The $4 Billion Dollar Man: Leadership Lessons from Jason Wenk

Cold open

Jason Wenk: There’s five to ten years for most of the big players to catch up to us, if they can attract the right people, stay committed to building, go for it pretty hard for quite a while, and sustain the amount of intensity it takes to build and then convert. So there’s, I’d say, a non-zero chance that maybe we are the only ones standing in another decade or two.

Intro

Aaron Klein: Welcome to Beyond the Meeting, the show about what happens after the room clears, where what gets said turns into what gets done.

I’m Aaron Klein. I spent twelve and a half years running a fintech company called Riskalyze. These days, I build AI to amplify the impact of human relationships at a company called Contio.

Our guest today is Jason Wenk, the founder and CEO of Altruist. Now, we recorded this conversation two weeks ago. Days later, Vanguard announced it’s buying his company. The two companies haven’t disclosed a price. Both the Wall Street Journal and Axios put it at four-plus billion dollars.

So what you’re about to hear is Jason’s most recent articulation of his vision for what financial advisors need. It’s consistent with what he wrote down in the pitch deck in 2018: clearing, custody, software, services, and an AI layer he was already calling Altruist Intelligence. Eight years later, he’ll tell you almost all of it got built.

Vanguard says the deal is about helping advisors serve more people without losing the human judgment at the center of good advice. Jason walks us through how they did it.

He started as an engineer, built advisory businesses, grew one of them to $3 billion in assets in about five years, and then walked away to build the layer underneath it. That layer was dominated by trillion-dollar companies. I love that move. It’s a specific kind of crazy. You have a working business. You look down at the plumbing it runs on, and you decide that the plumbing is the actual problem.

Sixty-six hundred advisors are on it now, and the rumored $4 billion says he was right.

Stick around to the end. I’ll be back with my three takeaways.

The 2018 pitch deck

Aaron Klein: Jason, thanks so much for joining me. You started as an engineer, then you built two advisory businesses, and then you left what was a very successful company to build the layer underneath it. That layer was dominated by trillion-dollar companies, and now all of a sudden you’re this up-and-coming AI-powered custodian. So I have to start by asking: what parts of where Altruist is today did you see coming, and what parts were good old-fashioned good luck?

Jason Wenk: Awesome to hang out with you on the podcast. What I’d say is, and it might be hard to believe, and maybe you had this with your companies, maybe not: if I look at the first pitch deck I created, with Nick Beim, who’s become a friend but was an investor at Venrock, in 2018, the first investor presentation, which really sets the tone. This is obviously pre-anything. Pre-product, ideation stage. What we laid out at that time, roughly eight years ago, almost everything we said we were going to do, we’ve effectively done. That’s been largely the roadmap.

So the AI part was in that first roadmap. We called it Altruist Intelligence back then. The different layers: clearing, custody, software, services, asset management. All those things were part of that first vision.

Aaron Klein: Wow. And what year was that again?

Jason Wenk: 2018.

Aaron Klein: Altruist Intelligence. A lot of people forget that machine learning was such an important part of, and really the center of, AI. Generative large language model AI hadn’t come along yet. I remember so clearly in 2018, on the one hand, this super interesting space where you clearly could do some interesting things, and yet I believe I said on stage at a conference once that the only successful deployment of AI in our industry had been via press releases. So it’s kind of cool that Altruist Intelligence was in the first pitch deck. Did the vision you were thinking about there effectively become Hazel?

Jason Wenk: You made an incredibly good point, which is that those first versions were very much machine-learning based, and we actually built and launched into production some of those first versions. But to your point, it was largely looking at large data sets and extracting useful data from big blocks of price and tax lots and performance attribution. Mostly numerical data, which is not what these generative LLMs are particularly good at. So, quite different.

We had this notion we called Altruist Rank, where we’d use this data to surface insights: here’s where you could save in fees for your clients, here’s where you could save in taxes. If we look at Hazel, Hazel is far more productivity-layer work. Some core administrative work, and then a lot of the planning work.

