Episode 76: Artificial Intelligence, Efficiency, and Cybersecurity with Fi360’s John Faustino

Revamping Retirement Episode 76

Please note: This is an AI generated transcription. There may be slight grammatical errors, spelling errors and/or misinterpretation of words.

Intro: [00:00:00] Covering the ever evolving retirement plan landscape to help identify the biggest opportunities for plan sponsors, CAPTRUST presents Revamping Retirement.

Jennifer Doss: hello everyone and welcome to Revamping Retirement. I’m Jennifer Doss and I’m joined by my co-host Matt. Patrick. Matt, how are you?

Matt Patrick: Doing well, Jennifer. Good morning.

Jennifer Doss: Yep, we’re on a roll together, so let’s just keep it going. today Matt and I are joined by John Fino. he is head of FI 360. That’s actually literally what it says on the website, and I love it.

Jennifer Doss: It’s very simple to the point. John is responsible for strategy and leading day-to-day operations at FI 360. He’s been there since [00:01:00] 2016. and then prior to that, I think you spent some time at Morningstar and John is gonna spend time with us talking about AI in retirement as it relates to that topic.

Jennifer Doss: And if you guys are long-term listeners, which I know you all are, you know that we talked about this with someone from Morningstar, Mike Allen last year,And things are moving really quickly. Things are changing. An update. And John, you have a perspective on the technology world as it relates to retirement industry and so.

John Faustino: Thank you so much for having me, Jennifer and Matt, very happy to chat with you today. And one thing that I wanna share is that in addition to FI 360 a few months ago, I took on responsibilities for. The product and strategy functions within Broadridge’s data and analytics group. So FI 360 is part of Broadridge, so my mandate has broadened a little bit.

John Faustino: and it’s given me, a broader purview on AI as well too. So I’m excited to share that with you.

Matt Patrick: just to add some more content. could you walk through what is Broadridge? What is F 5 360? What do you all [00:02:00] do? Just so we have a sense for your background in the space.

John Faustino: Sure. So Broadridge is an s and p 500 company, financial technology company. We do a lot with. Proxies, actually, that’s how we started. We spun out at a DP 27 years ago, and we, do a lot on the backend. we’re one of those firms that’s connected to a lot of broker dealers and RIAs and asset managers, but you might not hear our name as often as you hear some of the other.

John Faustino: marquee names in the space. So we’ve got 15,000 employees worldwide and we do a lot to support asset managers, a lot to support issuers, and a lot to support financial advisors as well.

Matt Patrick: as we get into, the actual topic of conversation, which is ai, before we hop into anything business related, walk us through how do you use AI in your personal life? Are there any specific tools you like to use, any projects you’ve done with it? what’s your background there?

John Faustino: so far I think it’s been pretty vanilla for me with ai. at work,for the first time at Broadridge here, [00:03:00] since we’ve been part of Broadridge for about five years, when there’s been an interest in me writing an article, I. They’ve gotten a ghost writer and someone writes a first draft of an article, you spend a half an hour with them.

John Faustino: And then I look at it and inevitably, I rewrite everything. So one thing that I’ve used AI for is to create first drafts of articles. So it’s not good enough to get a final version for me. But if I give it guardrails, Hey, please write me something on this topic using, and I’ll give seven or eight prompts and say within 600 words or 800 words.

John Faustino: that works out pretty well. I also used it to create a corporate charter or a, charter for an industry org that we created. We’ve got a retirement income consortium and I put some prompts in and that did a really good job of putting some guardrails, and giving me something fairly standardized that I could then tweak.

Matt Patrick: It sounds like some good process efficiency or process improvement items thus far, particularly on the writing side, overall at Broadridge, [00:04:00] is there any sort of AI developments internally or anything you’re willing to share there in terms of your all’s view on, how AI is gonna impact the industry?

