AI town hall: Implementation for today, strategy for tomorrow

Join the AI Town Hall to hear from industry experts on new and exciting tools to advance your business, improve efficiency, and reimagine the client experience. Artificial intelligence, machine learning, and predictive analytics have long been primed to reshape financial services. But as 2023 rolls on, it feels like we're reaching an AI tipping point. Tools like ChatGPT have suddenly become household names as AI use cases expand, and the "what if" possibilities are growing more exciting as advisors realize what the technology is capable of.


 In this session, panelists will discuss the practical applications of AI in today's industry, ponder where these innovations may take us tomorrow, and explain how the cutting-edge tech is helping firms grow in a more human way.

Transcript :

Justin Mack (00:06):

All right. All right. Good morning and we're continuing to roll with the agenda here at Invest 2023. I want to thank everybody for joining us for our AI Town Hall, all about implementation for today and strategy for tomorrow. By this point, you're probably tired of my voice, but if you're just joining us at Invest for day two, my name is Justin Mack, the Wealth Tech Reporter for Financial Planning, as well as host and Lead Editorial Producer of the Financial Planning podcast. And really excited to have a group of gentlemen that I've had the pleasure of speaking with before and bothering quite a bit on not only this topic but many others. Joining me is Dr. Mark Evans, President and CEO of Conquest planning, Scott Lamont, Director Consulting Services F2 strategy and Ritik Malhotra Co-founder and CEO of Savvy Wealth. Quick round of applause for my awesome panel this morning. All right. And I don't really think at this point we need to set the stage and explain why AI is, well a topic that we're talking about right now. It feels like you can't avoid it. Even if you're not interested, you're going to end up almost tripping and stumbling into the topic head first, but it's a big one. It's one that we already see kind of having an impact on how we do things more efficiently, changing the approach we take to not just this industry, but so many other businesses, industries are trying to wrap their hands and their heads around how exactly do we use this? What's the best use case? And all the head turning, head scratching, eye grabbing headlines that maybe cause some worry in a business sense, does that stuff really matter or is it something that we should actually be taking more seriously? So a lot to jump into today. So with this panel, I wanted to start by just talking to everybody about their personal experiences with AI and I'm excited because we have a lot of different perspectives on this stage. Someone representing the folks who are creating tools for advisors, somebody who is eared to the streets, always kind of analyzing the industry and seeing where we're headed next and someone who is starting a new way of being in this business, a tech-focused startup, RIA that has a brand new model where tech is the very underpinnings. So really cool perspectives. And Mark, I want to start with you not only just because right to my left, but for better words, you are kind of the OG here for our panel. If you don't know Mark from Conquest, he is also the architect of the NaviPlan financial planning software that first got its roots back in 1990 in your college days. So will say a far more enterprising college student than I was when he was in college. He's creating Financial Planning software. I'm trying to figure out which group callouts I can Goku to get free pizza. So a better use of your time, but Mark first a little bit about your experience with AI. How have you touched it so far and in what ways are you using it now?

Mark Evans (02:45):