Maybe where Hazel’s very different from that first version: I’m probably on record saying a bunch of times in the early days of Altruist that there are two things we’ll never build. We’ll never build CRM, and we’ll never build financial planning. Not that we’ve built CRM necessarily, but as you know, one of the big foundations of AI is you have to have really well-organized, canonical data. So we built this huge data layer behind the scenes on Hazel to power things like tax and financial planning.

And then we realized we needed to be able to show people the work. If you’re building a financial plan, where did this data come from? The provenance, if you will. We had it all in a database, but then realized we’re going to have to surface this on the front end. So when you’re looking at a client contact record, you can see all of the information about that client. Not just the demographic data, but even some of the psychographic data: in this conversation, this thing was said; in this email, this tone was captured. Those things then get built into the planning module.

That I didn’t see coming. Probably more like 2023, 2024, you started to see these early versions of ChatGPT, and by the time 3.5 came out it was all bets off. We need to go a hundred miles an hour at this.

Why FormulaFolios wasn’t enough

Aaron Klein: You’ve talked a little bit about your prior company, FormulaFolios, as a successful business, but one that still depended on a lot of custodians and a lot of workarounds. Let’s go back to that early founding moment. When did you realize that FormulaFolios was working as a business but was built on a challenging foundation, and you had to think about building that next layer?

Jason Wenk: I was perplexed by custodians, the clearing and recordkeeping infrastructure layer, going back twenty years. It just took me fifteen years to do something about it. The very first RIA I ever formed, it never made sense to me that you had to do billing by doing a daily download to reconcile all of these transactions. It wasn’t very clean data. You know this from your Nitrogen days: you get these messy overnight files, you turn that into something you could actually use for a billing calculation, then create your own billing flat file, upload it to the custodian, and then debit the fees. That just seemed crazy. Why wouldn’t you have a little fee schedule builder, build it inside the custodial application, assign it to the household when you’re onboarding them, and now billing’s just done automatically? That would be the very first-principles way to do it. And I could say that times a hundred other things where you’d scratch your head and go, why is this the case?

I put up with it, like most people, for a long time. I complained but put up with it. Where it got really painful was when we achieved some meaningful scale. FormulaFolios grew pretty quickly. It took roughly five to six years to get to three billion in assets. By that point we were adding over $100 million per month of new client assets. It was a TAMP, I guess, as the legal description of the business. Once you’re opening fifty to a hundred accounts per day, even if you’ve built a bunch of technology on your side, as soon as you get to that custodial layer, you now have 10 to 15% NIGO rates, not-in-good-order rates. And how long it took to open and fund an account and have it ready for trading would be four to six weeks. You start doing simple math on this and you’re like, wow, a thousand open tickets pretty much every day.

And then once you reach a critical mass of accounts and you start getting into trading and rebalancing, which I know you did as well, you’re building this across tens of thousands, hundreds of thousands of accounts, and all the interesting little cash management and seasonal things: contributions for retirement accounts, RMDs, QCDs. I could go on and on. To me it was obvious how painful it was. And once you start doing it at scale, a hundred times the volume of a typical advisor, you realize it’s the biggest pain point in the entire business. A huge friction point.

Then I started going deep. As I peeled back one more layer, I didn’t like the way the economics were set up. I didn’t like some of the general limitations, like the inaccessibility of fractional share trading. The frustration boiled over. Maybe that’s how a lot of businesses get formed: we grow so frustrated we decide to finally tackle it ourselves.

Aaron Klein: Absolutely. I can’t remember who said it, but somebody said a business is just an idea for making somebody else’s life better, and the most successful ones typically start with your own life.

