Good morning. Good afternoon. Good evening. Good whatever time it is. Wherever you are in this amazing planet of ours, I’m Malcolm Hawker, and I’m the host of the CDO matters podcast.

Thank you for listening. Thank you for watching. Maybe you’re on YouTube. Thank you for subscribing.

Thank you for downloading all of it. It is your engagement with this growing community that makes all of this possible. Hey. If you are into fintech, if you’re into financial services, if you are in the banking and insurance industry, and you’re looking to find ways to succeed with AI, maybe you are wanna get over that ninety five percent failure hump of POCs of AI, Maybe you’re looking at deploying some custom chatbots.

We’ve got a guest for you today. I am excited to be talking to Jeremi Carnell, who’s the head of data solutions for InvestNet. Jeremi’s joining us today from a hotel room, so thank you for the dedication to the pod here, Jeremi.

Usually a resident of my old hometown of the best city in America, Austin, Texas.

Yeah.

Thanks for joining, Jeremi. We’re we’re excited to have you here today.

I am excited to be here. Thank you so much for inviting me.

So Awesome. Awesome. Well, let’s let’s get into it. First and foremost, you know, I mentioned fintech.

I mentioned financial services. Yeah. You exist in a world where you are building and deploying database data dash solutions to a network of advisers. Correct?

Yeah. So help under help the help our listeners and viewers understand kind of your overall operating model. Yeah. How do you fit in the organization and the stuff that you, as the head of of data solutions, are building day in and day out?

How do all those fit together?

Yeah. Happy to. So I I am in a fortunate position. I I came to InvestNet through an acquisition. I was the CEO and founder of a company called Trulytics that InvestNet acquired back in twenty twenty two.

And right after that acquisition, they gave me the privilege and honor of working with some of the best data people I’ve ever been able to work with in my life.

And we get the privilege of taking the digital exhaust that comes off of six point five trillion dollars of assets under management and turning that both into business intelligence, so descriptive and diagnostic analytics. This is a lot of our fee flow performance benchmarks. But rearview mirror looking, and that’s what we’re known for actually leading up to, I would say, the last couple of years. That was our sweet spot is all of the business intelligence that we were able to generate from that data and provide back to three primary core segments: asset managers, wealth enterprises and their advisers, and RIAs.

And over the last about three or four years, we’ve been slowly making pivot not slowly, actually. I think we’ve been going pretty rapidly making a pivot to more disproportionately focus our attention on what we’re calling decision intelligence. And that is a collect and we’re trying to elevate the conversation within wealth and in financial services to get away from a looking at AI as like a tool set that’s episodically used and and look at it more strategically from an decision intelligence perspective.

You know, I think there is a really a sort of an overwhelming sort of focus on the use cases like note taking apps, things of that nature, and less, unfortunately, a focus on where AI and machine learning, deep neural networks can really weigh in and and and and take advantage of the mass amount of data that this industry puts off every year, like every day, you know, the trades, the positions, the valuations, all of that.

The the data stream that is a result of all of that on an on an average book of business is, that’s a a signal universe. That’s five thousand times greater than something like AI note taking. And and that’s where, really, AI and machine learning begin to shine. And so that’s and so that’s sort of why we are marching down this path of of framing everything through the lens of decision intelligence because it’s it’s let’s be honest.

Investors, clients don’t hire advisers because of how fast they are or how speedy they are. They hire them for their judgment, for their discernment, and helping them make some of the most important decisions of their life. And so we, with the data we have, with the approach we’re taking, with the the stack is is prescriptive, predictive analytics, so what hap what’s gonna happen? What should you do about it?

Paired with machine learning, deep neural networks, knowledge graphs, and AgenTic AI to pull all of that together in in really compelling experiences that are homegrown on the investment side or provide that data back to enterprises and advisers in in whatever interface they use on a regular basis, what their desktop looks like or or how they consume that. We wanna meet them where they are. And so we’ve made that also pretty easy.

So Well, so growing up, when I was a little kid, I had a giant box of Lego.

And I I was addicted to the Lego, and what I would do with it is I would literally just dump it out on the floor, and then I would sit there and, you know, something would come to me. It was almost always a boat. Sometimes I would build a car. But what you just said to me is you’ve got the biggest box of Lego, like, ever.

Yeah. And how do you go about figuring out what you’re gonna do with all of that, what you called the exhaust? Do you are you working backwards from those decisions? Is that kind of the rubric that you applied?