John Faustino: Sure. So we’re very invested in ai. We have, a couple of individuals that are focused solely on AI at Broadridge. It’s used throughout the organization, within the group that I manage. in particular, we have some things that are live. With AI already, primarily focused on the generative AI side of things.

John Faustino: And, I’d say we’re dipping our toes, in the water with things that are pragmatic, where we can create efficiencies, where we can save our clients money, out of the gate. But eventually we expect that it’s gonna be a really huge catalyst for personalized guidance, personalized advice for end investors.

Jennifer Doss: and John, just to level set with everybody because I think we throw out some of these terms and we expect that our plane sponsor clients know what we’re talking about. But when you say generative ai, what do you mean by [00:05:00] that? Is that the chat bt, that people think about?

John Faustino: Yeah. chat. GTP or other tools. So generative AI is essentially where you give. Artificial intelligence, some different prompts. So you give it a little bit of guidance and give it some guardrails and then it will generate some content for you. And one example of that, something that we’ve got designs on, if you think about plan sponsors in particular, the advisors that work with plan sponsors or the advisors that work with end investors, they spend a lot of time.

John Faustino: Developing the content that they’re gonna share on a quarterly basis or a semi-annual basis with their reviews. So the advisors that we work with,they’ll print out a monitoring report and then they’ll walk through that monitoring report and create notes in it and say, these are the funds that are doing well, these are the funds that aren’t doing well.

John Faustino: Participation’s up or it’s down. So one of the uses that we’re really excited about is to use generative AI to take all of the data associated with those monitoring reports and to create a first [00:06:00] draft of the commentary that an advisor will share with the plan sponsor. And we believe that’s gonna save a meaningful amount of the advisors time.

John Faustino: Allow them to serve more plan sponsors and also make it more profitable for firms so that they can actually go down market. So if you think about financial advisors working with plans or working with investors, there’s this relationship between the amount of time it takes to serve someone. And the size of the account, where the bigger the account, the more time you can afford to spend with them.

John Faustino: If we can lessen the time that it takes to serve each account, we can make it profitable to serve smaller clients and democratize access to advisors, which is what we’re really focused on.

Jennifer Doss: Yeah, I think that’s really interesting considering that there’s estimates that, an additional, 900,000 plans are gonna be created in the next, 10 years or some ridiculous number like that because of all the state plans and the push to, increase coverage, which is great.

Jennifer Doss: But then you go, okay, how are we gonna service these plans? And so things [00:07:00] like ai, like what you’re talking about, can really make that a more realistic goal to say, Hey, maybe we can just bring service, down market.

John Faustino: I totally agree. Again, it’s not only that there’s gonna be more plans, but the average number of participants expected in those plans is 15. So secure 2.0 in the state. Mandates are solving for these smaller organizations. And that’s where the scale is really important. So you can create an easy button for the advisors so they can profitably serve the plan sponsor.

John Faustino: So we’re excited at our opportunity to play a role in that and to be advisor, enabling, and to make sure that the human advisor still has a role in the process because. what we’ve learned, what we believe is that it’s really important to have that human still in the loop. it’s not realistic that these things are gonna be a hundred percent automated.

Matt Patrick: I think that point is, good to highlight just because when you talk about the impact of AI broadly, that’s most people think how am I gonna be replaced? Or How’s this gonna take out, jobs or, take over certain roles that I’m used to interacting with the human for. but then most [00:08:00] people that feels like in the industry, there’s some out there that feel like it’ll be that impactful, but most people I talk to now think it’ll be some sort of productivity aid, which seems to align with what your outlook is on.

Matt Patrick: Is that right?

John Faustino: Yeah, absolutely. I think it’s really gonna be a productivity enhancer. And we can look back a few years and get some guidance. From all the hype that we had on robo-advising in general, where people talked about, robos are gonna take over the world and there’s not gonna be any human financial advisors.