Breaking down the problem

Aaron Klein: I think about the work of building a custodian from the ground up, and it’s so monumental. Infrastructure, tech, regulatory. Then you multiply it by all the types of securities and investments you hold and trade. Talk to me about how you broke that problem down and figured out what to do first. I don’t say this lightly, but it strikes me as almost the same level of problem as these people who start companies to build nuclear reactors or rockets that are going to go to space. Not quite that bad, but it’s getting close.

Jason Wenk: Well, thank you for recognizing the difficulty. Sometimes I hear from customers who’ve built fintech companies, and inevitably you’ll have an occasional customer who says, this shouldn’t take more than a couple of days, why don’t you build this feature? Every now and again you have to educate them on the surface area, the way virtual teams work on different parts of the application. But you hit the nail on the head. I definitely underestimated it when I started. I thought we’d be able to move faster and it would take less money than it did.

The challenges are best illustrated by the old iceberg metaphor. You see the little part at the top, and there’s the huge amount below the water line. Building a business like Altruist is similar. The front end is the simplest: build an advisor desktop that has some UI elements and visualizations. But behind the scenes, you start thinking about building an order management system, you have to build all these different ledgers, an incredibly complex data architecture, a very complex user architecture because of the different types of firms. Even if you build this into a collection of services managed by vertical teams, it’s very intricately connected. Making one change to your product, every single component has connective tissue to that.

Where we started, we were very lucky. I was able to recruit a chief product officer back then. He was our VP of product. His name is Harpreet Ahluwalia, and I took an unconventional approach. Harp had never been in our industry. He didn’t even have a financial advisor. But what he had done is he was the first product leader at a company called ServiceTitan. They went public recently, but he was their first product leader; I think they had almost no revenue when he joined. ServiceTitan is the entire ecosystem that supports home service professionals, HVAC or plumbing or electricians, and they built this very end-to-end system that allowed them to do everything from managing their Yelp and Google profiles to booking, billing, job management, fleet management, inventory management. The whole stack. A very vertically integrated application for small businesses that were underserved. I thought, that’s a very similar problem set. He had no idea what that was, by the way; before that he was a product leader at Salesforce and other SaaS companies. So I thought, let’s bring someone in who can come into this industry with no bias at all and say, how would I tackle this problem?

He strongly advised early on, and I trusted him and we committed to it, that we needed to start essentially at the very bottom of the market in terms of simplicity and size of customers. There’s too much to build, so we have to find: is there some subset of people who would be okay with only five or six account types, who only need to trade stocks, ETFs, and mutual funds, who are one- or two-person solopreneurs where we don’t do a lot of user management? We descoped this incredibly complex surface area. What we do today would have been such a waterfall project. It would have taken years and years to get even a V1 to market. We needed to get something to market to prove there’s interest and demand, and that we can economically support even a smaller advisor.

So that’s where we started. In hindsight it seems like such an obvious and good choice. Others take a different approach. They build for enterprise from day one, and that’s okay too. That sometimes works really well. But ours was: if you built for enterprise from day one, it would have taken us eight or nine years and a billion dollars. Even with that strategy, it was still a pretty mammoth V1 just to do the baseline work.

Going self-clearing

Aaron Klein: If I recall correctly, you started by building on top of other rails, and then you got to the point where you had to make the self-clearing transition. You’ve talked about the fact that that was painful, that it exposed some gaps, that it stressed out advisors and clients and put a real burden on your support team. Take us into that moment when you realized you needed to do that, and how it went down.

Jason Wenk: The intention was always to be self-clearing. The thing about any startup, as you know better than most, is that investors don’t want to finance five years of your roadmap in advance. They’d like you to show some milestones along the way, some product-market fit, prove out a little of your unit economics. So the faster path to market was to use a clearing partner for the brokerage part of the business. We built a lot of our own software, but then sat on someone else’s clearing platform.

