Shaun Tucker: Hello and welcome to today's PracticeLab webinar. Thank you very much for joining us. I'm your host and moderator, Shaun Tucker, and I'm the director of practice management here at Capital Group. I'm really excited about today's topic where we're gonna discuss how high-performing teams use AI as a growth driver.

It's really a unique topic for us because usually on most of our webinars, everyone's more or less on the same page and has experience with what we're discussing to a degree. As an example, back in February, we talked about referrals. Every advisor knows that referrals are important and all of you are trying to get more of them to some degree. But what about AI? If you ask 10 different people, you'll get 11 or more different perspectives. If you think about it, the companies we work for have different tools, different approaches, different commitments to AI.

Some of you are deeply skeptical about it, some are just getting started. Others are all in and have been from the beginning. They're members of Sam Altman's fan club, right? Um, and then the rest of us are somewhere in between. But no matter where you are on this really wide spectrum, today we're gonna provide you with practical guidance that all of you can implement right away so that you can take a positive step in utilizing this new technology in your business. And, and we really think it can be transformative for you if you make a commitment to it.

Okay. To the good stuff.

Let me introduce our two expert panelists. And first I want to just, uh, give a shout-out to our special guest dialing in from back east, from Reston, Virginia, Michael Kitces. Uh, Michael is someone who needs no introduction for this group. He's become a key figure in our industry. Uh, Michael has over two decades of industry experience, is the Chief Financial Planning Nerd of Kitces.com. It's such an awesome title, befitting of the very smart insights that Michael and his team provide to their now over 65,000 subscribers and counting. Uh, Michael, hey, thank you so much for being a part of today. Thank you for the thought leadership you provide to our industry and thanks for being a great partner to Capital Group.

Michael Kitces: Uh, my pleasure. Appreciate the opportunity, Shaun.

Shaun Tucker: Yeah. Look, looking forward to this, my friend. Um, and joining me in the studio is my colleague Brock Sutton. Brock, welcome. Brock is a technologist, an AI specialist here at Capital Group. His formal title is head of emerging client capabilities, which just means that Brock is always working with clients, helping them understand and take advantage of new technologies, identifying better ways to implement them in their practices, especially pertaining to AI. So Brock, welcome. Great to have you here.

Brock Sutton: Thank you. Excited to be here.

Shaun Tucker: All right. Let's set the table. Here's the plan for today. First we're gonna talk about the why. Okay? Why should you care about AI and why are other advisors using it in their practices and how is it impacting them? Second, we're gonna talk about the what. What do you need to know about AI, what's important? And we're gonna dive into five distinctive layers of AI that are important for you to understand as you implement AI. And then we're gonna spend most of our time in the third part, the how. How do you get the most out of the AI tools that you have?

And Brock and Michael are gonna walk you through this really cool demo that utilizes some use cases that are highly relevant that you can go back and use with your teams and apply to your business. Okay, so I've set things up now. Let's start with the why, as they say. And Michael, I'm gonna start with you. And if we look at this next slide, um, it shows some research on adoption from advisors, but why is AI important? What's your take on AI with respect to advisors adopting it, and how are they benefiting from it today?

Michael Kitces: So in, in the simplest sense, I, I, I really just think of AI as another, another technology tool in the toolbox that we get to use to be efficient as, as firms, right? We have, if you wanna get in the grand scheme, like, we got calculators, then we got computers, then we got smartphones, uh, and now we have this, this new tool in the toolbox of AI that does certain things in a very helpful way and is not so great in other things. Uh, and, and has a, I, I find a, a, a very deep potential of what can be done and a learning curve associated with it. It reminds me in many ways of, uh, you just, when we all started getting computers on our desks 25, 30 years ago and we got spreadsheets. It's like, well, you could do all these amazing things with spreadsheets. But even now today, some people do very little with spreadsheets, some people are like spreadsheet wizard gurus that can figure out all sorts of amazing things I didn't even know how to do in spreadsheets, and most of us are somewhere in between.

It's a helpful tool to pull up and we've got certain calculations or models or analytical things we want to do, but then we don't, we don't really use it beyond that. It does well the things that it does well. And I think AI, we're still figuring out in some ways what the good things are to do with it. What are we actually going to, to do to leverage it in practice? And so we're seeing it crop up in a couple of different ways. I get things like, um, AI-specific applications like meeting notes tools. We had a lot of Jump users that highlighted. Okay, uh, I bring the tool into my meeting, it captures the transcript, it gives me a summary that I can either commemorate for compliance note purposes or maybe draft a follow-up email out to the client and queue up some follow-up tasks.

We're gonna see, I think a lot more layering of AI into existing applications and tools that we use. That's everything from Microsoft embedding Copilot, uh, Morningstar has Mo, tools that will simply say, "Hey, AI makes certain things in our applications easier." You won't buy it because it has AI. You'll buy it to do the investment research or analytics or financial planning analysis or performance reporting, whatever it is that you do. But there may be elements of how it interfaces or generates its outputs that happen to technically be generated by AI. And then as we saw in the, in the responses, a lot of us are just using AI tools directly, right? Chappie, GPT and Claude and Perplexity and the like to do more, I'll just call like o-openended questions and work that we may interface. But ultimately as I view it, I, I think it's both.

Um, candidly, uh, I think a little overhyped in, in the amount of predictions that are supposed to completely, like, change the world of what we do as financial advisors. I don't think the core of what we do really changes as much as everyone suggests, but the ways that technology becomes useful are sometimes really even hard to imagine, you know. When, when the internet showed up 20 years ago, not a lot of people were saying, "Well the future of this is I can manage my clients from a laptop and a smartphone on the beach and meet with them on the beach and they won't even know I'm at the beach 'cause there's a virtual background, so you can't tell I'm at the beach." All sorts of amazing unintended consequences that come from powerful technology. So sometimes it's gonna play out in ways that we don't expect, but I think that the crux of it is it's simply another productivity tool in the toolbox where sometimes I can pull out tech and it helps me I think get things done a little bit faster than I could have otherwise.

Shaun Tucker: Yeah. Uh, let me play that back. Appreciate that historical perspective just on the advent of new technologies. But what I heard here is this doesn't change the core of what we do as an advisor, there are some limitations, it's a bit overhyped. It is after coming up a learning curve, an efficiency tool. It-

Michael Kitces: Yes.

Shaun Tucker: ... can meaningfully change some things. It's gonna be good at what it's good at. Uh-

Michael Kitces: Yes.

Shaun Tucker: ... so hopefully that was a, a good paraphrasing, a summary of, of what you just said.

Michael Kitces: Yes, absolutely.

Shaun Tucker: Brock, I'm gonna bring you into the conversation and I would ask you just for fun, play devil's advocate here, okay?