Sure. Well my roots in AI go way back. I actually have a PhD in artificial intelligence in 1988. So been around the industry for a long time and seen a lot of changes in AI from the eighties through the nineties and so forth. Back in the initial days, a lot of it was rule-based expert systems trying to use AI to help experts or non-experts do their tasks more efficiently, look for ways to optimize that technology. At that point in time though, wasn't really advanced enough both from a hardware perspective like computing power perspective and even from a technology perspective. I think things have, if you fast forward, things have gone a long way on both those counts. We now have hosted environments that are extremely powerful and flexible and scalable and the technology of how we can integrate things into tools, how we can access data, and just the overall advancements in processing. So we're looking at using AI to enhance our financial planning tool. So as you indicated, built a tool called NaviPlan, sold that off in 2011, came back five years ago with Conquest, started from scratch learning what we had had learned at that time. One of the key things we did in the initial development of conquest was built in some AI search capabilities into the engine. So the tool looks to find the most appropriate strategies for a given client situation. So it's tailoring it to the client that you're working with at any given point in time and serving up strategies and actually ranking those strategies for you. So it's assisting the advisor in finding the best strategies for the given client. We're now enhancing that even further with the enhancements that have come out lately with chat GPT and large language models and so forth to build a natural language interface on top of our Financial Planning tool. So normally you would interact with the tool by going through menus and going to a planning page, going to a what if page, looking at different, here's our going to generate a client report. So the user has to learn how to use the tool by learning and navigation. We've enhanced that now by using some of the language model capabilities out there to build a natural language interface where you can say to our little built in assistant called Sam, Hey Sam, let's start planning takes you to the planning page. Hey Sam, what strategies are best suited for this client? I found the following five strategies, would you like me to apply One boom, boom, boom. So, oh, what if the client lives 10 more years than they expect to, I'll create this. What if for you client lives longer show you on the screen so it's navigating without you having to learn how to navigate the tool, you're basically interacting in a natural way with the software and we're using those language models to do that to facilitate that. So we built that on top of our existing tool.

Justin Mack (05:56):

And Scott, like I mentioned earlier, the consultancy work and the research work you guys do at F2 strategy, it's always that ear to the streets, finger on the pulse, trying to really help advisory firms figure out what's coming next and how to best respond to those changes. So with that, I imagine AI's probably been something you've been thinking a lot about just because I'm imagining you're getting questions and trying to figure out how exactly do we guide people through whatever's coming next? Because there's a lot of big concerns, but as Mark kind of greatly pointed out and something we'll talk about more is that on a more practical level, those applications are going to be way more, I guess tangible than some of the big things where people are worried about these overarching ideas. So for you, what are you focused on now and how's your relationship been with it and has that changed at all as more people have more questions and more headlines roll out?

Scott Lamont (06:41):

Yeah, it's been interesting in the last six months to a year to see, we get a lot more questions about it. We get, hey, I want to use this tool, or I think AI can be really great for me. Can I use that here and there? And when we get into the actual conversations about the underlying data that you're going to use and how it's going to make you more efficient and do you want to use it with your clients and you know, start to uncover a lot of other challenges that go along with simply implementing a new tool. But I think on the surface of it, being able to use it for, we talk with our clients a lot about this idea of time poverty and our clients being the advisor firms about making your advisor teams more efficient and how do you service as you're continuing to grow. We spend a lot of time working with smaller RIAs that are growing rapidly and they're trying to expand their client base, they're taking on more work, they just don't have the time to be able to go through and do all of the assessments and all of the data collection and all the preparation of new proposals of performance reports of projections and plans and so forth. There's a lot to do. How do you give them some tools to make it more efficient to build that initial proposal, to build the performance report, have that information prepared. We even three or four years ago, looking at natural language processing tools that you could layer on top of a performance data set and draft the advisor a statement of this week, this month, this is how your performance, this is how your portfolio performed. Dr. Crosby from Orion earlier mentioned, I don't remember the exact phrasing of the stat, but the correlation between engagement and satisfaction. So the more time that advisors are spending with clients, the more time that they're talking to them, that's much more correlated to satisfaction on the client side than simply what my portfolio did. An extra percentage point or two isn't going to sway satisfaction more than an extra phone call or two, a weekly newsletter of here's how I performed. How do you do that at scale? You start to use these tools like ai, the chat GPT tools and natural language processing tools. But at the end of the day, so much of it is, and we spend a lot of time thinking about that underlying foundation of information that's available to then layer the new tools on top of it.

Justin Mack (09:23):

Absolutely. Now I'm really going to you from the OG to essentially new kid on the block with Savvy Wealth. This is a new approach, digital first digital native of RIA. And you guys have, as we've talked about before, hit the ground, sprinting been around since 2020, is that correct? 2021, yeah, 2021. So just a few years but have done a whole lot and with tech kind of really being the core of the basis of everything you're providing to provide what you're doing with your. Give us a little bit about what savvy's about and how AI is now being a part of what you're providing in that service.

