Mark Wickersham, Host 0:00

All right, Ethan, welcome to The Wealth Tech Podcast. I am happy to have you on the show. Ethan, you are the founder and CEO of CFO Family. I think CFO Family is like a valuable and unique resource within the Family Office ecosystem. I'm looking forward to some of the things that we're going to get into today in terms of the evolving tech landscape, building your own tech stack and the considerations there. Owning your own data, data sovereignty, uh, AI, of course. There's lots of good information that we're going to be digging into today that I feel like you're at the forefront of. But if you wouldn't mind giving kind of a brief introduction of yourself and CFO family, that would be a great place to start.

Ethan Bonar 0:46

Yeah, you bet. Um thanks, Mark, and I appreciate you having me on. I I come from the RIA world, believe it or not. Uh I'm not necessarily a tech guy, although that's where I've always had a passion and where I've landed. So my lens that I see things through on the technology side is completely related to advising. So as I watch um over the years the same failure play out across um these families, and I'll call them complex families. Uh you got sophisticated wealth, the infrastructure underneath was not families with uh $500 million in assets across a dozen entities that were still running core reporting out of spreadsheets, uh QuickBooks, uh very basic systems, um or paying for platforms that were way more horsepower than what they needed. Um so my my missing piece was not necessarily, hey, these families need an another dashboard um or specific metrics. It was what what kind of structure can can they build out that's functional for the way that they operate, the way um their entities are structured, the way their family is structured. So looking at these families, I spend a lot of time up front um defining outputs because really that's the the function of any of these systems is to get the outputs that you need. And so we focus very deeply on what are the outputs that you need? Are you more accounting-centric? Are you more performance-centric? Are you somewhere in between? Just this kind of a general um assessment. And then once we determine really what are the functional outputs and who needs these functional outputs, we jump to where does this data live? Um, where does it originate? If it's custodial data, it originates at the custodian, pretty straightforward. If it's bank, pretty straightforward. How does it flow? It's all the miscellaneous pieces in between that usually make up these complex families that don't flow, they don't have structure. Um, that's where the problem lies. Um, and so once we define that out, then we get into things like systems. What systems are functional for this? What systems do you have? Can you leverage things better to get the outputs that you need? So I tend to stay agnostic to systems and software. I don't think there's a single system that um solves everybody's problem. Therefore, my job is to fit those puzzle pieces together uh day in and day out.

Mark Wickersham, Host 3:42

So, yeah, I I mean that there's there's a number of different types of family offices. I think I think an outcome-based makes sense, kind of you know, reverse engineer your way back based on the operational mandates of the family office, whether you know it's a founder's office, could have a simple, simpler entity structure. Some are more like private investment firms versus more comprehensive family offices that may be a little bit more accounting and bill payment centric. Um, so your tech stack is obviously going to be different based on those operational requirements. Before we kind of get into that, I'd love to get a little bit more background on your kind of founder story and and why you started CFO Family.

Ethan Bonar 4:23

Yeah, I mean, you CFO Family came out of um kind of a personal problem that I had. Um, I'm not uh a single family office by any means. Um, but I have complexity in the things that I have to keep track of. And so um after selling my uh my last RIA back in 2016, I went off and did completely other things, ran companies, built companies inside and outside the US, and oddly enough, turned most of those projects right about COVID time. So I had the luxury of sitting at the house um during COVID trying to figure out how to get all my stuff together. And I was thinking, you know, I know all the systems that are out there already. Uh, I know how to run them. Uh I know their price points, and yet I still have all my stuff on spreadsheets and QuickBooks. And I shouldn't, uh there should be something functional for me and priced appropriately for me that can gather all of my information together. And so I started looking out how do we solve this problem? And as I kind of backed into it, I I realized that um, you know, most complex families are trying to get the single family stage. Uh the reality is single family is really just a structure of a business, right? And so in a business setting, if you don't own the systems, if you don't own the data and the information, it's real hard to continue to make effective decisions about what you should be doing in your business. And the same applies to family offices. So I realize that it's core um you have to have the data and you have to have the ownership of the data and the systems in order to be closer to what a single family office is. And if you can scale that down and functionally do it from a price perspective, that makes sense. Now you've got more of the foundational structure of a single family office where you can begin to layer on the expert people in the positions that you want. And again, you've got ownership and control of your data versus renting it from somebody else. And in my MFO days, that's what was happening. Uh, the families were renting the data from me, they were renting the systems from me, they were renting the uh the staff from me. And if I didn't do my job and got fired, the family lost all of that. And so I wanted to flip that around and say, okay, let's have the family have control over it. So when these people are not doing their jobs and they have to terminate one of these positions, they don't lose all of those things. They stay in control of it. So that was the kind of the initial process of um of my thinking was let's create this uh service solution that will define out what systems fit best and at what price is going to be functional. Where's the most value cross to give them data ownership and data sovereignty?