Michael Kitces: (laughs)

Shaun Tucker: Um, we have a lot of people in the audience, you saw from the spot poll, that are on the bleeding edge maybe of adoption. Talk to us about what the high adopters are doing and if you disagree with what Michael said about the limitations and it's not gonna change what we, what we do on a daily basis, please weigh in.

Brock Sutton: Well, I, I can be a great devil's advocate so I can, I can certainly play that role. But, uh-

Michael Kitces: (laughs)

Brock Sutton: ... before we jump into that, what I think is helpful is to look at how advisors are leveraging this technology. And we've done two research studies to date, um, on this. We did one in the fall of 2024, and then one just in February 2025 looking at adoption. And one way to look at the evolution of AI is to look at these five levels here, and so currently we're at level three. And so we're gonna go through each one of these, but what we measured was advisor adoption of each one of these levels. So level one, conversational AI, this is what you experienced with the original ChatGPT that would've been, uh, what is it? November 2022. And so the analogy here is this is a PhD intern. They've got a PhD across every discipline in human history.

Michael Kitces: (laughs)

Brock Sutton: You ask them a question and they respond immediately off the top of your head, right? They sound, they sound really intelligent, they use a lot of big words, but sometimes they get questions wrong. Okay?

Michael Kitces: Yup.

Brock Sutton: And what-

Michael Kitces: It's, it's-

Brock Sutton: ... we see is-

Michael Kitces: It's a brilliant, it's a brilliant intern who has, well, studied every PhD and takes things ridiculously literally sometimes. (laughs)

Shaun Tucker: (laughs)

Brock Sutton: Yes. Yeah, and they're, they're-

Michael Kitces: And then you still have to shape 'em, give 'em feedback.

Brock Sutton: Exactly. They're new to your business.

Michael Kitces: (laughs)

Brock Sutton: And what we see is 79% of advisors are using that type of AI, at least somewhere within their business.

Michael Kitces: Yup.

Brock Sutton: Level two is reasoners. So this is what we got with OpenAI and their o-series models, which was in around September of 2024. So still relatively recently. And this is that same PhD intern, but now you give them a problem and as opposed to responding immediately off the top of their head, they actually take that problem, they think through it for a week, they break it down step by step and then they come back to you with an answer. And so this is really good for things like math, science, coding. Think of things with a binary right or wrong answer. And we actually see about half, so 48% of advisors leveraging this somewhere within their business. And then the bleeding edge is agents.

And so that I would say the, the really good agent products, which we'll get to one of them, one of my favorite in the demo, just came about in early 2025. And that's that PhD intern but now they're actually taking action without you there.

Michael Kitces: Mm-hmm.

Brock Sutton: So they're actually doing something, executing on your behalf and only coming back to you when they've done that. And what we see is actually 12% of advisors leveraging that, and that just came out this year. So, uh, I would say this, this technology overall is being adopted pretty aggressively by the advisor population.

Shaun Tucker: Yeah. And we're seeing the PhD intern get smarter and smarter and, and, and more nuanced

Michael Kitces: Yup.

Shaun Tucker: ... and, and, and more helpful.

Brock Sutton: I'm struggling to keep up with the PhD intern a little bit.

Shaun Tucker: There you go.

Brock Sutton: Yeah.

Shaun Tucker: Yeah, there you go. Um, alright, so Brock, I'm gonna stay with you. And you think about everything that you just said, the direction of travel. Agentic took place this year, it's moving at light speed. Um, I'm just gonna ask something that I know is an elephant in the room that's on everyone's mind. Let's just get to it. How is this gonna change the way advisors do business? Or more bluntly, is AI going to change a range of jobs in our industry? And more specifically, is it going to take my job? React to that.

Brock Sutton: I, I love this question. We get this question every time we talk about this topic, and I think we have to be realistic. So maybe Michael, I, I don't know if I disagree with you or not here, we can get into it. But I, I think it's not going to take our jobs, but I think you have to realize it will probably take a portion of your, of your job. So if you think of jobs as a collection of tasks that you do, some portion of that may be commoditized by AI. And so one frame that we can use, as Michael was talking about, it is just the next evolution of technology. How has technology historically impacted industries? And it has, has a similar dynamic, a similar impact. It commoditizes something, some job, and the beneficiaries then are the complements of that. So one frame for an advisor or for any, anyone in this room is to think about what is this technology good at, and let's make sure that my skill set is-

Brock Sutton: Is to think about what is this technology good at and let's make sure that my skill set is complimentary to that. So we've got five different ideas here where let's say we fast-forward five, 10 years that could play out that are kind of interesting thought experient-experiments. The first one is this idea of the idea economy. So if things like analysis and execution become free or become commoditized, where you point this technology matters more than ever.

Shaun Tucker: Mm-hmm.

Brock Sutton: Novel ideas, interesting ideas, agency, that becomes really important, so maybe, maybe that is one pressure point that becomes more important in the future. The other one would be private data and private information. So in a scenario where you are able to analyze anything and anyone is able to analyze anything, the data that you have that your competitors don't have becomes even more valuable.

Shaun Tucker: Mm-hmm.

Brock Sutton: So what is your client data strategy? What is your private data strategy and how do you use that? That could be more important. The third one here is this idea of task frequency. So, you know, Shaun, Michael, think of all the things that you wish you could do, but they're dated by time, money, resources, human capital, et cetera. What happens when you can do those on a consistent basis?

Shaun Tucker: Mm-hmm.

Brock Sutton: So if a client's looking through, uh, maybe they're gonna buy a boat, they can see real time how their financial plan and their retirement income changes as they scroll through those boats. The fourth one here is around client interactions. So today the internet is made for people. What happens in the future when the internet is made for AI? So when we all have-

Shaun Tucker: (Laughing).

Brock Sutton: AI assistants and we're telling our AI assistants-

Shaun Tucker: Yeah.

Brock Sutton: To say, "Hey, you go take action for me, you get this information for me." How does that change your strategy? How do you get through to new clients? How do you get through to existing clients? What, what does a website for an agent look like, right? Like, some really interesting questions there. And then the last one, and the maybe most important one just to pay attention to is this idea around business model. So today there's a lot of things that are high cost kind of flat payment, right? What happens when that trs-transition to a potential pay per outcome? And so the, the analogy here is what happened within advertising and the internet.So prior to the internet, um, I would go and I would get a billboard on Sunset, or I would get a-

Shaun Tucker: Yeah.