Ritik Malhotra (09:53):

Absolutely. So just again for context, we are an RIA and trying to implement technology to augment the client and advisor experience. So we actually have been using AI for the last 12 months to augment the actual advisor and helping them with work with their clients. And the three areas that we've found work really, really well have been around marketing have been around personalized communications and then also on compliance. Basically augmenting what the advisor's doing and making them much more effective. So on marketing what we've found is it's really useful for an advisor to build content, both written and also imagery that actually helps them advertise online. And that's actually been super effective, kind of reducing the amount of time they spend to do that by 80%. The second one on personalized comms that we've actually found is simply just responding to prospects or clients' responses. Actually coming up with a first draft of what to respond with and then introducing the actual kind of knowledge that they need to as an augmentation also does the same effect kind of helping them. And the third one on compliance, which frankly is very required but folks don't like doing. We found things like transcription, summarization, et cetera, actually help a lot kind of reducing that burden as well. So those are the three things that we found kind of practically working. And that's a combination of more traditional AI and also the new generative AI with large language models working in the wild.

Justin Mack (11:13):

Alright, Awesome. And after talking about your personal relationships, one thing I want to get into before we get into even more of the meat of the conversation is just your thoughts on the interest you, Scott kind of mentioned in the past six to 12 months, it's been interesting to see what you're providing as far as guidance has changed. And that's really been driven by I think the craze, how much interest has come from folks even outside of a business sense. The fact that I've got folks like my not so Tech Savvy aunts and uncles asking me what's about what's up with the AI and the robots and it's driving interest in a very human way just because it is eye catching when you see some of the stuff that's shared online, it's a, wow, I can't believe it's doing this, what is this going to mean? But as we've all kind of talked about, AI is not brand new, you know, got your PhD in it in 1988, it's been around but I think it's kind of broken out of the confines of the practical applications people have seen has become flashy, become flashy, rid. I'll stick with you just your thoughts on just that craze and what about it right now. Has anything new captured you? Has anything grabbed your interest right now and what are you all about right now?

Ritik Malhotra (12:16):

Yeah, I think if you think about the craze, there's been this explosion, we call it the Cambrian explosion of all these consumer generative AI applications, tap chat, GPT being the most popular one, but there's hundreds that have come out in the last six months. And one of the things that I think people get really attached on is there's a lot of application in this consumer AI world, but if you actually look at the data, if you look at month 1, 2, 3 cohorts of people signing up for these, there's a huge amount of churn because what's happening right now is kind of the natural speculation curve, a huge fad, it looks really cool, but the actual practical applications are lost on the consumer. So what we actually think is that the enterprise AI case is actually a lot more attractive. And I think all three of us and all of us here actually on the panel here probably believe that where we fundamentally believe that generative AI and just broadly AI can actually help the knowledge worker be much more effective at their job. And if you actually drive purpose-driven AI tools, so for example, if you are building something to help with compliance from an AI perspective, that's actually going to work well because the user knows exactly what the purpose is, it fixes a certain problem and you're not giving this open kind of ended tool to someone that's trying to figure out what the actual kind of key silver bullet use cases. So that's our view.

Justin Mack (13:28):

And Scott, that's a great point that really makes, I would love to ask your thoughts on it, that natural churn that after the wow factor kind of fades and the folks who aren't thinking about using this in a business system or even as part of their daily lives to maybe make just their personal lives a little more efficient, that's going to fall off, that's going to drop. But as we know the business applications are still being determined. How can this new version of it help us, make us drive more efficiencies? And what we've seen in our research, and I'd love to see if you saw some of the same is that there's a lot of wait and see approach. So while that consumer or external business is kind of fading off, we're still kind of in that law because waiting to see what that big application's going to be, your thoughts just kind of on that fade off and drop off and the wait and see approach that we're taking from a business side.