John Faustino: And I’m actually,tomorrow night is my last class in a doctorate program that I’m taking. So I’ve still gotta write my practicum, but I’m gonna be done with my classes as of tomorrow. And I’ve studied this quite a bit with robo-advising in particular. the roboadvisors tended to give really good guidance, relative to what the individual investors would select on their own.

John Faustino: The problem was there was a lack of receptivity from the human investors to take this advice from their computer. [00:09:00] So there’s a need to insert a human touch and a little bit of human context. In the delivery of advice. And, I think robo-advisors not taking off like people expected is a good clue in terms of how it’s gonna be incorporated in the context of plan sponsors and retirement plans.

John Faustino: And one other thing that I’ll share is. An area where AI has really taken off and been very successful already is in pharmaceuticals and specifically drug development. And there’s some neat research that has shown the three things that combine to make AI successful with drug development is when you’ve got consolidated data.

John Faustino: So you need to,have all your data in one place, first and foremost. But then you need human subject matter experts that can guide what you do with that data. And then you need to have some AI technology acumen as well too. So it’s really those three legs of the stool that make it happen. And that human subject matter expertise is a critical one.

John Faustino: And I [00:10:00] believe that’s,very much translatable to financial services and retirement plans as well. So that’s what we’re focused on, is taking our expertise. With how advisors work with plan sponsors and trying to create automation to make, everything more efficient and more effective.

Jennifer Doss: Well, I think there’s something to that in terms of. Maybe there’s a level that robo-advisors can give pretty good advice at, I’m just starting out. I’m 20 years old. It’s pretty clear what I should be doing. I don’t have a lot of money. I should be investing aggressively.

Jennifer Doss: There’s just not a lot of variability. And then we talk about as you get closer or you build. assets, then all these different variables come into your life. And that is where a robo-advisor or something without that human overlay is just not gonna understand all those different facets.

Jennifer Doss: and so I think that’s key too, which is, hey, maybe there’s a segment, but it’s not the majority and it’s not the important majority. So I think that’s a good point.

John Faustino: Yeah, very much aligned with you on that, Jennifer. And that’s where we’re having a lot of conversations with advisors that are working with plan sponsors on [00:11:00] getting more personalized advice to their participants and exactly what you described, a more of a stratified model where for some individuals that have less assets or a little bit younger, there’s things that can be delivered via email or via a chat bot that might be received in a fine manner.

John Faustino: And then as the needs get a little bit more sophisticated, you introduce. The human element in a way that helps you customize things a little bit for people without making it, unprofitable and just untenable to give that personalized advice to every single employee.

Jennifer Doss: the last point I would make on this, because this is very timely, I don’t know where we’ll be when this post, but we’re in some market volatility right now and we have been getting a lot of calls from participants. And it doesn’t matter if they’re 20 years old, it doesn’t matter if they’re 50 years old, they’re concerned and they wanna talk to a person about their concerns.

Jennifer Doss: And so that’s also a reality that you always need to provide that human touch when things go. Bump in the night. You need to have a person you wanna talk to, If things are going [00:12:00] well, maybe you’re fine, but if things are going, not well then that’s when you want that human reassurance.

Jennifer Doss: I, I believe is what we’ve seen

John Faustino: Absolutely. Staying the course is one of the greatest benefits that a financial advisor brings to their clients, whether an individual investor in wealth account or a plan participant, and we saw the trading activity, so another part of Broadridge. Is, matrix. So we own a custodian that works with many record keepers and the trading in 401k plans over the last week or so when there was a lot of

John Faustino: Volatility in the market was 10 x what it normally is. So,we saw a lot of individuals trade out of their, equities, go into a money market or stable value fund. At, probably a pretty bad time. we’ll know in a few months. but if you trade in and out, we know that’s generally, not a good thing.

John Faustino: Individuals aren’t good at time in the market.

Matt Patrick: I’d love to pick a little more at the education piece that you mentioned earlier, and then you need the human element in there. [00:13:00] So what part of education or advice to participants in retirement plans do you think the AI can. actually help the most, and where do we think we’ll see it most impacting that?