Mark Wickersham, Host 7:38

Yeah, I think that you're a little bit ahead of your time, certainly, I think, around the data ownership and sovereignty, right? I think with AI, it really highlights that. Uh, the need for that and uh how critical one is good clean data is, but also owning your own data. Um let's talk a little bit. There's uh the lack of you said you had a lot of it on Excel, a lot of um single family offices, regardless of the size, still have a lot in kind of a master Excel spreadsheet. Um, and there's a recent UBS survey, the 2026 Global Family Office report pegged it at um almost half of family offices are still doing this. I you know, I know I've I've talked to firms and people at Atabar and Mastro, and they'll be like, Who's your biggest competition? And they'll still say Excel to this day, which is uh amazing. But um why is this? Why is it why are they so dependent upon these kind of Excel spreadsheets? And what are some of the kind of main issues that arise when you don't have a proper reporting infrastructure?

Ethan Bonar 8:49

Yeah, I think it's just born out of that's the natural progression of things. I mean, if you think back 20 years ago, um an Excel spreadsheet was an invaluable thing to have to keep track of all of these things. And they could do the mathematical calculations you wanted it to. So it's just kind of the evolution of this. Um, most of the systems that that were out there, let's just say five or 10 years ago, were geared towards um uh corporations or investment advisors. They weren't geared towards families. So if you were going to be a family office and have something different, you you had to really have a staff to run it. You had to have some understanding of uh of that system and that value, whether that's a performance reporting system or some sort of aggregation system. And so it's just a natural evolution. I I see it all the time too. And I would say that that number is probably appropriate. About half of everybody that I talk to, families at whatever level they are, um uh are QuickBooks and spreadsheets. And and I venture to say it's probably more than that, but I I think that's a fair assessment. Um when you're running spreadsheets, um, you can get what you want out of them, you can get you can put them up what you want into them, but usually there's somebody that runs that spreadsheet. And so now all of a sudden you have some key man risk. What the minute that that person isn't around and you've got a spreadsheet with a hundred tabs, a pivot table on it, and nobody knows how to run it, it's gonna take three months, six months to get up to speed on uh the functionality of that spreadsheet, uh, that's a big risk inside of a family office. So streamlining um streamlining the ingestion of that data, the process of of how you're getting it and where you're putting it, the calculations behind it. There's tremendous value in that to keep things clean and clear, but still most families rely on a spreadsheet. Um I mean, I'm I'm doing a big data extraction um right now with a family that a single family office, they have roughly 300 physical properties um in New York City, and everything they have is tied up in these spreadsheets and documents. And so just being able to pull and extract data out of that information, get it normalized, and get it into a system that they want to use or that's functional for them is invaluable. Can they continue to use the spreadsheets? Sure, but they're using on this as a side note versus as the primary uh functional system that will give them what they need immediately.