And say, okay, our advisers need to decide how to allocate capital, so we’re gonna build a model that best recommends how to allocate is that kind of you’re working backwards from the decisions? How do you prioritize?

Yeah. So I I mean, we we first were wanting to diversify away from just business intelligence. Right? And so we’re like, what what is the thing that we’re trying to solve for an adviser?

The the the problem is that they have a book of business. Let’s say, I don’t know, the average adviser has one hundred and fifty, two hundred clients. The reality is is that they only really engage with a small percentage of them on a regular basis. They just don’t have enough time.

They don’t have enough time to meaningfully understand what is going on on a daily basis for their entire book of business. It’s complicated. There’s there’s a lot of financial planning objectives that they need to follow through with. There’s a lot of holding opportunities, like what brokerage accounts may be better served in a managed account, what held away accounts exist that I may be able to bring in, what if where’s where’s their money in motion?

Like, where is there an opportunity for me to help my client better invest that money than spend it, you know, on something frivolous or something that’s not gonna generate a meaningful return? So there there there’s there’s a thousand different things that an adviser faces on a day to day basis, and we thought, okay. With our dataset, what can we do to help them better make decisions? And and so that was with what we decided to do is build what we call our insights engine.

And we have a whole portfolio of business rule driven insights, machine learning predictive and prescriptive insights that well, I think we generate twenty five million next best action insights a day that we offer back to our advisers. And so now it’s no longer, like, with business intelligence, rearview mirror looking sort of analytics. That stuff is, you know, you look at that maybe biweekly, monthly, most likely quarterly.

We wanted to give the advisers something that they could that was daily active use. You turn on your computer. You get to see a beautiful dashboard that is organised across a number of of of different categories like tax management or non managed wealth or trust or CRM, held away assets, annuity, financial planning, any and then there’s tonnes of next next best actions that fall underneath that and allow them to really understand and prioritise where they’re gonna spend, you know, their time that day, show them the impact that that’s gonna have on their customer as well as the impact it’s gonna have financially for them, how they’re gonna grow organically.

And so it’s it’s it’s an it’s a it’s really an operating engine, a growth engine. It’s turned out to be for advisers. Those that leverage it literally grow thirty, thirty five percent faster than those who don’t. We’ve got all of these amazing lift metrics that we can point to, And we’ve got this longitudinally.

Like, the last four or five years, those those companies that and advisers that leverage Insight, they see a hundred and eight percent lift on something like driving UMA tax overlay assets versus those who don’t. It’s this year, it’s two eighty percent lift on FSP tax overlay assets. It’s it’s, like, twenty two percent year over year growth in driving brokerage to manage assets first. Like, it’s it’s it’s meaningful to their business.

And and because of that, we have enterprises now driving a ton of data back to us that’s off platform from InvestNet, which is like like, because InvestNet only has a certain amount of data. I mean, it has a lot of data, but it’s a lot of it’s the managed account data that that’s on our platform. There’s a lot of insurance and annuity data that we don’t have. There’s a lot of four zero one ks data that we don’t have.

There’s a lot of brokerage data we don’t have. And and because we have followed this path of solving for what the advisers need to do on a daily basis and making it their just making them more effective, it’s resulted in enterprises now saying, great. I want my entire book of business run through those models and and and and then open this up at scale. I’m not just gonna give this to five hundred of my top advisers.

I wanna give it to all four thousand of them and just watch, again, the revenue hockey stick just go.

So Well, so there you have it, my friends.

If you are doubting the ROI of investments in AI, if you’re doubting the ROI or maybe you still think that you can’t necessarily even quantify the investments in data or data management or data quality, here’s some advice. Go empower your salespeople. Yeah. Give them some tools to help them sell more.

And you don’t have to be in Jeremi’s business. You don’t have to be in financial services. Maybe you’re selling, I don’t know, eyeglasses. Maybe you’re selling iPhones.

Maybe you’re who knows what you’re selling? But we’re all selling something. And if you can go and help your salespeople sell some more, you’ll be able to to do a before and after, and you’ll be able to sow usage of your analytics, usage of your insights, and you’ll be able to draw a hard line between what you were selling before and what you’re selling now. Like, I so I Yeah.

I mean I mean, to to me, that that is the like a a dream scenario that you’re able to tie back to specific revenue goals that you’re doing just through this one thing, this next best action. Not saying it wasn’t hard. I’m sure it was quite hard. Yeah.