Interestingly, the pain wasn’t in the conversion. The pain was the other clearing firm. In our industry, if you’re building a simple consumer app, you can borrow the rails. There are two or three really great companies for building fintech products, especially for consumers with more simplistic needs. You want to do a trading app that has stocks and ETFs and crypto? No problem. You want to start building for advisors, and the list of requirements is very long. What we learned is that very few advisors, even if 95% of their clients only use four or five account types and primarily stocks, ETFs, mutual funds, if even 5% of their clients are using margin accounts or trading options or trading individual fixed income or alternative assets, they’re not giving you any of their dollars until you can meet all of their needs.

So the challenge with using other people’s clearing platforms is that they are not built for wealth managers. They’re built for consumer trading applications, and trying to jam that into advisors created a lot of complexity. We built a bunch of human workarounds. It’s like the Mechanical Turk: people thought it might have been automated, and it was really an army of people working eighty hours a week behind the scenes.

Aaron Klein: It was a beautiful interface, and when you clicked save, an army of people leapt into motion.

Jason Wenk: Totally, for certain things. Again, a lot of what we did was automated. But for us it was, as soon as we can possibly finance building self-clearing, let’s do it. And it wasn’t long. We were only on a third-party clearing platform for about a year. Very grateful for them. Great company, Apex Clearing, who we worked with predominantly. Bill Capuzzi is probably one of the nicest guys in the whole industry.

Aaron Klein: Oh, for sure.

Jason Wenk: Nothing but respect for what they do. But it did not work for what we were trying to do. I don’t think we were the first people who’d tried to make it work for wealth management. That’s not the core competency, at least it wasn’t; I think they’ve made a bunch of progress. This is five, six years ago.

So that was the journey. Building self-clearing is non-trivial. It’s a whole layer deeper into the infrastructure, and precision is very high. You can’t mess things up like cost basis accounting. You don’t want to trade the wrong tax lot or the wrong account. So there’s a lot of precision required in deciding to launch that business, and a lot of capital required too. We’d fortunately had enough success. We’d built enough demand, there was a real fervor around the business that got advisors excited, which allowed us to prove to investors that this would be worth the additional capital. The rest is sort of history. It turned out to be very wise to make that investment, but it was definitely not easy.

Why a custodian builds AI

Aaron Klein: So Altruist is rolling along. The generative AI revolution arrives. LLMs start coming on the scene. You decide to build Hazel. Talk about what led you to take on that part of the challenge. Was it opportunity, or is that what survival looks like for a custodian in 2026?

Jason Wenk: I don’t know if I’d say survival, because if that’s the case, we’ll be the only one who survives.

Aaron Klein: I was looking for that bold prediction, Jason.

Jason Wenk: There’s a non-zero chance. If you’re a mobile telecom company right now, and you’re staring at Starlink getting ready to launch Gen 3 satellites and undercut your pricing and improve your latency and speed, I’m not real bullish on those folks. The odds of them being able to build a rocket company, build a satellite company, deploy said satellites into low orbit, and do it all before their business is completely wiped out aren’t very good. They’re going to be focused on very different things.

What we do isn’t quite that level of complexity, but there’s five to ten years for most of the big players to catch up to us, if they can attract the right people, stay committed to building, go for it pretty hard for quite a while, and sustain the amount of intensity it takes to build and then convert. So there’s a non-zero chance that maybe we are the only one standing in another decade or two.

But separate from that, on the AI front, my logic was very simple. When we built the company, you’re building a culture, and part of the culture here was this flywheel to make it very simple to our team what’s going to matter. Our flywheel was very simple. First, let’s drive 10x innovation for advisors. That could be 90% cost savings or a 10x improvement in some way: speed, cost, simplicity. Second, let’s drive 150 basis points better outcomes for end investors, their clients. We assumed if we do those two things, it’ll lead to the third, which is that we’ll have the highest net promoter score, the happiest, most loyal customers in the industry. That allows us to capture more assets, which gives us the revenue to invest in further innovation, to drive more return for the advisors and the clients, satisfaction, and assets.