Scott Lamont (14:11):

Yeah, I think it's interesting. I heard someone talk the other day about this idea of this efficient chat GPT tool that's going to make us all more efficient was actually wasting people's time because they were all spending it talking about like, Hey, can you tell me what vacation I should take or this fun thing to test it out? And that we were ending up spending more time doing sort of mindless tasks instead of being more efficient. So I think that's the churn, you'll see the drop off, you'll see people kind of play with it. I think the counter to that, one of the things that we come across in working with our clients is great ideas in technology. We walk through the process of it, get new tools, stood up the challenge of getting people to then use the tools and adopt it into their everyday practice to make themselves more efficient or to process portfolios or trading in a more effective way. That can be a challenge in getting adoption to really pick up. I think there's an interesting opportunity here with the AI tools where you have people, I have an advisor that I've spoken to who will not use a CRM, will not, I use Excel, they keep everything in pap on paper boxes everywhere. Talking to me about how they can use chat GPT now in their life. They're intrigued by it because they've used it personally. So the fact that we've got this tool now where people are really interested in the opportunity that it could present from their personal life, I think we're going to see more of an uptick in, hey, can I actually now use this to make myself more effective? The challenge and the compliance piece of it is the question of where can I use it, how much can I do with the tool, how does it interact and does it just about making me more efficient? Can I use it with my clients? What's the data privacy pieces? So there's a lot of elements of that that I still think is why we see slower implementation and wait to see. But I do think the opportunity there with the buzz early on is an interesting twist to see how this gets adopted.

Justin Mack (16:24):

Absolutely, and Mark would love your thoughts on that buzz. As you even mentioned too when you came back with conquest in 2018, AI was built in IT from the jump, having those search capabilities and developing SAM in the way that you guys have, but then again, you're already doing your thing for many years Chat GPT hits the masses last fall at the end of the year. Things change a little bit. Your thoughts on that and has that changed the trajectory of what you're do at Conquest at all or are you kind of sticking with the course and understanding that the appetite is there so it only maybe will strengthen what you're working on? How do you guys respond to that?

Mark Evans (16:54):

I think a couple things. One, I think it strength is what we're working on because one of the problems you have when you're building a software application is as you add more features and capabilities, the application naturally becomes more difficult to use for the new user, right? Because there's more features, more things you can do. How do I find how to do those things and even realize that they're there, you have to navigate around. Whereas if you can just say I want to generate a client report, and then your little intelligent assistant comes up and says, well you can generate a PDF report or would you like to generate a PowerPoint like story? Oh, let's do the PDF report. Okay, I see you don't have a summary page. Would you like us to generate a summary page for you and then we can go out and pass the data and use generative AI to create a summary of that client situation that's personalized to them for the advisor. Again, Tori's point streamlining that instead of you having to type that in and do it, let a Chat GPT like engine generate that for you. But there's still a lot of practical limitations and questions about using things like that organizations, financial services organizations are worried about, well the data, to your point, Scott, about data privacy, if I take this data and I fire it out to open AI to do a Chat GPT like generation of text, what happens with that data and how do I, well, you can anonymize that data, sure, you can do that, but organizations still look and say that data is going off over to Open AI. I'm giving them access to my data, my information, and in large quantities that's valuable to Open AI. There's other ways you can do it. In our situation, we're an AWS shop, we can use their tools, we can build it inside our hosted environment so we can use their large language model tools and so forth within our own environment so that we can guarantee that the data does not go outside of the organization's environment. And that's something that's really powerful in terms of ensuring that the information is kept secure and they have control over all their data.