Let’s let’s talk to that a little bit. I mean, so obviously, you’ve got a tea you’ve got a team of data scientists. Yep. What does that team look like? And do you have do you have people that are working on integrations, infrastructure? What does the team kinda look like?

Yeah.

So we’re we’re organized from a data solution perspective sort of bifurcated between data strategy and acquisition and management versus data product and commercialization, sort of two legs to the the stool.

And so on the on the data strategy acquisition and management side, they’re the ones that are, you know, that are bringing in all of the data, that are replicating the data from our trading platforms into Snowflake, running it through both our business and decision intelligence models. They’re the ones that are that are actually also, you know, driving a lot of the new insights manufacturing. Like, they they manufacture new insights, and and and they’re spending a lot more time on the data science side doing the prescriptive and the predictive analytics. Like, they we just rolled out a couple of of predictive next best action insights that with ninety percent accuracy, we can predict, I think, twenty two billion dollars of money in motion events. With seventy percent accuracy, we can predict, I think it was, like, forty five billion dollars of held away assets.

And all of that is data that they’re bringing in, data that, actually, I think one of the biggest, most strategic pieces of intellectual property Envestnet has right now is the knowledge graph that they built off of that data. So that group, they took, again, digital exhaust off of a six you know, trillions of dollars of assets under management and created and so you have over a hundred thousand wealth advisers that make up the adviser knowledge graph. So that’s a third of the wealth industry.

And you’ve got twelve million clients that make up the client knowledge graph. And the ability for them to create really robust profiles, client and and and adviser profiles, have led to some really, not only really cool predictive insights, just amazing propensity models. And and, like, providing those back to Salesforce is, for example, and how we operate with asset managers and helping them focus on where their wholesalers should spend time with advisers, we just released our at the beginning of this year, based off of the data that was derived from the knowledge graph that we put into place, we had about five thousand predictions, ninety six of which resulted in at least one buy transaction. Ninety percent drove new account sales.

Our prediction, we were able to predict at the beginning of this year with, again, between ninety percent to ninety nine percent accuracy across products like FSP, SMA, direct indexing, almost seventy percent of the net flows that came into InvestNet. Like, we knew the adviser and what they were gonna buy six months ahead of time. And so we’ve got and we just we’ve taken that same approach, and we’ve created a new high net worth propensity model. And so we’re just gonna continue to that group’s gonna continue to lean into sort of those approaches and drive innovation around sort of new insights and new propensity models and do a lot of collaboration with our customers to find out, you know, what’s meaningful to them and continue to track their ROI from that. On the product on the data product and commercialization side, that’s where we’ve got a lot of people focused on how we’re delivering this data.

We’ve got new infrastructure there. Again, Snowflake is a major partner of ours. We’ve got new React based interfaces that we’re driving a lot of this data into.

We’ve we’ve now integrated they’ve been instrumental in integrating our insights across InvestNet’s wealth tech platform. So it used to only be available to, say, I don’t know, two and a half thousand, three thousand advisers that whose wealth enterprise paid a SaaS subscription to access those insights. Now we have them seamlessly integrated into InvestNet’s trading platform in both on the the unified managed account side as well as on the RIA side with Tamarac. We have it integrate we’ve got it integrate targeted for integration to be in our proposal engine and our financial planning system with MoneyGuide Pro in the first quarter of next year.

So the total addressable market of people that have access to those insights went to a couple of thousand up over to of over a hundred thousand. And so it’s it’s been a massive change in in how they interface with that. And then, you know, building out data products that we can put into something like a Snowflake marketplace, building out new interfaces and and data share capabilities with our customers and how they ingest the data. And again, we’ve got a number of different ways in which we monetize it these days.

And again, the ROI is pretty dramatic for those who leverage it. There’s a lot of companies that still, I think, need to build that culture, that data driven culture.

I think they’re gonna be forced in that sooner rather than later just given the velocity of where, you know, how people is consuming data, how it’s necessary. It’s the oxygen for AI related experiences. So, yeah, we’ve got a we’ve got a team of people dedicated to that specific function as well.

So, again, I I find this I find this rather interesting. It’s something that I’ve been talking about for a while. I certainly talk about it in my book.