Mark Wickersham, Host 11:47

Yeah, I I think there's a there's a place for Excel. I think it's a great last mile type of reporting tool. Um, obviously the flexibility is is tremendous, and every family office has a certain report they want to see a certain way. But then the other factors in terms of um you know, there's probably a material error in that that spreadsheet. I think University of Hawaii, or there's been some studies out there that say, you know, 95% of spreadsheets have some sort of material error in there. Somebody copies over a formula with a hard value. Or this happened with um working with a single family office on a search and selection uh years ago, and they had the master spreadsheet and had an issue with uh this back when Microsoft wasn't compatible with some earlier versions and they had to upgrade. And going through that conversion process, they realized they had some uh performance numbers that they're a little Fungazi, um, had to recast everything. And never mind the audibility or lack of auditability. I I understand it, it's easy to stand up and it's super flexible. But if you're relying upon it as a system versus a last mile reporting tool, I think that's where uh family offices can get kind of sideways. And you know, QuickBooks is just as bad in terms of the retail software, there's no auditability, there's no permissions and controls on it. So um maybe it's a hair better because it is a system versus a spreadsheet, but it still gets you sideways pretty easy. Right. Um let's talk about owning your own data. This is one of the themes that you speak out about fairly regularly uh around data sovereignty. Why should family offices own their own data? What are the what are some of the benefits to owning your own data? And what are some of the the risks that they have when they don't own their own data? And and even just the point like where family offices fully understand whether they own their own data or not.

Ethan Bonar 13:43

Yeah, great question. I mean, this is probably the theme I'm I'm most vocal about. I mean, at its core, it's very simple. Your data should be portable, exportable, and yours. That's it. Um not held hostage inside any vendor's uh proprietary format. I mean that's that's my standard theme across everything. Uh too many families build um years and years of history inside of a platform, and then then they discover when they want to switch providers or bring the reporting in-house, uh, that data doesn't come with them cleanly, or the contract that they have makes leaving leaving um painful experience, expensive. So I see this over and over, and I've experienced this myself. Um, and one of the pieces that I always look at is the out. How do we get out of here? What if we're gonna use a system? What's the out? I want that written right up front. I want it clearly stated and understood. How do we get out of here? What is the data that we're getting? What does it look like? Not just a description on a piece of paper or a contract that says we're gonna provide you something, but I want to see an actual output so that we can understand is this going to be functional for us to get out uh of a system? Um, AI has changed that quite a bit, and I'll dive into that here in a minute. But clearly owning your data means that you can leave that vendor relationship without losing your history. It means your team can query, audit, and manipulate the data instead of only viewing what the UI that the vendor has chooses to show you. Um, and it means your your position to actually use the AI on top of your own data rather than be limited to whatever AI features a specific vendor bolts onto their platform. So in my mind, data ownership isn't just a nice to have, it's the foundation. Uh everything else, including AI adoption, gets built on.

Mark Wickersham, Host 15:52

So data sovereignty is is that kind of key pillar. I I think it's a good point. I I think you know, too, with AI, it's it's really kind of highlighting some technology trends that were or principles that were already there. Like it doesn't define gravity or kind of create new uh laws of physics, right? Like garbage in, garbage out still exists. We've just got a you know, it could just be a force multiplier both ways in a good way and a bad way. So I I think having your own data, owning your own data, I think, with a best of breed tech stack to the um the risk management component of it, being able to switch out the nodes on your your tech stack and and be able to uh change vendors. If you don't have control over your data, to your point, that extraction process can be painful. Um, still the the implementations of these systems is still hard in terms of the backdated uh historical data loads that are going on, but that's getting better with AI. Um I think there was some big news this week with with AI with Anthropic announcing an advisor desktop. We got, you know, many vendors have an MCP link. Certainly software is changing, but how is this changing the landscape for family offices and and how has this kind of ultimately changed the kind of user experience?

Ethan Bonar 17:16

Um, great question. I mean, AI is at every second that we we tick on something new and something more efficient is happening. But I'll jump back to what you said just a minute ago about clean data. Clean data is critically important for all of this. Uh AI is just going to give you bad responses faster with within. That sounds more plausible. Yeah. So so when it said in a nicer way. Yeah. When we talk about um we talk about uh vendors and vendor selection, the idea behind that is that they're going to help you keep your data clean. Um because if you own your data and you've got a data warehouse, your job is to keep it clean. So let's say an example of using an accounting system and a performance reporting system, and you've got custodial feeds coming in, you've got bank feeds coming in. Their job is to clean that up and get it organized and make sure that it's functional. If they've got an MCP into Claude, for example, on both ends, now you can query Claude with functionally clean data. And that's the idea that somebody has to clean it, whether it's it's the vendor, whether it's you, uh irrelevant, it has to be clean for any of this to work. So you family office, the anthropic rolling out claude for financial advisors is not really going to change the family office uh world in in my mind. Um it is gonna hopefully allow some of these advisors to be deeper um in in their value prop versus just kind of selling a suite of services across the board. Um so that's my hope, but I don't think it's gonna change much on the family office side. Um the the the jumping back to the to the clean data, um that's that's the thing that doesn't change, no matter how good any model gets, you have to have clean data that that data fed to it, has to be clean. Um it if there are, for example, too many guardrails, and and I think that's what's gonna happen with this squad for financial advisors right now, too many guardrails um as it relates to what they can and can't say, then there's a limited amount of value that's gonna come out of it. Um again, clean data, back to clean data. And we'll keep jumping clean data is the new oil. We we'll keep jumping back to clean data and back to clean data and back to clean data on just about any of these topics that we're talking about, because that is is has always been and will always be the critically important piece of this. Faster uh faster bad information helps no more, right?