Justin Mack (19:03):

Absolutely. And that safety and security piece I know is going to be a huge one, even big one I think to Dr. Crosby's keynote this morning showing the level of trust in the financial services industry compared to others. So you introduce something else that maybe comes with some concerns when folks who aren't thinking about it from the business sense, again, regular consumers who see it, they have their natural human concerns coupled with an industry that already unfortunately still has some low trust. So what you talk about as far as how is this being implemented in a way that is making sure that the data is secured and we can show that to consumers upfront is really important. It even brings an example, something we had a chance to cover. Morningstar for example, launched a new AI Assistant, a named Mo. It's built on Chat GPT structure. One of the features they built in it, for example, is that it has an awful memory. You talk to MO after you get done talking to MO, he forgets everything you said to them and they said they built that in, for example, A bad memory is a way to keep information safe. So good to hear that that's upfront center. Something else I would love to get your thoughts on is just the way we talk about AI and as a reporter, and as I'm sure you guys have had these conversations, it's important that we're broaching these in the appropriate way because I find that AI is almost talked about in this extremely monolithic way, AI is this. AI is that AI will do this period strips the nuance out of the applications that we've already spent 20 minutes talking about because it is AI full stop. That's it. Talk to me about how, one, is there any way to chip that down and really communicate more clearly those smaller practical applications? And are there some ways that you're seeing AI leveraged, maybe even outside of wealth management that are being overlooked that I think could help communicate the fact that hey, it's not just one thing, it's not just coming to replace us and whatever, Skynet, all of that. How do we break through and communicate that a little more clearly? Mark, I'll kick things off with you. Is there a way to figure that out?

Mark Evans (20:52):

Yeah, I think to your point, AI is a right now is treated a lot like cancer. People think cancer, oh cancer, this is just one disease when there's 200 plus different types of cancer have their own nuances, their own ways that you have to deal with them and so forth. And AI can be broken down to different areas as well. I think the other issue is that AI is now in mainstream and you get the hype and you get the interest level of that. And for us in our day-to-day lives and what we're most of the people in the room here, we're trying to break that down into, well how can AI help us in what we do in our business lives? And that's a whole different question than what can it do for us in our personal lives. And I think looking at the technology from that perspective, it's a lot like social media. Social media in your personal life is one way of doing things. Social media from a business perspective, you have to have way more discipline, you have to have way more focus. It can be very powerful for you, but it can also be your worst enemy for sure. So I think there's a lot of that aspect of it going on right now where the hype is really taking over because it's worldwide hype and all industries and all application areas and we need to hone that down into for your specific purposes that you're looking at, what's most practical for you and what's the best way to implement that?

Justin Mack (22:13):

Absolutely, yeah, The mainstream hype can really have its impact. AI was an indie band, it went mainstream and sold out. AI is like green day and sometimes it doesn't help when the mainstream gets their hands on it. Scott, your thoughts on how do we break down the monolithic thought of AI and show that nuance a little more clearly?

Scott Lamont (22:33):

Yeah, I think when we talk to our clients about and about a piece of technology and it goes beyond ai, we get prospects or clients who walk in and they want a CRM. And when you ask them, well what are you going to do with A CRM? We get 10 different answers and we start to dig deeper and end up, we end getting to the point where they're really focused on business development and they want to be prospecting more. And so that's why they want to use a CRM versus managing client relationships and the value proposition that they offer. We spend a lot of time talking to clients about what are you trying to do with your technology? What makes you as an individual advisor, as an advisor shop as a team, as a portfolio manager, as a stock picker, as a financial planner, why are you different? Why are you unique? And then let's look at all of the technology and capabilities that are out there and find the right combination that accentuates that value. And I think that's the direction that we take with AI. Any other piece of technology is you, you're not going to just go and use AI because it's the cool new buzzword to use. What are you trying to do with the tool? Are you trying to engage with more clients or more prospects? Are you trying to build smarter portfolios? Are you looking at direct indexing and personalized portfolios? Can you use an artificial intelligence capability in that manner versus in writing performance reporting that tells a different story. So getting to the value proposition of the individual firm and the advisor. And then let's look at the specific capabilities that are available to you, the data sets that you have, the other tools that you're using, and find ways to create a whole ecosystem of tools and not just buy a cool new AI tool because everybody said this is the cool new tool to use.