Seems to me like what you’ve done is you’ve kind of separated the data supply chain Yeah. Loose loosely saying. Right? The data supply chain, which would include data integration, data engineering, all the ETL stuff Yep. Even acquisition potentially working with all the various sources and get and what I’m hearing you say, you separated that from kind of product and go to market and and and productizing and operationalizing that data.

I this is something I don’t understand why more companies aren’t doing this, particularly companies that get a little bit bigger. I love this approach. Yeah. I specifically love the product management approach.

You used the word data products. Yeah. What what role does product management play here? Do you have folks on your team who were, like, out in the field talking to advisers, do getting all the requirements, defining what these products look like?

What is that?

My yeah. My team rolls up into product at Envestnet. We are part of the product Yeah.

We’re not eve like, we’re not on the engineering side. Like, we it’s interesting. I mean, obviously, there’s there’s some small overlap, but my data analysts, like, they they’re amazing at SQL. Right?

But they’re not. You know? And we’ve gotten some really capable data scientists as well, but they don’t like, we were really intentional about making sure they didn’t get consumed into an engineering function, which is a cost function in an enterprise. And so they look at every when you’re a hammer, everything looks like a nail.

And so everything is like, how what how can we reduce cost there? And and we’re like, no. No. No.

They’re like, we’re on the product side. We drive revenue. We build data products that drive revenue, that that deliver ROI. And because we look at the world through that lens, we can be very strategic about the investments we make into our infrastructure, into new innovation.

We’re out in front of our clients constantly getting their feedback on what they and it’s awesome. Actually, what’s really interesting to me and why I think we’re in this great position is that we’re seeing this maturation in the wealth side where, like, data plumbing is there’s don’t get me wrong, data plumbing is still a big thing. This is a this is this industry is a laggard on so many different levels. And there’s still people that, you know, just want their data in week like, biweekly extracts, etcetera, and they’ve got old systems. But there’s a lot now that are moving to, like, more robust cloud based data fabric systems like Snowflake or Databricks.

And we’ve got this shared infrastructure now, and I’m getting calls from chief data officers and chief technology officers that are like, hey. We’re voracious consumers of your data.

We now are both on Snowflake. Can we start talking about how we can get that not in a weekly, biweekly, monthly basis, on a daily or near real time basis? And then once that’s in place, let’s start collaborating around what things we can do together with our data. And those like, it’s awesome that we have, again, a team that, like, really focused on driving those conversations. And I think we’re gonna see more of that as data maturity absolutely continues to happen in our industry.

I I I love it. This is one of the reasons why I’m talking to you because when we had talked before, all the things that you were saying was just like I’m, like, screaming, yes. Yes. Yes.

Yeah. What you just mentioned was kind of maybe data sharing ecosystems. I mean, yes, I guess you could call it data monetization, and that’s fine. And you’re selling data products, and that’s fine too.

But what you talked about was something slightly, maybe even a little different, which is more of a kind of a sharing, you know, I’ll scratch your back. You scratch mine. We’ll have access to each other’s data, potentially getting kind of to somewhere that looks like that. We may even get into shared governance model, shared in you’ve already talked about shared infrastructure.

Yeah.

One of the things that that just kind of struck me the most was that I I could I could hear a lot of our listeners and viewers saying, okay. Well, you know, I’m in the manufacturing business. Yeah. I’m not in the data business.

Yeah. Right? And Yeah. I’m not I’m not in financial services where, you know, know, stock pickers, datas, trade data, like, I mean, that’s kinda that’s intrinsic to what they do.

You know, my company just makes widgets. Right? Yeah. But in an era of AI, what I’m starting to see more and more, and I think our boards see this, I think our c suite see this, and I’m not entirely sure that we as the practitioners see this enough.

Data can become the differentiator. Day data can become the business driver, especially in industries where you are highly commoditized and where your margins are getting pressed and pressed and pressed.

But Yeah.

Where you have thirty, forty, fifty, maybe even a hundred years of manufacturing experience of how to do things the right way, how to procure things the right way. I mean, these are all like, it’s all data.

And and what I hear you say, Jeremi, is that, like, hey.

Look through data through that lens. Look through data through the lens of a product. Look through the lens of becoming a p and l. Right? And, like, to me, this is this is kind of fundamental.

What do you I mean, if if you’re in manufacturing, I’m so I I did I I did some courses at MIT.

And and one of the things they really highlighted about this this next generation of innovation that we’re in is was sort of a trifecta. They’re like Internet of Things, AI, and blockchain. Block no.

Stayed away from cryptocurrency.