John Faustino: I think there’s a lot of opportunities and one area that I’ll share. We saw during Covid, so I’m going back a few years now. But during Covid, there were a lot of participants that lowered their contribution rates in their 401k plans. maybe they weren’t getting the same, sales commission or they were extra concerns.

John Faustino: So They backed off of their contributions for a little while. So one of the things that AI is really good at is identifying cohorts of individuals. there’s some tools that we’re working with where we can identify the individuals that aren’t maximizing their contributions and we can send them a note, to remind them, Hey, maybe you turned this down last year or a couple years ago.

John Faustino: It might be time to turn it up now. And there’s other things that you can do with. Artificial intelligence [00:14:00] to gamify things a little bit with the participants and increase their engagement. So sending messages, in that same context, perhaps on increasing contribution. Saying that right now you’ve got a. let’s say a 70% chance of hitting your retirement goal, if you were to fully contribute, you can bump that 70% up to 85%. we’ve seen really good success based on how people interact with their smartphones and their smart watches, that they want some of those nudges. that’s one of the things that we’re looking at.

Matt Patrick: you talking about, the AI analyzing different cohorts within a producing population I’d imagine that could lead to some pretty interesting. Problem solving at the plan sponsor level in terms of identifying ways to better engage certain groups, or specifically targeting certain groups.

Matt Patrick: Now it’s a lot of just Hey, we’re gonna send everything out and mass to everybody and, hope it works, or hopes it gets people more engaged. But if it could start to. Tailor some of that specifically to certain groups, just in terms of what kind of messaging works well, what seems [00:15:00] to drive positive behavior in certain areas.

Matt Patrick: Maybe we can get a little more refined and actually get people engaged in the plans in the way that we want.

John Faustino: I think that’s a great point, Matt, and we know that If we spam people with messages, if there’s, 10 messages that we wanna get out to an entire population, if you hit people with messages that don’t resonate with them, you run the risk of them unsubscribing or just not paying attention to the one that might really be a need for them.

John Faustino: And another area that I’ll mention that’s, not specifically retirement related, but one of my professors. Has done some interesting research on emergency savings and the impact that has on people’s, financial precarity is what they call it. So there are concerns about retirement and the like, and it can create, anxiety at work that’s really detrimental.

John Faustino: and this particular study that my professor did talked about a trucking firm and the individuals that had an emergency savings account. Got in less accidents than the drivers that didn’t have [00:16:00] it. So you could literally justify funding these emergency savings plans based on the lower insurance premiums that you’d pay.

John Faustino: so using,big data to identify those relationships and actually, connecting. Employee behaviors to company costs, are things that,the plan sponsors that we’re connected with are really excited about It’s extending beyond just the mutual funds and the contribution rates, but getting into broader corporate expenses.

Jennifer Doss: that could have some interesting applications on the healthcare side too, Where you’re looking at what people are consuming and how they’re consuming it. And the more data you have that the more you can offer them, that really speaks to, getting them to actually do things that you want them to do.

Jennifer Doss: so we’ve hit on a few things that we think will change and will get better. Personalization, communications, using the data to better drive behavior. You talked about one thing that AI probably won’t do, which is replace human advisors, and we talked about why.

Jennifer Doss: Is there anything else you think people are hyping up about ai? that it’s just not gonna do It’s just [00:17:00] not realistic

John Faustino: I think the, the big thing really is that it won’t be able to replace, the human factor. I’m also, not concerned so much with. AI taking over the world

John Faustino: the gentleman that runs, AI for Fidelity was at a conference that I was at a couple of weeks ago, and he is like, Hey, the computing power to do that,

John Faustino: It’s not realistic. So I believe that,there’s appropriate guardrails in place and there’s just computing capacity challenges that make that very unrealistic. And one of the negative things of the Skynet,hype is that it’s getting some people to back off from some of these pragmatic solutions that are gonna save you some time operationally.