Justin Mack (24:33):

Absolutely, and I would love your thoughts on this as someone who's running an RIA thinking about supporting your advisors with that tech, that growing team of advisors you have, I know the practical and actual's got to be the biggest thing for you. How are you guys tackling that and showing that difference?

Ritik Malhotra (24:47):

I think Mark and Scott touched on a lot of the things that we do as well basically for us is we always start with what is the end impact? So whether it's 80% reduction in compliance time or whatever else it is because that's actually the thing that resonates most. And then that obviously raises eyebrows and then they ask, Hey, well how's that possible? Then we'll kind of uncover, well here is the underlying technology, let us show you kind of step by step what it's actually doing to take that person on the same journey we went through when building that as well. And that's probably the most effective way that we found. And if you think about the history of the most successful companies out there, when you think about Google, they tell you, Hey, we're an index for the world's knowledge base. Effectively they don't talk about, hey, we're a big database that you can search using this technical page, page rank algorithm, the algorithm or the AI technology underneath is simply the tool we use. That shouldn't be the front and foremost kind of point that you lead with, but I think that's exactly what's happening in the kind of market today where everyone hears AI and they think that that's the problem they need to solve somehow. But that should just be more practically introduced where you lead with what the actual impact and problem is first.

Justin Mack (25:52):

For sure. Absolutely. Now something else I really would love to talk about is just what we can learn from other industries as we continue to take that maybe wait and see approach to AI to really have it used in new ways, different ways going forward in wealth management. And I'm a huge consumer electronics nerd. I always love looking outside of the industry to see what might be next. And I think about using VR as another example of an emerging technology that's going to be led by the gaming industry and the tourism industry and obviously those kind of experience industries before it comes around here and we find a way to use it, which it might be in trouble as Apple's rolling out $3,500 headsets, but that's a topic for another conference. But are you seeing anything outside of financial services or even outside of business at all that you see AI being used for and you're like, man, that's cool. We should get some of that. Scott, I'm actually going to start with you and catch you off guard. What are you seeing out there? And I'm interesting because I know F2's going to be incorporating that into consultancy and guidance going forward. Anything out there that's really grabbed you like, boy, that's awesome.

Scott Lamont (26:53):

Yeah, we always talk about with client experiences and digital experiences, the experience that you're designing for your client or for your advisor shouldn't just be informed by what's other wealth firms are doing. But your client is looking at your digital portal page and then they're on Starbucks or they're on Apple or they're on Google or they're somewhere else. So that's what they're comparing your pages to. And I think similarly in the AI area, understanding how it's being used, there was an interesting TED talk that a colleague of mine posts the other day about use of AI in education and this concern, significant concern that AI is going to help every kid cheat and get through their tests and no one's going to learn anything more. And the angle that they took was you can program AI to the point of it being a tool is how you deploy it and how you program it. And this, it was Khan Academy was programming it as looking toward moving to a one-to-one tutorial relationship between the student and this tool and to program the language or the chat G P T tool to coach the student through asking them a problem. And when they then say, can you just tell me the answer? The AI comes back with, well I can't just tell you the answer and let's go through the problem set and coach them through how to think about it. So I do think that you know, look at an example like that. You look at an example in the writing community where you've got this idea they were calling somebody called it sandwiching where you give AI the Chachi five ideas, it comes back with several different scripts, then you spend time adjusting it and then you feed it back into AI to modify it and adjust it and help finalize it. Can you do the same thing with planning? I've got these four ideas, come up with something, then put your own twist on it and then have it reviewed. So I think there's lots of ways that we can be thinking about using this that aren't these sort of daunting, scary Skynet scenarios that are out there. For

Justin Mack (29:12):

For sure. For Ritik, same question to you. What are you seeing outside of our industry? That seems pretty cool.