They’re like, blockchain is an We can go there, man.

I’m I’m ready.

Let’s go. I’m totally cool with that too. But it was it was blockchain as an infrastructure layer. And so the again, I think about it through the lens of digital exhaust.

So in a manufacturing plant, what are the things that you can derive data from through sensors, through IoT? And how and and is there value like, looking through the entire value chain, how does that data get captured maybe on an immutable ledger so that, you know, there’s there’s providence and governance of that data that could potentially be used in in different ways, could be tokenized potentially in different ways, all the way then for AI to sit above all of that and ingesting it and and and creating new insights and creating new ways to think about it, like, especially especially in manufacturing or especially in transportation or logistics.

Like, I think there’s a just a wealth of opportunity there with data and AI. Yeah.

Oh, I I’m I’m I’m with you. I mean, I fell down the blockchain rabbit hole about six or seven years ago, and I had I haven’t fully come out. I’m just waiting for everybody else to get to me. Yeah.

Because I’m sitting at the bottom of that rabbit hole saying, hey, y’all. Water’s water’s great. Come on and join me. When you start talking about tokenizing things, like, even if we just wanna talk about data and data as a product.

Yeah. There’s a lot of people out there talking about data contracts. Yep. You know, exposing data through data marketplaces, and that that’s great.

And centralized marketplaces are are great too, and data catalogs are great too. But I view a a future state where there is this decentralized data marketplaces where all of these data products are flowing around that have the data contracts, where you’re actually getting paid for them Yeah. Maybe even in Bitcoin Yep. Maybe maybe in USD too. Yeah. But but, I mean, is it am I am I crazy? Are you been are you thinking about some of the same things in your industry right now?

Oh, completely. Absolutely. Like, that is there is I mean, this is the next renaissance. And and I and I think and and it’s funny because you see ebbs and flows of of the sort of hype around things. Like like, I remember Internet of Things went through its thing, and then blockchain went through its thing. And I think it got a little disintermediated by everyone’s fascination with cryptocurrency.

Three would be another one. Right.

Right. That’s Web three o. I guess Web three o almost encompasses a lot of these things sort of all wrapped up underneath. Right?

Now AI is at the table. Like, that is whatever it’s taking up all the oxygen in the room, but these other things still exist. And they’re so, like, instrumentally they’re so instrumental instrumentally. They’re so instrumental to new modern contemporary data infrastructures that, you know, people people have got to widen the aperture a little bit and and really look at the opportunity here.

It’s not all just AIML. I mean, there’s there’s so many amazing innovations that all sort of tick and tie together, actually. If you’re really strategic about the architecture, there’s amazingly like, amazing new ways to generate new innovation, new new revenue streams.

Yeah. Couldn’t agree more. I mean, you know, historically, one of the things from a kind of a marketplace and monetization perspective is, you know, marketplaces are are are a dime a dozen. Yeah.

And, you know, you could set up a data and try to market it through a a marketplace. But to me and I’ve seen marketplaces come and go like a thousand times. Yeah. Right?

Yeah.

Every major software provider has has had a data marketplace. Everybody’s tried it. But to me, the reason why data marketplaces have not taken off is because of governance. Yeah. People will create a marketplace and say, well, here’s my data, and you just have to go trust it.

Yeah. Right?

Yeah. But, again, getting back to blockchain, and I don’t wanna make it all about blockchain. Yeah. But one of the things that blockchain is particularly good at is governance.

Yep. And you can actually integrate maybe it’s through a data contract that exists as an NFT, but you can actually say, you know, here is the governance that went into this thing. Yeah. Here are the acceptable uses of of this thing.

You can even put controls around, like, time based controls about when it can be used when it can’t be used. So Yeah. I’m I’m I’m I’m with you. Yeah.

One thing we actually didn’t talk about know, we talked about some of the stuff you’ve done with Nextbase Action. We had talked about some of the other ML stuff that you’ve been doing. Yep. And all you you the graph based stuff you’ve been doing around your your your your network graph and your client graph.

Amazing. What about Gen AI? We haven’t talked about Gen AI.

Yeah. So at the beginning of this year, at our large conference, it was in Vegas this year, it was called Elevate, we introduced two new generative AI capabilities into production.