John Faustino: and one other thing that I’ll share. That I think is hyped up a little bit too much, which is, the people that go on the other side of it and say, Hey, you need to embrace ai. You need to do all this stuff right now, or else you’re gonna get way behind your competitors. The reality is there’s gonna be a set of, vendors, service providers out there that are going to enable others to [00:18:00] consume ai.

John Faustino: I don’t think every company is gonna have to have their own in-house ai. Technology experts at a super deep level. I believe that this is great example of democratization and where, vendors are gonna be able to get people access to things. And I think chat gpt is a great example where,for free you can get in there and you can play around and you can use things that can actually make you more efficient.

John Faustino: on both sides of the continuum,there’s been some hype that isn’t gonna end up materializing.

Jennifer Doss: Matt and I were opining on the dawn of the internet and Google and things like that and how there’s some similarities here where people thought. Oh, I can’t trust anything it says. I can’t even use it as a source in my paper. my teacher won’t accept it. and now here we are.

Jennifer Doss: And you can’t imagine doing your job without it. I can’t imagine actually picking myself up and going to be like, Matt, I’m gonna go to the library. I’ll be back in a few hours, I’m gonna go research some stuff. it’s out of a history book. But, there’s some similarities too,

Jennifer Doss: Where people think that this is a huge deal, but it’s, history rhymes, So I think there’s a lot of similarities there.

John Faustino: [00:19:00] starting to look a little further out, timeline wise, the things we’ve addressed Making participant advice more scalable. There’s some efficiencies on the backend from a data perspective for those in the industry, hopefully aiding in better client support.

Matt Patrick: What do you think long-term the impact of AI looks like across the retirement industry?

John Faustino: the personalization is gonna be a big impact. and one thing that I’ll share is that in addition to the benefits that human advisors give, that Jennifer mentioned to. Having people stay the course, when the markets go down. Another thing that helps people stay the course is when they have personalized portfolios.

John Faustino: And we saw this during covid with those folks that had, for example, ESG investments, whether you’re pro ESG or anti ESG, if you know that you’re stacked towards it or against it, the individuals that had those personalized portfolios tended to hold, more so than those that didn’t.

John Faustino: So in the long run. I’m really excited at AI creating [00:20:00] personalization that’s gonna connect end investors more directly with what they’re holding and just make them, more secure. there’s a couple of ways that it’s gonna create some positive I. Behavioral impacts and the iterative learning is something that I’m really excited about as well too.

John Faustino: It’s not a fixed in point,scenario, but the improvement that I’ve seen in a short time with some of the AI tools I’ve used is really great, which makes me feel very good about refinement. it’s not perfect today. It’s, not gonna be perfect a year from now, or, 20 years from now.

John Faustino: But, we’re getting to a point where we’re gonna make investment outcomes better for planned participants based on our leverage of technology once we’ve got the right guardrails in place. And we’re confident that the hallucinations aren’t worse than the benefits we’re getting from them.

Matt Patrick: highlighting some of that stuff is really important because at least the way that I hear. AI efficiency talked about Sometimes if I was sitting in the plant sponsored chair, it’d be like, [00:21:00] what does it matter to me if writing your emails goes faster, I want to hear from you and, some of the elements that get highlighted, like the.

Matt Patrick: Robo-advisor feel that,we already acknowledged just didn’t resonate with people, didn’t take off. But you could better identify campaigns or solutions at an individual level or at a small group level within your plan.

Matt Patrick: Or you could start to personalize the portfolio or personalize the experience in a way for participants that would help them feel more engaged and have it feel more like a high touch. Wealth service or high net worth advisory service within your plan that registers with people? I think those are the benefits and those are the goals.

Matt Patrick: And it’s imperative on us in the industry to not lose sight of that. and, as we make things more efficient and be like, Hey, We’re more efficient on the back end, so then we can be more forward facing and, more service oriented on the front end actually interacting with plan sponsors.