Ritik Malhotra (29:17):

I was thinking about a conversation I had a couple months ago on the question a few of friends and I were talking about was what would be the best application for AI kind of once the speculation goes away? And the answer we came up with is anything that has a deterministic or mostly deterministic output actually makes a lot of sense for the new generative AI or large language models. And so if we think about probably my top three kind of favorite outside of wealth management AI use cases I found are number one for software engineers to write code actually ends up being a very, very good use case because code is very deterministic. And number two, I think to a lesser extent counterintuitively in healthcare there's a subset of problems in medicine that are relatively deterministic If you go to the doctor that actually would be can you can actually see some studies where AI actually is producing equal or better outcomes just purely from an actual answer perspective than the top physicians. And that was also very fascinating. And then the third one that I think is a little bit out of this realm is an application called Whisper, which is effectively a transcription application. And it's shocking because back in the early two thousands there was this kind of wave of many, many kind of transcription softwares that were coming out and it was very difficult to make improvements. But if you look at Whisper today and the way that it uses artificial intelligence, it is near perfect transcription. You can have an extremely noisy room in someone talking and it'll get every single word down. And I think those are probably the three that have me the most excited.

Mark Evans (30:54):

So that's great. Awesome. And Mark, what are you seeing out there? Yeah, I would say the one that resonates with me the most is the medical domain. And I'll give you an example. Radiologists who are looking at CT scans or MRI images of potential like someone looking at their lungs or looking at their abdomen and looking for tumors and that if you've ever seen a CT scan of a lung, it is a very difficult path to determine whether or not there are nodules that could be cancerous there. Everybody has nodules in their lungs that are two, three millimeters in size. Anything bigger than that, okay, that's easy, that's obvious. But when you're in that range there, it's difficult. And to look at these scans and look at all the three D images of that, for a radiologist who has a throughput of all the scans that are happening, that's really difficult. Whereas that's a problem for AI where if you pass those scans through historically and you train the AI engine to actually be able to recognize these tumors, you're not replacing the radiologist, you're enhancing the radiologist as the radiologist is going to be doing this hour after hour after hour. You've got different levels of radiologist experience. Some have been doing it for 30 years, some have been doing it for five years. And if you can make that diagnosis more consistent worldwide, not just in the best hospitals in the world, but anywhere in the world, that's going to be an improvement for society. And I think that's the kind of application that if you look back to other business applications like in Wealth where you're saying can we have build financial plans that are equivalent to what the best advisors would produce, not bias, not based on what I just saw with the last client that came through the door and so forth. That's where I think, and same thing with portfolio construction and so forth is that's where I think you can learn from other domains and other industries and how to best practically apply these new technologies to improve mankind across the board.

Justin Mack (32:57):

Absolutely. And then kind of the last question I want to get to as we're here hitting our last seven minutes or so, and I will shout out to the audience saving some time for some Q and A, so we actually have some questions, think about them now and we'll try and get 'em answered. But the last question is always something we want to talk about when we're talking about bleeding edge technology, the regulators regulatory environment. It's highly regulated and I sometimes think about that when I see online people getting concerned when they see these examples of the negative sides of AI, Dr Crosby's video, yes, I'll kill all humans, all that kind of stuff. But one thing that I always think about is that this regulators aren't just going to let you use ai however the heck you want. We know good and well that SEC's not going to let you just do that federal's not just going to let you do that. So when we think about the regulatory framework that we have to work within, and Mark, I wanted to start with you on this, your thoughts, do you think that the regulators are going to really start to focus their attention on this technology, try to provide some guidance? Because one, I think it could be helpful without that framework as we continue to advance and do so in a very rapid way, we'll get to a point where we have potential AI solutions that might be well ahead of what regulators have in place as far as use in our industry. So your thoughts on, do you think you're going to start seeing a little more focus on AI from the higher ups there?