Brock Sutton: Advertisement in the New York Times. It was a fixed fee. I had no idea how much revenue I was gonna drive, I didn't know if it was gonna be profitable, right? I hoped it was, but I wasn't certain. Then comes along the internet and then comes along, Meta, comes along Google, and they change it to a pay per outcome. I could say, "Hey, I am selling a $10 widget. I'm willing to spend up to $5 for every customer that you can find me that is going to actually purchase this $10 widget." So pay per outcome. So what potentially happens if work starts to transition, more types of work, to this pay per outcome dynamic? So again, some really interesting things to start to think about.

Shaun Tucker: Yeah, I, I mean there's a, there's a ton there. What I wrote down was we need to evolve, and I like that idea of your skillset needs to be complementary to the strength of the technology wherever it is. Um, I heard a lot about the power and importance of ideation, and that being still something that, that advisors can lean into and differentiate around. Uh, data and proprietary data, that's a big deal, time being freed up, efficiency, there's, that's massive. Client interactions chachanging and shifting, and of course, bus-business model shifting towards outcomes, that's all really great stuff. Um, Michael, jump in here. I know you're super passionate about this. Um, do you have a different take on the impact AI is gonna have on advisors and the way they do business?

Michael Kitces: I, I actually start in a very similar place to Brock's framing, that, uh, you know, if, if you think of the jobs that we do as a long list of tasks that we do, technology just has a fairly standard process of chipping away at the sort of the lowest common denominator, simplest, easiest, most repeatable tasks, because those are the ones that are most straightforward to, uh, uh, to delegate to technology to eventually automate in full. So I, I think about this in the context, so I, I, Brock mentioned some historical examples, I like to look from this lens as well. So, 20 years ago I was part of a, uh, advisory firm in the early 2000s. There were three lead advisors, they did about $1.3 million of production of GDC, which back then was a, that was a pretty good-sized shop, and there were about eight of us in support positions helping them out.

Three, like paraplanner associates, uh, two client service admins, someone handling all the bookkeeping and billing and, and, and, uh, payments reconciliation. Uh, we had a wonderful woman named Betty, and Betty's primary job was to take all the mail that came in every day, open every envelope, pull every client statement, put it in the physical paper folder of the client, check for any checks, 'cause God forbid you can't sit on any of those overnight, and then maybe greet the occasional clients as they came in for an in-person meeting. If I were to look at a firm today that had three advisors doing $1.3 million of, of gross revenue, that firm probably has at most two support staff instead of eight. Right, Betty's job doesn't exist anymore. Many of us meet with clients virtually, even if they come in, we don't need a full-timer to greet people, and there are no paper statements coming in the mail that we put into physical file folders, nor are there any physical file folders.

All of that's been transformed, but advisory firms today have remarkably similar profit margins to what they did 20 years ago. We've been running 25% to 30% profit margins in this business for decades. All that technology change, we actually reduced the number of people but we increased the amount of money we spend on technology, three advisors buying portfolio performance reporting software today probably passed more than what Betty received in salary 20 years ago, and along the way, our value proposition upgraded. Right, the true, the dirty secret was, yes, we manage all of our clients’ portfolios, but we were pulling out like Morningstar Principia Pro and analyzing each client's portfolio 30 minutes before the meeting, just trying to get up to speed on what was going on with them and what they're doing, portfolios did not get reviewed much in between meetings. Now we have continuous monitoring, continuous rebalancing, and tax overlays, and all these digital things we do on top. Financial planning software was still very spreadsheet based, now we have all this amazing collaborative tools that we can put in front of clients. And so when I look at that evolution collectively, first of all, like it's the same three advisors, almost all the technology change didn't actually change the advisor's job, it mostly changed the back office jobs.

Shaun Tucker: Mm-hmm.

Michael Kitces: We don't need as many people to support an advisor to get the things done, and we can do more for the clients. And along the way, we do more for the clients, there's no question to me today that we do more holistically for clients than we did 20 years ago because of the ongoing march of technology. Yet through it all, the core value proposition for clients is really not all that different. I'm gonna provide you comprehensive financial planning, I'm gonna oversee your portfolio, I'm gonna make sure that your investments are aligned to your long-term goals and that you can achieve them, I'm gonna give you advice about retirement and tax and estate along the way, I'm gonna make sure your insurance and risk management is in a good place, and we're gonna take you along to achieve your goals and be your guide along the way.

And so there is, again, to me, this dovetailing that, yes, I fully agree with Brock, like the, the technology always inexorably marches forward and takes out the bottom layer, uh, the simplest things that we do and automates them or commoditize them and pulls 'em outta the way, that's what gives us the time to do more for our clients now than we did 20 years ago, but it doesn't necessarily change the core value proposition of what we do as financial advisors and why we get hired. It lets us, over time, incrementally do more with a little bit less support staff to achieve the outcomes. And that's still like very much the path that we're on as we get the next wave of AI as we got with robo-tools, as we got with smartphones, as we got with internet, as we got with the personal computers. You know, we were, we were supposed to be destroyed in like four different waves of technology innovation right now, and we're still here, we just use all this amazing technology to do more for our clients.

Shaun Tucker: Yeah, very well said. Um, I just wrote down, uh, marginal contribution to the bottom line, albeit that it's changing some of the organizational structures, bringing streamlining, um, you know, changing the back office, if you will. Uh, it's interesting your comment about while it is absolutely helping, uh, concurrently you saw an increase in complexity and sophistication of our business, and so maybe that's the reason you didn't see the scale and the ROI from it as much.

Michael Kitces: Well ...

Shaun Tucker: Um-

Michael Kitces: Well, no, I-

Shaun Tucker: You know, and I think... Go ahead.

Michael Kitces: I think that's actually a really important thing to highlight in, in this dynamic, right? Because so much of the discussion, I mean, every time we go through on these technology waves, like, oh, the efficiencies, and the margin expansion, and the enhanced profitability, and I always come back to like the old Jeff Bezos saying, you know, "Your margin is my opportunity." Right, what happens when these technology waves come through, yes, they make us incrementally more efficient. The old things that we used to do get more efficient, right? We used to need Betty, uh, we don't need Betty anymore, she retired. But when you get those incremental improvements, you essentially get a choice, any firm, any business owner gets a choice. I'm freeing up time, I can either rest on my laurels and bank the profits, or I can take the freed up time and resources and reinvest it back into more things that I'm going to do for my clients.

And the reality is that for most businesses that want to keep growing, you don't rest on your laurels, you reinvest. If only because you wanna reinvest to do better things for clients, so you can take clients from the firms that just try to sit back and run giant profit margins instead of re-investing in their client proposition. And when you look at that writ large across the industry, to me, that's why if you look at, uh, uh, industry benchmarking studies five years ago, 10 years ago, 15 years ago, 20 years ago, and you think about the astounding amount of technology change that has not happened, margins have basically not moved in any material way in 20 years of technology innovation because we take all these technology improvements and we reinvest them right back into the services that we provide clients to do a little more, a little better, a little faster, a little, uh, more effectively.

