a power plant near some data center developments or you're investing in land near the power plant for potential future
[09:18] uh development itself.
[09:20] There's no identifiers everyone shares going into that deal.
[09:23] There's no one that regulates the entire structure of information you might get out the other end.
[09:27] And there's no one data vendor that has a single API where they've thought through the ontology of these entities in particular.
[09:34] Everything from the repayment schedule to the collateralization, all of this lives in one-off agreements or amendments.
[09:41] It's not fungible.
[09:42] The asset is captured and defined entirely by the documentation.
[09:47] And this is interesting because now the data scales almost one-to-one with the assets and the burden to scale the the Sec Master itself is constantly increasing.
[09:59] Great.
[09:59] So, I'll pick up back here.
[10:02] Now, Cliff makes a really good point.
[10:03] The private assets are an absolute pain to manage.
[10:06] They just don't fit into the Sec Master all that well.
[10:08] But all that information still needs to actually be collated in some way.
[10:12] It's not like your audit requirements and your cleaning requirements and your standardization requirements disappear.
[10:17] If anything, it's actually a lot worse now because it's just so much less standardized.
[10:21] You can't rely on the people who are doing your risk modeling to know whether or not that data is ground truth.
[10:26] You certainly don't want them to be the ones going through and munging through all your documents.
[10:30] So, somehow the data that lives in that PDF somewhere, that information needs to make its way back into your SecMaster.
[10:38] And right now, that is just an extremely high labor task.
[10:42] I mean, generally it just takes a lot of people hours and a lot of domain expertise.
[10:47] I'm also going to really quickly move this mouse out of the way.
[10:49] It's been bothering me.
[10:51] All right.
[10:52] So, I want to ground this in a specific example.
[10:55] We work closely with a Fortune 200 asset manager here in New York, one of the benefits of being New York local.
[11:01] They recently started owning a bunch of private assets, who hasn't, and it was causing a lot of problems for their data engineering teams, the people who were maintaining their SecMaster.
[11:10] Now, when any assets in their SecMaster changed, they had to make sure that change was correct.
[11:15] So, they had a system.
[11:17] They did a day-over-day change analysis of their table.
[11:21] If you have a table with 50,000 assets and each one of those assets has something like 200 columns, even infrequent changes are going to happen on some daily cadence.
[11:31] And every single one of those changed cells was then sent to a team of people to triage because, again, all of this data is super unstructured.
[11:39] It just lives in documents somewhere.
[11:41] So, that team was then responsible for going through docs and going through various data sources to determine if any given change was good or bad because upstream data providers
[11:52] may have fat-fingered something or may have gotten something incorrect or they may have just been conflict.
[11:57] And I want to highlight that this is not just every row, not every column, literally every single cell.
[12:03] So, if an asset changed price or if an asset changed its name or if a company went bankrupt or if it changed ownership or you renegotiated a contract or whatever,
[12:12] every single one of those changes in that investment universe would have, on average, an hour of work behind it.
[12:17] An hour of somebody somebody's time, somebody that you're paying to go and just lunge through data.
[12:23] That is hundreds of thousands of hours every single month.
[12:26] Hundreds or thousands, sorry.
[12:29] So, we're at AIE, how would an AI engineer solve this problem?
[12:34] Two words.
[12:35] Cloud agents.
[12:37] What are cloud agents?
[12:38] This is more for the finance folks in the room.
[12:40] In simplest terms, a cloud agent is a coding agent like Claude Code or ChatGPT, but instead of running on your local machine, it lives and runs in a remote environment instead of on your computer.
[12:52] Because it lives on the cloud, we can do a lot of really interesting things.
[12:54] We can uh assert a bunch of audit requirements, we can get auditability and trans um transparency, we can get observability, and more importantly, we can have it spin up and do work on a schedule without a person being involved.
[13:06] For example, when new data hits the SEC master.
[13:10] So, whenever there is a change in the investment universe for this particular asset manager, a Nori cloud agent will spin up, in this case running a locally deployed open source model under the hood, so all of the data stays on prem.
[13:22] There's no concerns about information getting out of the underlying environment.
[13:26] The agent here has explicit instructions to figure out how the source uh figure out the source of this change.
[13:32] And it also has access to a few useful tools, Box for PDFs, public APIs like Bloomberg, and even a small segment of email.
[13:41] The agent then assembles an explanation of the change, gets citations for why that change happened, and then provides a recommendation for whether to keep or discard it.
[13:50] And finally, it pings the owner of whichever cell changed in Slack for review.
[13:55] So, you have this agent that is operating basically entirely autonomously, providing tasks for a human that the human just has to say yes or no to.
[14:03] And of course, if necessary, the human can go and do further research to make sure that the information provided was correct.
[14:10] Now, you still need a human to review those changes.
[14:13] Of course, this is necessary just for accountability among other things.
[14:17] But, this kind of work dramatically cuts down the response time because most of the previous slowness was just people having to manually search through a bunch of documents.
[14:25] Agents are really, really, really good at that, and humans are pretty bad at that.