So when we think about that, operating primarily at the custody and software layer, think about a thirty-year client journey. The custodian doesn’t even get into the mix until the client decides to hire the advisor. There’s all that work that happens beforehand. And then after they become a client, if you’re using Altruist, the onboarding process is two or three minutes, and you can set up almost everything automated: cash management, tax management, securities exclusions, model portfolios, billing, reporting. So in essence, we can’t give them any more time back if we only live in the custody layer. Then, after you open accounts, you’ve got twenty or thirty years of serving people. That’s a lot of time you might spend being responsive to emails, to meetings, following up. A lot of the work you’re doing at Contio is relevant to this.

We looked at it and said, if we’re trying to create 10x returns for these advisors, we have to find ways to help them with all of that other stuff. I’ve always felt a custodian is uniquely suited there. To power AI models you need a lot of things, but two in particular. One is really clean, well-organized data. Custodial data is well organized and clean, and it’s very much the source of truth. If the custodial records say your date of birth and it differs from the CRM, the only one that matters is the custodial record. Same with your beneficiaries, your account titling, your positions. And then you need context. We have 6,600 financial advisors that use our product every day, and tens of thousands of their employees. If we start getting involved in what they do outside the custodial layer, we can build context very quickly, combine that with the data, with an AI API-powered custodial platform, and start making agentic connections between the front end and the back end.

So the vision wasn’t, oh, it’s the only way to survive. It was: how could we provide way more value than we can if we don’t do this? And we decided to go for it.

Aaron Klein: That’s something else. I think it’s interesting how you connected that vision into the core flywheel, and it wasn’t just opportunistic. To me, that’s when it starts to get really interesting, when you can drive the core flywheel of a business.

The day the stocks sold off

Aaron Klein: And then you have these unexpected moments where the flywheel starts to spin out of control. I think about when Hazel’s tax capability rolled out. As far as I understand the story, but you’ve got to tell it to us: you’re on vacation in Mexico, and all of a sudden wealth management stocks are crashing because Wall Street is nervous that AI might disrupt this industry. That must have been a surreal day.

Jason Wenk: I wish I could say it was seven days into the best vacation. It was literally one day into a weekend getaway. I take very little time off from work, so I got one day off, and the next morning I woke up and my phone was on fire. I live on the West Coast, so by 5:30, 6 a.m., whenever I woke up, it was already pre-market, and it was going to be a bloodbath. And we were being cited in those reports, because of the launch of Hazel.

A fair way to put it is it was like the canary in the coal mine. AI broadly will impact wealth management stocks. We happened to be an innovation that was cited, and all we did was a standard press release. Some people ask me, how did you get that story into the Bloomberg terminals? We did the standard press release, but we get followed by a lot more tier-one media than most startups, just because of what we do. It started making its way around the Bloomberg wires, and traders started paying attention: this is the tip of the spear right here. They’re going to start building knowledge work that can replace humans, and there’s a whole long tail of things AI can do, doesn’t do yet, but likely will do fairly soon, and it will be pretty disruptive to these firms. Everything from automating cash management to automating wealth management.

So my vacation ended early. I retreated to my hotel room, spent the next thirty-six hours there, and then hopped on a plane and went home, back to work. A very short-lived vacation.

Aaron Klein: Not sure that was the most popular routing of that vacation for the rest of your family.

Jason Wenk: They enjoyed their time. I heard all about the fun stuff they did. I didn’t have to do any of it.

Aaron Klein: Oh, man. Hopefully you get another one to replace it. What an incredible story so far, and I know it’s nowhere close to over. You’ve got a lot more to conquer, and it’s exciting to watch.

The one value

Aaron Klein: There’s a standard closing question I like to ask. Zoom up to the big picture and think about the people you lead and the clients you serve. We call this podcast Beyond the Meeting. When the meetings are all over, what is the one value that you most want your leadership to be remembered for?