I think clients benefit tremendously. Again, I mean, there's no question we do so much more so much more effectively for our clients now than we did 20 years ago because technology, but we run the same margins.

Shaun Tucker: Yeah.

Michael Kitces: And so when I look at it from the business end, like, yes, the technology still becomes an imperative, because if you don't use it to get more efficient, your competition will, and they'll do more for their clients than you can do with yours if you don't leverage technology. But I don't look at technology as a pathway for margin enhancement, just, I think it's the wrong framing in our business in particular. We're not, uh, factories producing widgets, where automation just lets us ship an unlimited number of widgets that come off the line. We are fundamentally a services business, and so if you don't reinvest your margin into your services, someone else is going to do that and try to take your clients, or at least win the new clients that you're not gonna be able to win anymore, and that's why when we look over long periods of time at the evolution of technology, we get better and do more and continue to be successful, profitable businesses, but margins for the industry and the aggregate don't improve through these technology waves. It doesn't show up as more efficient for the advisor, it shows up as better for the client.

Shaun Tucker: Yeah. So I'm gonna put a pin in this 'cause I think-

Michael Kitces: Yeah.

Shaun Tucker: There's a debate brewing here.

Michael Kitces: (Laughs).

Shaun Tucker: We're not gonna do it on this call, but let's just put a pin in that.

Michael Kitces: Sure.

Shaun Tucker: 'Cause I think the million-dollar question is, is it different this time?

Michael Kitces: Uh-huh.

Shaun Tucker: Brock, I know you think it is, Michael, it sounds like you don't think so. I'll just quote you and then we're gonna move on to part two.

Michael Kitces: Sure.

Shaun Tucker: But Michael, I don't mean to embarrass you, but you said in your recent podcast, AI ultimately cannot replace the empathy and personalized problem solving skills that the, that form the foundation of the client-advisor relationship, and, and I, I think you're a, a strong proponent of that. Um, great discussion, great banter, table it, maybe we'll put a wager on it-

Michael Kitces: Yeah.

Shaun Tucker: And come back a year from now and decide who's right.

Michael Kitces: Sure.

Shaun Tucker: Um, let's, let's move on. Get folks oriented. Brock, over to you. Let's talk about the what. There's a ton going on in AI right now, platforms, tools, players. Hard to get your arms around all of it, so can you just give us a framework for understanding the world of AI, ground us and what's important, if you could?

Brock Sutton: Yeah, absolutely. And I think this is a, uh, a pain point for a lot of folks. I think, uh, people think of AI as this kind of amorphous, you know, black box.

Um, and so I think it's really helpful to understand the five different layers of AI, not just as an investor, but also as a business operator, and so we're gonna go through each one of these. The three that we're gonna spend the most time are, are gonna be the model layer, the platform layer, and the application layer, and we'll talk a little bit about when you would use each one of those.

So first up is the foundation and the chip layer. Um, so this is going to be companies like Nvidia, uh, and Google with their TPUs and GPUs.

Um, and so this is going to be the foundation of AI. This is really the hardware that everything is, uh, built on top of, and this layer is really complex, so lucky, luckily for us, none of us have to necessarily worry too much about this layer. Next up then we have the cloud layer.

So the cloud layer is typically gonna be data centers, it's gonna be those hyperscalers, so companies like Amazon, companies like Google, companies like Microsoft. And really what they're doing is they're solving enterprise problems with scaled solutions. So typically, again, the only people that are going to be engaging at this level would be, you know, founders, CEOs, CTOs, people of that nature. So again, not necessarily something that we have to worry about today. Layer three then is the model layer, and so I'm sure folks on this, on this call have heard a lot about this layer.

This is where you have companies like OpenAI and their GPT series. You've got Google with Gemini, Mark Zuckerberg with with Llama, which is a open source, uh, uh, flavor. You've got Claude, which is really big in the Bay Area. They have a really strong partnership with Amazon. Uh, you've got Elon Musk and Grok series, and then there's a whole host of Chinese competitors in this space as well, with the most notable one being DeepSeek. And the model layer is the layer that is actually creating, you know, human-like capabilities, it's actually the algorithms, et cetera, that is creating intelligence. And so the reason that you'd want to engage on the model layer is if what you're trying to do is very differentiated from your peers.

So if what you're trying to do is truly unique to you, you may want to build that at the model layer. So an example could be your portfolio construction process or your wealth planning process, let's say that those are unique to you. That may make sense then to actually build that at that model layer, so highly differentiated. The next layer is the platform layer.

So we're gonna go back to some of the hyperscalers, so think Google, Amazon, Microsoft, and what they're doing here is they're giving you flexibility to solve a range of problems with a little bit simpler tooling. So one way to think about this, if the model layer is too complex but the application layer is not flexible enough, the platform layer may be the place to build.

Shaun Tucker: Hmm.

Brock Sutton: So an example is, let's say you've got a very convoluted client onboarding process, it works across multiple systems, and you wanna replicate it in the exact same way, that may make sense to actually then build at the platform layer. And then the last layer we have is the application layer, and so this is really gonna be products.

So this is where your ChatGPT, Microsoft Copilot, you obviously have this explosion of startups in this space that are solving problems, and what you're getting here is solutions that are easy to use, but typically they're solving very specific problems. And so if you're using the application layer, it's a great layer to use when your use case is very similar to the masses or it's similar to your peers. So maybe after this call, you want a summary and next steps, that makes a lot of sense to use just the application layer. And so one way to think about your use cases is an upside down triangle. As you leverage this across your business. Most things are going to be in the application layer, and this is where I really, I agree with you, Michael. Like this is where, um, this stuff is eventually gonna be commoditized, it's like you have to do it. You don't really have a choice, right? But then the bottom layer, the platform layer, the model layer, those were, those are where you should potentially build things that ...

Brock Sutton: The platform layer or the model layer, those were, those are where you should potentially build things that are actually differentiated within your business that are maybe harder for other people to replicate.

Shaun Tucker: Re-really helpful. Um, thanks for the frame. It's an education. I th-I think we all need to start thinking about what our strategy is as it relates to AI and that will inevitably kind of move you down that triangle into these other layers of AI, but really good, really good education for all of us. Um, let's stick with the application layer. Let's get into the how really important, and Brock, I'm gonna come back to you. Um, what are the basic things we need to know to start using AI tools?