Mark Wickersham, Host 20:13

I do think it is um, yeah, structured, reconciled, clean data is is the foundation which everything's built on. I I think the the Cloud MCP links, I think there's gonna be some really interesting artifacts that can be built with that. We're doing that today with you know, especially around CapCall and cash flow analysis that can get really complex, and everybody has their own slightly different view of it. Um it doesn't, to your point, necessarily change a whole lot, I think, with the family offices, maybe gives them a little bit more control and customization. It does, I think, change the world of software. It just disintermediates uh some providers a little bit, and that the UI um gets changed uh to anthropic. The uh you became and some of these vendors become headless. I think the pricing models for certain vendors um you know need to change if it's a it's a user license. We were just talking before we turned on the record button, right? About Salesforce and where they're going with their pricing model and and how they they they anticipate their future. So I do think it's it's really interesting where where um where this is going, but I I think ultimately it's a good thing for family offices.

Ethan Bonar 21:36

I think it is. I think it's a good thing for family offices. It gives more control, clearly. Um, but it's not without uh its own risks and and potential problems. And that's why um you know, just rushing in to adopting an AI system and going, this is gonna solve all our problems. It's not accurate. Jump back to clean data. Uh Without clean data, it's not going to help them at all. So I think we're going to see a lot of stumbling and a lot of falling that happen.

Mark Wickersham, Host 22:11

Yeah, they could potentially blame uh AI adoption because AI didn't work, but it's not necessarily around AI. It's you don't have that foundation in place. Let's talk about adopting AI. I think there's a couple different uh threads I want to pull on this, how family offices should adopt AI. But let's talk about the vendor due diligence when you're looking to select new new vendors. How should you evaluate their AI capabilities and what questions should you be asking vendors in this area?

Ethan Bonar 22:44

Yeah, this is a good one. Uh definitely you need um to take kind of the same discipline uh approach that you would do with any other fiduciary decision. Real due diligence, not hype-driven urgency, not um uh not any of these things that are gonna run you off your typical um uh rails. So I I kind of go to three questions. First one being, where does our data go and who else can see it or train on it? So whether that is an internal AI in uh in a system that you're you're vetting, or whether that's your own AI that you're you're looking at for your own data warehouse, uh, regardless, that's a question. You need uh a kind of a contractual specific answer, not just marketing language about data residency and retention and weather information to use to improve models. So, number one, where does our data go? Who else can see or train on it? Um number two, what happens when the AI is wrong? Who's accountable? So every vendor that's serious in in the AI space, including anthropics, own and published guardrails for financial services agents, is it's very explicit in that these tools draft work products for qualified professional to review, they don't execute trades, prove onboarding, bind risk. Um so it if a vendor can't clearly describe the human and loop checkpoints, I that would be a disqualifying piece for me. So what happens when the AI is wrong? Who's accountable? Second. And then the third uh being does this integrate with the data infrastructure we already have, or does it require us to change how we store, structure our data to fit the tool? So very important on that one. So the best AI tools that are out there should sit on top of clean owned data. They shouldn't require you to re-architect your entire stack around one vendor's assumptions. So those are the three due diligence questions that are probably the most important when you're uh when you're looking at adding and adopting AI into your into your structure.