Jason Wenk: I’ll give you an answer, but first I just want to say, I remember when you first started Contio and you were talking about fixing meetings, and I was like, this sounds amazing. I can’t wait to see what you’re building. Because if I’m being really honest with myself, and if businesses listen closely to their employees, you can gain a lot of insights. For us, one of those is that we move incredibly fast, we give people a lot of empowerment to act with autonomy, and we don’t do a great job of communicating. Whether it’s a meeting like today, where we have an all-hands and I’ll be addressing the whole company, or a small executive meeting, or a one-on-one, there are all sorts of things you discuss, and then there are things you want to broadly communicate, sear into the culture of your team, and ensure that you execute against what was just talked about. When we do our employee surveys, it’s the number one thing people tell us we have to get better at: the communication after the meeting. The follow-up. The follow-through.

So in terms of what to be remembered for, I don’t think a ton about that. We’re so heads-down focused on the work. But an area I want our leadership to get better at is: maybe we become the world’s greatest organization at communicating the shared vision and mission and tactical work that comes out of our meetings, versus the meetings themselves.

Aaron Klein: I love it. Jason Wenk, thank you so much for joining us.

Jason Wenk: My pleasure, man. Thank you so much for having me on.

Aaron’s three takeaways

Aaron Klein: I hope you liked that one. Jason is one of the sharpest infrastructure thinkers in this industry, and we recorded this two weeks ago, before the announcement. Experience has taught me that deals are not done until they’re signed. So you can take what he says about the next ten years as what he actually thinks. He had to say it in a way that worked if the deal didn’t happen.

Three takeaways from me.

First, and it’s the reason I wanted him on the show. Altruist is fast, well funded, it gives people enormous autonomy, and when they survey their own employees, the number one thing that comes back is they need to get better at the communication after the meeting. The follow-up. The follow-through. And that’s measured. That’s not just his hunch. A company that good, with a problem that basic. A big part of that is that those things probably were communicated in the meeting. People can’t remember the specifics from that moment. And this is so common: they’re too embarrassed to go back and ask again, because it makes them look bad. That’s a big part of how I’ve been thinking about Contio’s compounding workspace intelligence, because you never look bad asking AI to bring back the ambient details from those conversations. So steal this one. Run his survey question on your own team. Ask them the one thing you’ve got to get better at, and count how many of them say what his people said.

Second. I love how he talked about cutting the scope until the thing can actually ship. Their version one was five or six account types. Stocks, ETFs, mutual funds. Solopreneurs only. And his counterfactual is the part that really landed for me. He said if we’d built for enterprise from day one, it would have taken eight or nine years and half a billion dollars. Eight years and half a billion, or ship something small and real that you can build on. That’s not just a startup lesson. That’s every project on your to-do list right now.

And third. Once you’ve automated the transaction, the only thing left is the relationship. Here’s his frame. In a thirty-year client journey, the custodian doesn’t even enter the picture until the client has already hired the advisor. Onboarding now takes two or three minutes, and then there are twenty or thirty years of actually serving somebody. He said it plainly: once you’ve made the transaction instant, you can’t give anyone more time back by making it faster still. The remaining time is all emails and meetings and following up. That’s where the leverage moved. That’s where it’s staying.

One more thing, and this is the part I keep thinking about. Partway through, Jason says the big players in his industry are five to ten years behind him. He wasn’t bragging. He listed the conditions. They’d have to attract the right people, stay committed, sustain that intensity long enough to build it, and then convert everybody onto it. And this week, one of them decided to buy that capability instead.

Salim Ramji, Vanguard’s CEO, said the point of the deal is preserving the human judgment and the relationships at the center of good advice. That’s the argument of this entire show, coming out of a firm that manages trillions of dollars. My read, and it’s mine, not theirs, is that what changed hands this week wasn’t a custody platform. It was that argument, built out far enough that somebody could actually buy it. I just didn’t expect to have a $4 billion proof point before the podcast even dropped.

The second half of the work starts now. We’ll see you next time.