[14:29] So, if most of the time is just assembling a case, well, you can take something that used to take an hour and make it just 5 minutes of human work.
[14:38] But, wait.
[14:38] It's worth pausing just to ask, how does the agent actually get domain expertise?
[14:43] How does it know what a good change looks like?
[14:45] I'm concerned that some of the folks in the room might think, "Oh man, I have to go and build this like custom agent infrastructure that's going to do this really, really specific one-off thing."
[14:53] Well, I don't think that's actually how these things should work.
[14:56] Agents get information the same way a new analyst does, from written instructions about where documents live, which sources to trust, and how to judge a change, all in English.
[15:05] Agents are pretty good at self-modification these days.
[15:08] So, you can change how an agent behaves by just editing natural language English markdown files.
[15:13] You use a generic agent like Claude Code or Open Code or anything that's open source like Pi, and then you skin those agents with configs like skills or agent MD files.
[15:22] And most of the customization actually comes from the environment in which the agent is running in, which of course is just about what kind of configs and what kind of credentials you're dropping on the machine.
[15:32] So, I want to take a step back and zoom out a little bit.
[15:35] Everything that we've talked about so far, this agentic SecMaster concept, this is essentially a database with a bunch of cleaner fish automatons that are dynamically making sure that everything is up to date.
[15:45] They can spot problems quickly and proactively surface them and even solve them in some cases, all autonomously.
[15:51] And one of the things I want to highlight is that this technology is not limited to SecMasters or even data cleaning in general.
[15:58] It can be useful across a firm because it is a generic primitive in which we are having agents do work that was previously loaded onto humans.
[16:06] In my mind, the incredible potential of AI tools in finance is that they can make a huge range of tasks proactive instead of reactive.
[16:14] Right now, every time anything needs to happen, and a human has to spend time and energy evaluating whether or not that task is worthwhile and then sending that work to an agent.
[16:23] If you ask me, the future of enterprise AI is not use a robot on your laptop to solve a problem.
[16:28] It's having the robots come to you with tasks, advice, analysis, and recommendations.
[16:35] I think that in order to make that future a reality, you need to use flexible infrastructure that can be used for any task and make it really, really easy to modify what kind of tasks it's being used for.
[16:45] Your SecOps master agent shouldn't be a one-off unique agent.
[16:48] It should actually be running on the same cloud agent infrastructure as your audit agent or your code review agent or any of the agents that your engineers are using for their day-to-day job.
[16:57] Each of the Each of these agents will have different instructions and tools and triggers, different context.
[17:01] That, of course, is owned by you, by the company.
[17:04] But the orchestration, the security, the integrations, all of that should actually just be the same across these different tasks.
[17:10] And really, in my mind, should just be owned by a vendor because everybody is inventing these problems or solving these problems over and over again in the same organizations.
[17:19] So, what were the things that we learned by building AI for finance?
[17:22] I want to leave with three lessons that we that we had that I want to impart on everybody here.
[17:27] First, if you ask me, AI adoption is primarily a collaboration problem.
[17:31] People want to buy solutions, engineers want to sell building blocks.
[17:36] We have to find a way to meet in the middle.
[17:37] We found the most success when we work with brilliant folks who are experts in their domain, who are able to make the leap, that intuitive leap from cloud agents to wait, this actually fits into a problem that I have right now today.
[17:52] Second, I think it's really important to aim for 95% automation and it's okay to keep a human in the loop.
[17:58] People want to use AI to do 100% of everything and this doesn't work.
[18:02] It doesn't work in software, we've seen that with software factories over and over again, but it really doesn't work in finance.
[18:07] It's okay to have a human in the loop.
[18:08] In fact, I think we should lean into it.
[18:10] There are way more problems that you can solve if you're okay with having one human spend 5 minutes looking at something than trying to have an agent do something 100%
[18:18] end-to-end every single time.
[18:21] And finally, I want to point out that the cost of accidental complexity has dropped significantly.
[18:26] Fred Brooks used the phrase accidental complexity to describe the difficulty that comes from how a system was built as opposed to the problem it solves.
[18:34] And I think in the back office, a lot of work is actually shaped like this.
[18:37] Teams record the same data in different ways, and vendors use different formats, mostly because that's just how things evolved organically.
[18:44] I think agents are really, really good at handling this kind of work, this low-complexity, high-effort work, which actually then allows people to spend more time on the problems that are essentially complex, the problems that are actually hard.
[18:56] So, that's all I got, guys.
[18:58] Thank you so much for coming.
[18:59] Uh before I go, I want to just quickly highlight two things.
[19:01] One is if you're interested in more talks like this, we run a community called Agentics NYC.
[19:06] It is New York's largest agentic engineering community.
[19:08] We run regular events on building with agents for beginners through experts.
[19:12] And second, if you are interested in cloud agents, software factories, agentic back office, SecMasters, all these different things, uh you should reach out.
[19:20] Cliff and I will be in the hallway track immediately after this.
[19:23] Uh and of course, you can check us out at noriagentic.com.
[19:26] Thank you so much.