Mark Evans (34:13):

Well, I've already seen it with the organizations we deal with when they go, well how are you going to use ai? You're not going to generate the financial plan by sending it out to Chat GPT and getting the response back because right away the compliance goes, well, how did it actually generate that? We don't know what it did. And if it does something that's offside, we've got a big problem. So I think that there's going to be a lot of, particularly in our area where there is a lot of on the line and there's potential for lawsuits and so forth, that the compliance department is going to be very, very skeptical about where it's being used, how it's being used, how do we have a track, what it's doing and how it's doing it in order to have a comfort level that yes, what we're providing to the client is good advice and we can stand behind it and we can explain that advice.

Justin Mack (35:03):

Absolutely. Yeah. Scott, your thoughts on where you think we'll see or where we'll be going as far as regulation and then I guess how hard is it too for firms to chase these ideas when there might be a potential that something that they're really into is going to get swatted down as soon as they start getting around to thinking about implementing it? Thoughts on that?

Scott Lamont (35:21):

I'm not sure that I would be as concerned about that piece of it because I think the preventative part of using this is going to be more the internal trust of I put this information into the black box and I got something out. Do I really know for sure that that's the right answer? So even internally as an advisor or a consultant, and if I use it for my own research, I'm using it to get a bunch of information back quickly and then I'm probably going to go validate a fair amount of it. So I think the constraints will be there. I think my sense would be regulators, SEC and OCC and others will spend a lot of time kind of sitting and watching the use cases and understanding where it's being used, asking the questions of you're not just taking it and turning it right around and delivering it as the approved financial plan. We even were playing with it yesterday and it will tell you that it can't produce a financial plan that should go out. There should be that human engagement piece. So I think it will take some time. I think we're in this era right now of feeling out how it can be used and with each use case that comes forward, we'll review to see can it be monitored? Is the data visible? And clearly you sort of identifiable that you can prove out what it's suggesting and being comfortable that it still meets the regulatory requirements.

Justin Mack (36:55):

Absolutely. And I really want to ask you kind of a variation of this question just because I think about your role at Savvy and again, leading that team and providing and introducing new Tech to your team of advisors and what Scott just mentioned about that trust coming before we even get to the point where you've got to worry about regulatory frameworks as you're looking at new things to add to what you're offering, obviously AI based solutions. I know there has to be an upfront conversation or full buy-in from your advisors about they're going to have to use this. Do they trust it? Do they want to use it? In which way? How do they want to layer it into their personal practices right out front or maybe quietly in the background. So how are you approaching that going forward, that trust component?

Ritik Malhotra (37:33):

This is something we thought about even just before starting the firm, which was we would never force anything on someone. And it's a lot of it is show not tell. So we would always want to show the power of something and maybe have one or two kind of test cases first that really bring an advisor, whoever is using the software along for the journey they need to be bought in, not told that they need to use something. And if you kind of think about that, that's kind of transpired. And even talking about from a compliance perspective, which is the first thing that everyone thinks about is any type of technology, not just ai, what is it doing that I will it get me in trouble? So the other kind of philosophy we took upfront was nothing that is AI generated will automatically be put out. The way we always think about it is it, it's a reduction of time, but there's always a human element, whether it's someone on the compliance team, the advisors, someone else that reviews it and makes sure that it is good to go before they're sending it out. And I think the other thing we've found is that it actually can reduce the amount of errors and actually help reduce compliance errors as well. Because for example, if you have a call with a client summarizing that call or kind of transcribing the call to a certain degree, it actually might be easier for an AI to do that versus doing it for memory. And so those are just the three things we've talked here.

Justin Mack (38:53):

Awesome. And I wanted to open it up for any questions. If you do have a question, I think we have someone with a microphone who's able to facilitate a quick q and a if we have any questions at all or did we answer all of your questions in advance because we're just that darn good? I think we're just that darn good gentlemen. So I want to thank everyone for joining us today. Another round of applause for Mark, Ritik and Scott.