One is called generative business intelligence, and the other is insights AI. So I’ll I’ll deal with GenBI first. So at I we started this call by saying, like, our group historically has been known around our business intelligence data. And and and we are we are shifting now to really focus our time and intentions to be more on the decision intelligence side. One of the ways we need to one of the things we needed to do to help us sort of make that pivot is get out of the business of manufacturing business intelligence, you know, experiences, be they charts, be they grids, be they widgets, be they dashboards.

We we have about a hundred a little over a hundred fit to purpose APIs that we know the wealth industry wants to see from a business intelligence perspective, these benchmarks, its fees, flows, performance, etcetera. But if they wanted to see that data differently and it wasn’t part of our API library, it would be a professional services engagement. It would take my really limited and important resources off the line to build something bespoke that will never be seen again that is actually tech debt for us. We’ll have to maintain this over time.

So as we build new experiences, like, oh, we forgot about those, like, those dashboards that we now need to build only for this single customer. And so we wanted to and and it cost them. It cost them time and money. So it it it added friction to their ability to derive value from the data that they have.

And so we looked at the the situation, and we came up with a solution that said, let’s introduce a natural language query interface where they can ask to see their data in any way they want. And as long as we have the data, then it will send that query securely to OpenAI, use their structured output to dynamically generate an API. So it doesn’t send any data. It just sends the query, creates the API on the fly, sends it back to us, and then we display it real time.

So it gets us out of the business of having to manufacture any new views. Like, again, this can be down at the the grid level. It could be down at just a simple graph. It could be a widget of graphs.

It could be an entire dashboard. It all dynamically gets generated now by generative AI through OpenAI.

It’s integration into our our business intelligence dashboard. So that’s that was generative business intelligence, and that’s now being that that’s literally being, like, pushed through every one of these experiences that exist at InvestNet. If if there’s a if there’s anything BI reporting basis, this is sort of core to that now at at InvestNet.

And then on the insights AI, there, we wanted to deal with what we saw as decision fatigue, generating twenty five million insights a day, giving an adviser a beautiful dashboard to see all of their insights.

They still have it’s a hunt and gather experience. They still have to pick which insights they should go after. And because of that, we wanted to give them a more natural language interface to their insights to be able to sort of talk to their insights, prioritize their insights, get a lot of generative content around those insights that they can’t get from a dashboard. Like, well, help me understand what a UMA tax overlay is or why it’s important, and how should I actually frame that insight?

Oh, this brokerage to manage insight or this single stock concentration risk that exists in my client’s portfolio, what’s the best way to frame that for them and generate that email or generate that that PowerPoint presentation or generate what, like, all of the client prep that goes into that. And so both and there, we leveraged Snowflake Cortex as our primary infrastructure. And just this month, we rolled out an Ingentic framework within Insights dot ai. So before, it was it was a lot of prompting back and forth between the end user and the large language model to get the answer that they were really looking for.

Now we’ve got six or seven different agents that all have very specific roles to play. That allows for really sophisticated multipart questions.

We’re dealing with some really sophisticated concepts. And so these advisers are gonna have sophisticated questions. And we’ve got now a half a dozen agents that now are able to manage that using chain of thought reasoning and having the best outcome possible in the first try instead of having to prompt back and forth to to really get to sort of the, you know, the output that they were looking for.

So All right.

Well well, there you have it. If you doubt for a second that LLMs and natural language queries will not take over legacy BI platforms, believe it or not, Jeremi, I’ve had this I’ve had this Yeah. Debate a number of times on LinkedIn where there’s still there’s still a lot of people who think that BI will continue to play a role. And maybe maybe it’s sitting under the the cover somehow. Maybe it’s helping facilitate some of these these APIs. Or I mean, but I’m I have a hard time seeing how kind of traditional BI lives in a world that you just described.

Yeah. I mean, I I think it does. I just think it changes a little bit. You know what I think needs to happen in the industry is that we we seem to be overly focused on natural language query and chat. And that’s that’s because everyone like, chat has made AI accessible. Like, that’s what now everyone loves to that’s our primary interface.

In in in really highly regulated industries like financial services where the outcome, the quality outcome of the advice from AI matters dramatically, the the open ended nature of Q and A and natural language queries presents a like, we’ll be the first to say, like, this whole year has been a learning curve for us. Like, it has presented a number of different challenges in in not only the quality of the response, the time of the response. And and so where we’re headed, I I think where there’s a lot of green field to take advantage of is, like, embedding AI into the interface itself with the use case already there, like, removing the need for natural language query and having AI anticipate what it is, but have that under the covers and get rid of the open ended nature of some of this.