Brock Sutton: Yeah, so we'll, we'll do two things here. First, we'll focus on how to get the most out of these tools and then we'll focus on where to apply them. So this is a slide, you guys may have heard of something called, uh, prompt engineering. It's just a fancy way to say how do you get this technology to do exactly what you want? So we have six tips here, um, but uh, I-I'm sure there's others that are really helpful, but these are, these are great ones to get started. So the first one is around providing clear instruction. So, if a stranger cannot understand what you're typing into the chat, then AI isn't going to be able to understand either. So make sure very clear, very direct with what you're asking for. The second one is to break down complex tasks. So one way to think about it is, every time you ask a question or leverage this technology, because it's probabilistic, there's a percentage chance that it gets it wrong. So if you're taking a very complex question with 10 steps and asking it to do all 10 steps, you're essentially compounding that risk.

Shaun Tucker: Okay.

Brock Sutton: So what you should do is break that apart, validate the answer before moving on. The third one here is using examples. This technology's awesome at pattern matching. So the more you can provide examples of what good looks like, what bad looks like, again, the better the output you're gonna get. The fourth one is around giving context. So one of my favorite tips here is to tell the technology, tell AI what you're trying to do, and then have it ask you 10 follow up questions around that. So, "Hey, we're doing a webinar for financial advisors around AI. Could you ask me 10 questions on that?" I then answer those questions and you get a much richer answer. The fifth one's around specifying format and tone.

So be really specific with what you want: how many words, who's reading it, do you want charts, do you want graphs. The more detailed you can be, the better. And then the last one's the most important one, and this is around just testing and refining. So, it really is right now a skill set. So give yourself the space to play around with this technology to figure out how you get the most out of it. So what we recommend is take an hour a week, block that hour, take two tasks, and then try to do those two tasks in AI during that hour. That's a bit about how to get the most out of it.

And then the next slide is where you can think about, uh, app-applying it. So, we talked about conversational AI, we talked about reasoners and we talked about agents. I just want to give you guys a sense of where this technology is best, best at.

So conversational AI, this is gonna be great for things like brainstorming, ideation, content creation, summarization. Again, it's that intern responding off the top of your head. So, where you have tasks like this in your business, that's where it makes sense to leverage conversational AI. Reasoners is gonna be great at things like data analysis, decision-making, complex scenarios. So, anything that has a binary or mathematical right or wrong answer, again, reasoners is going to be a great tool that you can leverage within that. And we'll show some of this as we get into the demo. And then the last one here is agents. So again, we're just starting on agents, so there isn't a ton of super compelling agentic tools yet. Um, they're coming out, and this is when you wanna fully automate something out. So let's say, you know, client review, you don't wanna focus on scheduling that, you actually wanna have an agent go back and forth between your client and your calendar. Things like that are great use cases for the agentic layer.

Shaun Tucker: Fa-fantastic. I think it just gives us a way to think about getting started and focusing on that application layer where we're all, I think, kind of living, but, but understanding what's gonna make that better and enhance it. Um, I would just say that, you know, we're, we'll, we'll make ourselves available to help with some of these application layer, uh, engagements. You know, prompting is a great example. There are a million better ideas in terms of getting into that, um, and, and, and utilizing it, uh, to solve specific problems, uh, for your practice. So to that end, let's segue into the demo and let's get into, uh, you know, what an actual use case would look like to showcase the power of what we're bringing forward in AI and, and how, uh, folks can apply that at their practice level. So, so Brock, let me take it back to you.

Brock Sutton: Awesome. Okay. So, uh, just, just need to warn everyone. Whenever you do a demo, there's some risk it could go wrong so that, that-

Shaun Tucker: No-

Brock Sutton: ... may happen. So, uh, that, that's my biggest fear. Um, yeah.

Brock Sutton: So we're gonna do a use case around client acquisition. Uh, there's gonna be six steps within this that we're gonna leverage. So first we're going to use it as a research analyst. So, we are going to use something called deep research. We're gonna use that to find our ideal client who we should target for organic growth. Then we're going to use it as a creative agency to build on that research and actually create a compelling value proposition based on that research that we just did around who we should target. Then we're gonna use it as a content director to actually develop a quarterly calendar based on those two artifacts. Then we'll use it as a writer to write a blog post and a visual designer to actually create a, uh, design, uh, ad creative. And then the last section we'll use reasoning AI t-as a digital marketing analyst.

And so what we'll do is we'll give it paid, uh, LinkedIn campaign data and we'll have it analyze that data, tell us what's good, what's bad, and what we should do next. So let's get into it here. So if we could share the, uh, the screen here. What we're going to be using, uh, for this demo is ChatGPT. And the reason that we're using that is because, again, based on some of the research that we've done, that is the number one tool that is still being leveraged by financial advisors. And this first question is really focused on, um, organic growth. So, I'm gonna set up this situation. This is something that we ran prior to the session today. The reason that we did that is because it takes anywhere from five to 20 minutes to actually do this, so we didn't wanna have to sit here and, and have dead air time.

Um, but we'll show you what we prompted, we'll show you the output and then we'll do everything else real time. So, we set up the situation and gave context first. So we're an advisor in the Chicagoland area, we work on a three-person team. We talked a little, uh, about who we serve and we said we want sources of organic growth. Okay? Can you help me build this research report on who we should target? We want geographic and generational data. We want trends that we should take advantage of. We, we want underserved segments that are high net worth, but our f-competitors don't focus on, we want their top financial concerns, digital channels, common topics, et cetera. Okay?

Then typically what happens within this deep research tool is it comes back with some clarifying questions. So then we answered those and then what's really cool down-

Michael Kitces: And-

Brock Sutton: Yeah, go ahead Michael.

Michael Kitces: And Brock, just, I was gonna highlight quickly even for that, that segment that you were queuing up for those of us who are not the fastest typists, you can talk that in a ChatGPT as well. There's a little audio thing in the lower right-hand corner of that, uh, uh, of where you t-of where you type. So you can dictate those questions or you can full-on use voice mode and it will talk the answers back to you. Just like noting, for those of us ... I'm a fast typist, so I love this. Um, Carl Richards from our podcast likes to highlight this. He is not a typist. He likes to talk. So, he does all of this in voice mode. I, I prefer to type it, so to each their own, but it's a nice option for those who are not typers. You don't have to anymore.

Brock Sutton: No excuses not to use this is, is, is my take with that.

Michael Kitces: (laughs)

Brock Sutton: Um, okay. So what's really cool here is it says research completed in 14 minutes in 20 sources. So this means that this agentic tool went out to the internet, researched for 14 minutes, read 20 different sources, and came back and built this report. And what's cool is you can actually click here and you can see the activity at a high level.

So, this is actually what the agent is doing at a high level, what it's thinking about, where it's visiting.

And then if you want to visit any of the sources or s-or source check those, you can go down here as well.

So, we're gonna go through this p-report fairly quickly because it's anywhere from 10 to 20 pages long. And so we'll just give you a few of the highlights and then we'll do a summary here at the end, but it's focused on the Chicagoland area, exactly what we were asking for.