Mark Wickersham, Host 25:14

I think those are all good questions to ask their um vendor on that. I I think obviously understanding your vendor's AI product roadmap, if they don't have an AI product roadmap, I think that would be a red flag. Yeah, I think obviously SOC 2 type 2 becomes kind of table stakes for vendors now in terms of data security, understanding privacy, understanding what happens with your data. Are they using your data in in ammonized? I'm not saying it correctly, but you know, uh blind slices to be able to provide analytics at a at a broader slice, and are you comfortable with that? Um, is a is another question to ask. I think uh adopting AI for family offices, what what should firms be thinking about? I think at the you know, obviously beyond the note-taking, which which was been fairly successful. Like I think firms have really struggled with having success with with adopting AI particular use cases. Uh, what are some of the common mistakes that you see firms making and and where do they get sideways?

Ethan Bonar 26:22

Yeah, where where they go wrong usually is messy data and expecting that AI is gonna fix the mess. So I mean AI doesn't clean data, it just uh amplifies whatever's already there, good or bad. Um the second, probably the second mistake is uh treating the adoption as a technology decision instead of a governance decision. So nobody has defined Uber views AI generated output, what's what it's not allowed to touch, or what the escalation path is when it's wrong. Those are governance pieces, those aren't technology pieces. So define those out. You have to to clear that's that'll get you in trouble real quick. And then the third is you know chasing the shiniest feature rather than the normal boring infrastructure questions underneath. Yeah, does it reduce actually reduce work? Does it just create a new place for staff to double check things? Um and then uh maybe a fourth would would be uh tied kind of closely to data ownership. Um is signing on with a vendor without understanding what happens to your data and your history, if you ever want to leave that, and I would go back to what I said before. Firms get excited about the demo and then they skip the exit clause conversation. And so again, backwards. Uh uh that's one of my starting points with with uh indie vendors. How do we get out?

Mark Wickersham, Host 27:59

Yeah, I I think it's could be easy to chase that that shiny object, right? Like you can really get some impressive looking results from it from AI, but taking a look at those kind of painkiller applications of AI, can you alleviate that low value for a smile? I know it's boring, but that infrastructure to be able to free up the human to be able to provide more, they're always gonna have better judgment, right? They're gonna have that context. Um, how can you you know provide that level of automation? And to your point, it's like if they're second-guessing that the results of AI that in in any system, if you don't trust the results of the data, you're spending more time proofing out that data than then you're not getting the the ROI that you expected from such a such a system. And I I think it's really interesting the governance point too, because to your point, like in the and the the the pounding on the fact that you need good clean data. Well, who determines what data is good and clean? It's that that's a governance issue as well, and that governance and data and AI are all interrelated, and there needs to be a structure for all of it for it to have some good results around that. Um let's talk a little bit about the the future of family offices and the family office structure. Obviously, I think AI is as the ability to kind of reduce the barriers to entry for family offices for a while there that was creeping up was 200 million, 250 million. The way it was progressing without AI, you were gonna see like, oh, could you is a family office only at a billion? But now you're seeing you know that that number come down lower again. This kind of lean two-man or two-person uh family office is is viable, the hundred million dollar threshold is viable. I mean, what persons define the definition of a family office can be wide widely open too. So um, but how do you see firms thinking about how they should structure their family office? And what are those three kind of main layers that they need to have regardless of the size?