I mean, there it’s important. Don’t get me wrong. That experience is really important in certain use cases. But in other use cases, you can just have it naturally embedded in the back end, doing all of the things it needs to do from a thinking and acting from an agentic point of view, And it’s gonna deliver just as good, if not a better experience in many ways, shapes, and form.

So Well, but but what you just touched on is what, you know, the kind of thought leaders and, you know, advocates in this space are are saying is the natural next evolution, which is it’s not about individual productivity per se. I mean, it’s not not. I I I challenge people to say there isn’t additional productivity out there. There most certainly is.

Right. I just think that individual line managers aren’t reporting it because they don’t wanna lose their employees, but that’s a separate issue. Yeah. The natural evolution here is from individual productivity to workflow engineering and process in reengineering. And what you just said, I’m paraphrasing you, is now you wanna go from a Google like ask me anything. Right. Right?

And being able to sort any potential query by any investment adviser anywhere on anything, including tell me about the weather today Right.

To to this, hey. I’m doing a portfolio review with client x y z, and I’m looking reallocate something. Right? Like, right.

Is it what you’re saying is is that make that part of that flow Completely. Basic portfolio review. Yep. Or maybe an invest an investment disc or or initial discovery call with a maybe a new client or who knows what but you know your workflows as well as anybody, better than anybody.

But to embed in there, that’s what you’re saying.

Yeah. Yeah. That’s right. And because you have the context. You can you can have AI follow the workflow to that content. You can have AI step in with a really rich explainability.

You can have, especially in our industry, have AI be connected to real time streams to the SEC and FINRA for compliance related and so it can check just real time what normally a human would be involved in to understand, like, okay, is this output also compliant? Like, that all of that can be automated. It does not need a chat window. Like, it it’s it’s, you know, it’s it’s it’s it’s the next phase. It’s going from search and synthesize and generating content for you as an individual to now AI thinking and acting either in a decision making framework all the way to just agentic, like, without any rails, like, letting it do what it wants to do, which I I don’t know whether or not there’s ever gonna be a use case for that type of AI within financial services, AI on rails most or agentic on rails most certainly.

But the autonomous AI, I think it’s gonna take a while for for financial services in particular to feel comfortable with with autonomous AI.

Well, you know, what you just said to me, again, really highlights the requirement, I would say, to focus from a product management perspective and to focus from a customer perspective, a workflow perspective. Understand stand over their shoulders.

Understand how they use their tools day in and day out. Be deeply ingrained with the people that are building those CRM systems or Right. Workflow management systems or what due diligence systems, whatever they are, whatever tools that you’re using. Yeah.

Be deeply ingrained with them and define ways to integrate AI into those processes. And interestingly interestingly, when when you kind of look at some of the companies that are being the most successful here, that’s exactly what they do. Yeah. The layer on top of that is that governance isn’t this separate thing that is over here.

It’s a part of everything. It’s part literally woven into your operations, because it has to be in your world.

Do you agree? That’s a half hundred percent agree, especially in our industry. Yeah. Yeah. Yeah.

Alright. This is this has been amazing, Jeremi.

I Awesome.

Thank you so much for the conversation. Welcome. Really, really enjoyed it.

Really, really eye opening about the possibilities there. And and, you know, why I wanted to talk to you the most is because I wanted to to highlight, hey. Here’s a success story. Here’s a company that’s doing pretty leading edge stuff. It’s pretty cool stuff, and we need more success stories in our our our world. How can people connect with you, Jeremi?

LinkedIn is the best way. This Perfect. Jeremi Carnell at LinkedIn. I’m the one with the big investment band. Now, I have to warn everyone, I’m an identical twin.

Oh, that’s cool.

And his name is Ian. My name is Jeremi. We’re both in wealth management. So you may come across his profile. We look identical. So, yeah, choose the better well, I’m the better looking one. So choose that.

We could probably defer to smarter, maybe. Don’t No.

No. Actually, my brother is definitely the better looking one. Alright. I’m just joking. Alright. Like, he is definitely.

Thank you so much for joining the community today. I really appreciated it, Jeremi. If you’ve listened this long, take a moment to like. Take a moment to subscribe. We’d love to have you join this growing community of CDOs and people who want to become CDOs. So we will check you out on another episode of CDO Matter sometime very soon. Thanks, everybody, and bye for now.

Thanks, everyone. Cheers.