You can see here it's talking about affluent city neighborhoods, exclusive suburbs.

So where should we actually target?

Now we're getting into emerging trends and wealth transfer patterns here, the great wealth transfer.

And what's cool here is you can also see where it's sourcing this information, uh, is at.

So, if you have any questions or you don't believe the sources, you can actually click through and say, "Let me, let me validate that."

Younger high net worth segments.

Again, we're gonna go through this pretty quickly, but we'll give you a summary here at the end.

Now it's going to underserve high and, and high net worth segments, specifically within that Chicagoland area.

So younger high net worth professionals, first-generation wealth creators, niche professional segments, cultural communities.

Now we're getting into some of the behavioral and demographic insights, top financial concerns. So, very detailed here and exactly what we had asked for. And then we've got digital channels.

So where should we actually target these folks?

So I'm gonna scroll down here to the bottom, and just to give you a sense of some of the insights here, I'm going to actually type in, "Okay, so this is great. We really like this research report, but hey, we don't have time. Can you give the 10 most-

Michael Kitces: (laughs).

Brock Sutton: ... surprising insights from this research report that would be valuable to a Chicagoland financial advisor?" And then we're gonna put them in bullet format.

So, Chicago's millionaire growth outpaces peers, next-gen high net worth clients want to give early, self-made wealth is now the norm in Chicago. Women are gonna control 30 t-uh, trillion by 2030, et cetera.

So some really interesting kind of insights here that we then wanna build on top of. So, we just got done using it as a research analyst to understand who we should target in that Chicagoland area. Michael, I'd be curious, have you, have you used Deep Research? If so, have you used it for anything that, uh, that you're willing to share? Anything interesting?

Michael Kitces: Yeah, I've been doing a lot of, uh, experimenting with, um, uh, with Deep Research lately. Its, its ability to go and find lots of resources are helpful.

So, we've been experimenting anything from like, "Hey, I, I need a pri-I need a primer on a unique, uh, tax strategy that, uh, maybe unfortunately we haven't written an article on, on (laughs) our own site yet." So, you could say things like, you know, "Give me a primer on qualified small business stocks, section 1202 stock because I've got a client with such and such a situation." And it will give you like, "Here's what the rules are, here's how they work. Here's some follow up questions about your client situation. Let's see if it applies in your client situation." And can kind of talk through a planning strategy or a planning scenario. Uh, we've also used it, uh, probably less geless dee-less Deep Research and more the main model itself.

Uh, I've used it a lot for these kinds of brainstorming sessions around marketing strategy, right? Who are we going after? How are we communicating with them? Uh, we did an internal exercise where we were trying to update the strategic plan for the business itself and trying to figure out how to present it to the team. And so basically told ChatGPT like, "You are, you are my management consultant that's helping me do a strategy rollout for the team. Here are the objectives we wanna talk about over the next 10 years, three years and 12 months. Help me figure out a way to present this to the team. We have to organize it into four logical categories 'cause I can't go with like 17 initiatives." The team, we kind of, kind of boil it down to a couple of main themes and it helped with all that thematic brainstorming as well.

Brock Sutton: Yeah, those are, those are, those are great use cases. Um, so, so kind of on that front, the next thing we're gonna do is take this research and we wanna build it into a value proposition.

So, can we build that compelling and unique value proposition for a financial advisor targeting these recommended niches? Use best practices from creative agencies and marketing firms and focus on key messages. And you can see here it did exactly that. It gave you a great first draft that you could potentially leverage to target those clients. So it's got brand positioning statement, unique value proposition, messaging pillars for those specific segments here, specific to Chicago suggested CTAs that we could leverage, et cetera.

So let's say we like that, we think that this is a compelling value proposition for us. Next, what we're gonna want to do is say, "Okay, let's use it as a content, uh, editor. Let's say we want a quarter's worth of potential topics that we can talk about and different types of events."

So we want a blog post title, we want a social media headline, we want a webinar topic, and we want event idea and we want this in a table that's easy to read. So we're being specific about how we wanna get this information.

Shaun Tucker: Yeah.

Brock Sutton: And then, you know, tariffs is in the news. I'm not sure if you guys have heard, so week one we are going to actually focus that on tariffs.

Shaun Tucker: Mm-hmm.

Brock Sutton: And what you can see here is it came up with exactly that.

So it's given us all these topics throughout here that we wanna focus on for these different 12 weeks that we could potentially leverage within this. So again, this is a great brainstorm partner to get you started. I'm sure you're not gonna like all of these, but I'm sure a handful of these ideas are gonna be compelling enough for you to actually go ahead and leverage there.

Shaun Tucker: Michael.

Michael Kitces: And Brock, I would, I would highlight as well these kinds of frames of you give me, give me some ideas of topics I can, you know, write about, speak about, video about, audio about, right, depending on what your marketing, uh, uh, strategy is. AI tools to help ideate on those are really effective. I mean, we're increasingly using this internal for our teams as well to say, "Okay, you know, we want to cover this topic. What are, what are three angles that might be interesting to the target audience that I'm trying to reach?" Uh, okay, we wanna go with this one. Gimme five headlines that you think would be engaging for them."

And I find it, it's usually not, not perfect, particularly if you ask it for one that it, like, it gives you the one perfect thing. But when we ask for three or five at a time, we almost always find one that's effective or one or two that we can blend together. So I might still know my audience better than the AI, but it's so good at filling a blank page when most of us get like paralyzed by a, a blank page, right? As, as I joke sometimes with the team, like lots of people have writer's block, no one has editor's block. If it makes the first draft, it's easy to edit. We get stuck making a first draft. So, it's really good at getting started.

Brock Sutton: Yeah. And, and that makes me think of talking about how your team is using it. I think the other week, uh, it was, I think it was the Shopify CEO came out and basically said, "AI is, is not a, it's not optional anymore. Everyone has to use it." And you essentially had to prove out why you needed head count by trying AI first, and you had to prove that AI could not do what you were actually asking for.

Michael Kitces: Yeah.

Brock Sutton: I thought it was a really, it was really aggressive, but a very interesting perspective. So next, let's say we like this, we think this is compelling, but we actually want it to write an article for us. So, draft a short LinkedIn article.

Michael Kitces: Yep.

Brock Sutton: We want it 500 words or less. Introduce this weekly topic, focus on the tariff theme and then we're being specific here. So, we want a friendly yet professional tone. We want a clear call to action and we want it attractive for our target client. And what you can see here is it came up with exactly that, "How the 2025 tariffs could quietly reshape your portfolio." Got a space for our name, it's got the 500 word article here, and then as you scroll down here, it should have a CTA as well.