Ethan Bonar 30:04

Yeah, uh great question. Um the the lower barriers to formation that that you're talking about are are real. I mean, you you you can run a lean, what I'll define as a modular family office uh these days, uh maybe even all the way down to $50 million. And if it's not there, it's getting very, very close. Complexity obviously is a driving factor behind what you need and and where you need it. But I mean, the costs have come down, the cost of operating have come down. The ability for you to put good people, uh, the appropriate deep value people in the right places uh is here if you own your own data for sure. So I I I want to make it clear lower barriers to that formation don't mean lower barriers to doing it well. So just because you can do it doesn't mean you're gonna do it. Doesn't mean you should do it well. Exactly. So if I'm gonna tell, if I'm gonna talk to a family that's standing uh up uh an office, a modular family office, a single family, whatever you want to call it, um you really think in in three layers. And the first one is governance. So the legal and decision-making structure, um, the entities, the trusts, and the state framework, who has the authority over what, how to dispute major decisions, uh uh how how do they get resolved? That is the very first layer that you need. The second layer behind that, in my mind, is the data and reporting infrastructure. So the single source of truth uh for consolidated balance sheet, entity ownership, alt tracking, um, that everybody else depends on. So governance, data and reporting infrastructure, and then the third layer being kind of the advisory and execution team. So now you've got governance, you've got data, uh, ownership and sovereignty and infrastructure. Now you get the experts to come in. So whether that's in-house, outsource, the people that actually manage investments, tags operations, day-to-day. Um a lot of new family office build that layer first because it's the most visible. Then they discovered the second layer was never built uh properly, and then the first layer was assumed rather than documented. And so they typically get the order backwards and and then you spend years retrofitting and cleaning up. Um the the future of where things are going is kind of the next piece of that. Um, I don't know if you've got any thoughts on it, but I think over the next 12 to 18 months, we see AI you know move from a novelty feature to table stakes across the entire industry. Um I mean that that announcement from Anthropic about building uh financial services specific tooling. Um that's the leading edge uh right there. Even though it's not affecting directly family offices in my mind, in any great degree right now, we're I think 12 to 18 months out from the the evolution of that being able to uh uh affect and and power these family office uh offices even more. So uh I I expect a real split to emerge between firms that already have clean owned, uh, well-structured data and can genuinely leverage that new intelligence layer uh versus the firms that just adopted AI on top of a legacy mess. Uh so we're gonna see a big divide in there. Uh that's my expectation over the next 12 to 18 months. Beyond that, um, and I I think from a technology standpoint, the the platforms that treat data ownership and portability as a core product principle rather than you know a competitive threat. Uh those are the ones that are going to win, and families are demanding it, and they're going to demand it even more.

Mark Wickersham, Host 34:28

I think you know, the future, to your point, I often look at the RA industry as kind of a leading indicator of what's going to happen to the family office industry. It's bigger, it's more dynamic, it's more profit-driven, but there are trends that kind of emerge there first and kind of uh come over, not 100% part and parcel, but it certainly is a good industry to keep an eye on, and you see what's going on with anthropic uh and the their visor desktop. I do think um it, you know, the software is gonna radically change in the next 12 to 18 months, and that deep vertical solutions are gonna be the winning way to go versus kind of broader horizontal solutions. So I I think regardless of what industry you're you're seeing that in, that if you're finding a vendor that focuses on your particular problem and gets deep on that versus vendors that focus on you know small businesses, horizontal businesses, whatever that may be, I think it is um it is is how these firms are gonna win. Obviously, the ability to have structured clean data. I get to your point with family offices, I do think there will be a split. I think it's already happening. There'll be the haves and the have nots. The have nots are gonna be the ones that can't modernize, their data is a mess, their data's not structured, the data's not clean. Um, they they rely too much on human capital to get the results uh of that. And that AI is forcing family offices to modernize, and the ones that do modernize are are are gonna be uh you know, really separating themselves from the ones that that are that are not uh in that case. So it is gonna be a uh a dramatic shift in the next 12 to 18 months within the industry and how that they the industry is gonna gotta basically bifurcate itself, um, which I think is gonna be interesting. Uh any comments on that before we kind of wrap it up? Or I got one more question for you.

Ethan Bonar 36:34

I mean, I I I think I think you're you're spot on. I I uh I think there's so many variables and unknowns that we don't know exactly what's gonna happen in the next six months. Again, going back to the fact that if you own your data and you have clean data that you can pivot and adjust, and you're gonna be right on top of everything.

Mark Wickersham, Host 36:57

I think to it's it's a that's a good point. I think like you know, Amazon does an interesting planning process where they take a look at what's not gonna change in the future because it is really hard to predict the future, right? People are always gonna want lower costs, people are always gonna want quicker delivery, you know, those are the things that they look at. Well, I I think you know, having a need for clean data and owning your data, it's a good bet that that's how that's to become even more important in the future. Right. Because that's one of those things that's not gonna change. In fact, it has greater importance. It was great to talk to you, Ethan.

Ethan Bonar 37:31

Same here. I appreciate it, Mark.