Brock Sutton: ... have a CTA as well. So the last step in this process, Michael, have you used the, uh, the image generation feature, the newer ina-image generation feature much yet?

Michael Kitces: No, we have not played with the new image generator yet.

Brock Sutton: Okay. Well, this is gonna be a fun one then. So-

Michael Kitces: Cool.

Brock Sutton: This takes a little bit of time, so we'll, we'll let it sit here, but I'll give you the prompt beforehand. So we've got the article, but now we actually need an image to go promote on LinkedIn. So we're gonna say, "Okay, can you create an image that would be used on LinkedIn to promote the article that's about tariffs?" We want it to be attractive to the target market, and we want make sure it's focused on the Chicago market. So we're being really specific there. And then we're saying we wanna make sure that it, it shows that we provide holistic wealth planning, and we want to call ourselves the Michael, Shaun, and Brock Group. So we'll make sure that that's

Michael Kitces: Yeah.

Brock Sutton: ... included within the ad as well. So we'll let this do its thing. And in the meantime, let's say that we went out here and we actually ran a LinkedIn campaign. So we went out, we ran a campaign and we got data back from a paid campaign targeting those clients with that ad on LinkedIn. You get that data back and what do you do with it? Are you a marketing analyst? Probably not. So let's use AI and a reasoning agent as that marketing analyst. So what we're gonna do is we're gonna click on a new model here. It's o1. Anything with an O in front of it, that means it's a reasoning tool. No, I don't know why, uh, OpenAI has the naming problems that they do, but that's just the world we live in. The other thing that we're gonna-

Michael Kitces: (laughs).

Brock Sutton: ... The other thing that we're gonna do is we're gonna upload our LinkedIn data. So this is an Excel file with typical data that you would get from LinkedIn. See if it'll go here for us. There we go. And then we're going to tell it that it is going to be a marketing digital analyst for us and actually analyze this data.

Michael Kitces: Yeah.

Brock Sutton: So we'll go through, here, we've uploaded this data, we'll click through and essentially what we said is, "Hey, this attached data is from a paid media campaign on LinkedIn. It was a campaign promoting financial advisor targeting their service to niche markets. We want you to analyze this, tell us what's good, what's bad. We want charts and graphs." And then at the very end, what we're doing is we're being very specific and saying we want five recommendations. I don't know about you guys, but five is probably the highest number of recommendations-

Michael Kitces: Yeah.

Brock Sutton: ... that I can manage within this. So what you can see here is, again, because it's a reasoning agent, it's actually thinking through the problem, which is pretty cool. You can actually see at a very high level what it's thinking about, how it's making those decisions. And then the other thing that I would wanna call out here is what it's leveraging here. So it is using a tool. So it's using Python to analyze this data. And so it's actually writing Python directly to the table to actually better understand the data, to give us the recommendations that we would ask for. So we'll scroll down here. You can always close these out if you don't wanna look at it.

And we've got our first chart here. So impressions by campaign, you can see biz owners and tariffs has the highest number of impressions.

Then we've got the click-through rate here.

Michael Kitces: Mm-hmm.

Brock Sutton: It looks like we've got a tie here between younger high net worth tariffs and estate planning, the cost per lead.

So the most effective cost per lead is for estate plan at $50 per lead.

And then conversion rate here. We've got some key observations, potential reasons. So it's, it's thinking through the differences.

Michael Kitces: Yep.

Brock Sutton: And then at the very bottom here, we should have recommendations. Perfect. So here's our, here's our final recommendations here. Great.

So I want to go back really quick and I want to go back, uh, to this one here to see what we got as an image. Perfect. So now we have an image-

Michael Kitces: There you go.

Brock Sutton: ... that we can potentially leverage on LinkedIn, how the 2025 tariffs could quietly reshape your portfolio, holistic wealth planning. The Michael, Sean, and Brock Group. There it is. Look at that.

Shaun Tucker: Need some. Uh, there you go.

Brock Sutton: So just a quick re

Michael Kitces: I must be financial planning.

Brock Sutton: ... So just a quick recap to-

Michael Kitces: (laughs).

Brock Sutton: ... to talk about what we showed everyone. So we, we first used it as a research analyst to research who we should target. Then we use it as a creative agency to create the value prop. Then we use it as a content director to come up with that content calendar. Then we use it as a writer and visual designer. And finally as a digital marketing analyst, to actually analyze the campaign and give us a recommendation. So a fairly holistic soup-to-nuts use case.

Shaun Tucker: Um, lemme jump in 'cause I'm sensitive to time here. Gentlemen, thank you. It just showcases the power of this tool, right? And you're not alone in this. I heard a lot of things that I think are important encouragements and, uh, and sort of best practices, not the least of which is start, make a start, right? And, and start kind of rolling up your sleeves and playing with this. You'll get better as you go. Uh, rest assured to everyone in the audience, the prompts that we just went through in this use case are available in the appendix. So that's one.

Michael Kitces: Yeah.

Shaun Tucker: Two, there are some really cool tactics that you can utilize. So if you're saying, "I would never be as good as kind of putting these specific prompts in, you can actually ask AI what questions haven't I asked? How can I make these questions better or more sophisticated for my target market? Tin-Things like that that actually help refine your prompting. And then we're standing by to help. Michael is available, the Capital Group, our team and practice management is available. So I just wanted to mention that.

Um, but let's get started. And to that point, maybe we'll finish just with how to get started. And Brock, maybe you can hum a few bars on that. And then we're gonna go to questions where I think we can tease out some of the specific concerns that people

Michael Kitces: Yeah.

Shaun Tucker: ... might have and, and anything else.

Brock Sutton: Yeah. So one of the things that we have been helping advisors with is, I think right now there is a really reactive approach to leveraging this technology, which makes sense. So you've got a job to be done and you go to this technology to, to help you do that. But what we're helping advisors with now is think of a systematic process for actually applying AI strategically to their business over the next three, six, nine, uh, 12 months.

So up here on this slide, we've got, uh, at a high level what that process is, and I'll just touch on it briefly, but this is something we can certainly help, help you with. The first step is prioritizing where you want to apply AI. So understanding your business processes and where you should start. The second step is breaking those down and recording key details. The third step then is matching that to AI, the tools, the capabilities, all the things that you have access to based on what you're doing in each one of those steps. Step four then is taking this, sharing it with your team, validating it, and starting to have them record, how much time is this taking after they've started leveraging these tools?

And then step five is optimizing scale. So this is actually measuring where am I getting the most efficiency within AI within my practice? And then how do I move back to step one to figure out where I should apply AI next? So holistic process for thinking about applying AI strategically top down to your business.

Shaun Tucker: Terrific. And, and I think this process will be key to getting started successfully. And, and again, we'll help you along the way. Uh, let's move to Q&A for the sake of time and, uh, I'll throw this first one out there, um, because, you know, I think it's, it's some, it is somewhat emblematic of a concern that, that a lot of folks have, which is on the quality of the data and whether or not there are kinda hallucinations out there, inaccuracies in the data. Uh, how do you know whether you should trust it? So here's this specific question. LLMs search the internet, there's so much bad financial information on the internet. How do we determine whether an AI tool has good inputs?, you want to comment on that?

Michael Kitces: So, so from my end, like it's, it's, it's, that's not what I'm asking it for typically in the, in, in the first place, right? In, in, in the context earlier that I think Brock framed well, right? There's, there's a long list of job tasks that we have and technology in general tends to knock out the lowest rung. Like, the first thing I want to use AI for is not here's my comprehensive financial planning situation. Please, uh, analyze the client's scenario and develop the financial plan for me. That's like a really sophisticated, complex thing.

It might be, here's the transcript of the discovery meeting with the client, please summarize the three things I should make sure I cover in the plan. Because I may not be the best, best note taker in going through this and I don't wanna miss anything. Or you, Brock's good example here around the marketing plan. I mean like, look, if it is slightly off about the optimal wording of the best article you should write for next generation wealth creators in Chicago, you're probably still gonna be close enough. Like, I don't think it's hallucinating the fact that people create wealth in Chicago. This is probably like reasonably on target.

Uh, maybe it was two percentage points off in the actual growth rate of millionaires in the Chicagoland area, that's not really gonna materially change the output, the outcome of your marketing strategy. So you don't start with the things that are like the highest stakes, most complex, most nuanced answers. You start with, "Hey, I just have really never had a structured marketing plan and I'd like to create one. So I'm gonna let the AI be my brainstorming buddy." And, and to the comment of, well, I don't even know if I can write the length of prompts that Brock wrote.

There's also a version of this where the first prompt is, "Dear AI, you are my marketing consultant, what should I have in a good marketing plan?" And it'll tell you. And then you say, "Great. Make that for opportunities in the Chicagoland area."

Shaun Tucker: Yeah.

Michael Kitces: And then it'll-

Shaun Tucker: That-

Michael Kitces: ... do all the things and you don't even need to know what to ask it. So I, I start there, right? The, the, or, those are domains where if it is not perfectly accurate in every technical bit of information, it does not really matter. Like, you're not extracting advice to which you have fiduciary liability, you're trying to figure out the title of a blog post that might connect with people who'd be interested in doing business with you.

Shaun Tucker: Yeah, I really like that Michael, let AI do what AI does well and stick to what you do really, do really well and, and combine together, I think you have a really unique value proposition. One more question. I'm so sorry that we're, we're kind of nearing time here. This is something we could spend a long time on, but great question came through. What is a specific example of using AI to improve the client experience? Not the back office, um, but really the client experience? Uh, any quick thoughts?

Brock Sutton: I mean, I think, I think one of the holy grails for folks is, uh, the interactive, uh, chatbot, uh, through the website where they can get immediate answers that's a representation of their firm. I think one, one of the concepts that's out there right now is, so it started with websites, next, it was apps. Each company had their app.

Shaun Tucker: Mm-hmm.

Brock Sutton: And then next is it going to be each company has their own chatbot? Right? And is that just how you interface with technology and get some of those immediate answers back and forth? And I think, whereas apps were still, you know, somewhat cost-constrained where small businesses, you know, wouldn't, wouldn't want to develop their own apps. I do think some of this technology is going to be cheap enough where anyone could develop a version of them that their client could immediately get information back from. And especially for clients who don't like-

Shaun Tucker: Yeah.

Brock Sutton: ... going through the internet, that could be a simple, a simple example. It's a little bit farther out there, but I, I think that's a, a compelling one.

Shaun Tucker: Quick, fast follow. Another question that was asked are, does this mean that all of our outputs will be the same? If we're all using AI, are we all gonna be in the sea of sameness? Any quick 30 second response?

Michael Kitces: I, I, if we all ask the same generic questions, perhaps it starts to homogenize a little bit. But first just, I would say, I mean we're, we're all gonna have a different angle around how we ask this the moment we start teaching it, how we approach financial planning, how we approach client service, the clients that we serve. We should have a different business simply because we serve differently and, and work with different clientele. And if you teach the AI that that's what you do and who you serve, we will start to look different on that basis alone.

That being said, I think the other layer always is we still get to add the value on top. Like, we ship nothing that the AI outputs, we take the AI output 'cause it is faster to edit than create and then we make it into what we want to be. And so even if we start from a similar, you know, slab of marble, we can all craft a unique sculpture out of it based on what we do.

Brock Sutton: And, and the one thing I would also say is-

Shaun Tucker: That's good.

Brock Sutton: ... uh, ChatGPT just came out with a feature called memory. And so it remembers your previous chats now. And so-

Michael Kitces: Mm-hmm.

Brock Sutton: ... it's creating a, a version of you, right, of a picture of who you are. And so your outputs now with memory turned on are gonna be different. You know, all three of us would have a little bit different outputs based on who we are.

Shaun Tucker: Yeah. Well, and Michael, you said it really well. Everybody has writer's block, nobody has editor's block. And so, you know, I think that we reserve the right of what we actually put in front of the client. Okay. I hear the Oscar music, it's sounding-

Michael Kitces: (laughs).

Shaun Tucker: ... uh, it's time for us to wrap up. So let me close. Uh, Michael, Brock, great discussion. We can spend a lot more time on this. Thank you both for your time, your expertise.

Um, and before we wrap, I just want to bring everyone's attention to some resources that you can take advantage of immediately. The first, uh, that you mentioned that we show here on the left is an article that we recently published. It details five specific ways that advisors like you have been using AI in their practices. It talks about the note-taking function for client meetings. Uh, Michael, you've mentioned that. Client emails, how you can use it for tax planning and summarizing your, your different advice.

Uh, if you want more on all of that, go to our practicemanagement.com, uh, website and, and it's all organized for you there. Do not forget the document section on the right-hand side of your screen. Again, if you wanna get CE credit, the quiz is there as well.

Um, and let me just stop there. On behalf of Michael, of Brock, myself and everyone at Capital Group, thank you for joining us today. We really hope that you found it useful and we hope that we did the three things we set out to do, which is illuminate the why AI can be a key growth driver in your business. Hopefully, we covered the what and educated you on what do you need to know that's important, those five layers. And then the how, guiding you on how to get started, a practical framework, that five-step process, and then ultimately a strategy. And we're here for you through our partnership to help you with all of that.

If you have any questions, follow up with your Capital Group team. They're standing by to help. And with that, thank you so much. Have a great day.