Matt Patrick: But, it’s nice to highlight some of these that feel a little more tangible from a plan sponsor end, at least for me.

John Faustino: you know, I mentioned before the human element of it. I think the process is really important. So how you get from. Where we are today to the [00:22:00] more automated place, and one of the things that we’re doing is we’re eating our own cooking. So instead of saying, Hey, we’re gonna use AI and we’re gonna automate all of our emails to you, we’re having, in some instances our support team is using it and they’ll use it to do an initial draft of an email or to do research, and then they’ll check it to make sure that it’s correct and pass it along.

John Faustino: So it’s saving us time from a support perspective. But you still have a human that’s meticulously going through things at the beginning stages. And as we get more comfort with, for these use cases, it’s right all the time, then we’ll start to have some more direct connection with our, end users.

John Faustino: But we’re inserting a little bit of a safety net in the interim, because we know that it’s not perfect and, we wanna be thoughtful about the use cases that we expose our clients to directly versus those that we use. More internally for us,serving them more efficiently.

Jennifer Doss: I think the safety net, to me, that guardrails a couple of times too. having [00:23:00] the right governance structure in place is gonna be important for. Everybody’s gonna use AI to some degree within their organizations. Either, like you said, they’re gonna engage with other companies to facilitate that or they’re gonna do it themselves in-house and this all ties into cybersecurity

Jennifer Doss: You have to understand how they’re using the data, how they’re using these programs. so talk to us about. any notable impacts of AI on cybersecurity issues. I think there’s all these positive aspects of what we talked about, but there’s also some things to be aware of. and, one of the things I always think about, which is, you can make these videos that make it look like, Matt sent me a video and I’ve worked with Matt for, 10 years and I might not know the difference.

Jennifer Doss: and so how do we think about, enabling cybersecurity and protecting people in this.

John Faustino: Sure. Yeah, and I’m, I am very concerned about that and I’ll share. There that our neighbors have, an aunt who’s elderly and there was someone, and I’ve seen some of those videos, if you look really close, you might be able to tell that it’s a video and it’s not the individual. it’s even tougher if it’s just a phone call.[00:24:00]

John Faustino: And we’ve had, a family member of one of our friends who got a call and it said, Hey, this is your grandson calling and I’m in jail. And, You gotta post the bond in cash and these people got scammed outta $5,000 cash, so they met someone and handed it off. there are aspects of AI that can make it easy for this, socially engineered, crime.

John Faustino: And that is a concern for me. And I think that’s why when you think about who you’re partnering with. AI where you’re putting your plan participant data, where you’re putting your advisor data, you need to really prioritize the cybersecurity capabilities of whoever is storing that information for you.

John Faustino: and I think there’s a lot of things on the best practices front. get all these spam robocall. don’t talk,wait for the other person to talk. Don’t let them get your voice. and maybe educating some of your elderly family members and friends as well too, on these things.

John Faustino: So I get really offended by, especially elderly people being taken advantage of. [00:25:00] And that’s something that I think there’s likely gonna be a lot of malfeasance with, and we’re all gonna have to be vigilant and do what we can to help people out.

Jennifer Doss: there was a brief period of time years ago where. One of the big innovations in cybersecurity for a call center was voice recognition. we’ll, recognize your voice and that’s how we’ll know that it’s you. it was so cool and hyped and that was very short lived because obviously you can’t do that anymore.

Jennifer Doss: And so the industry has to keep evolving and thinking about ways to solve for these bad actors. and ai just makes that faster,

John Faustino: And there’s, some annoyances that come with it. Like the multifactor authentication. If I log in and then I gotta get a text or I have to get an email. but in the big picture, that’s where we’re at. That’s what we need to do.

Matt Patrick: You were heading in this direction, John, a little bit about being thoughtful about where your data is stored and all that. on the cybersecurity front, but as more parties start to use AI in some capacity within their business, how should plan sponsors be thinking about evaluating, record keepers, advisors, asset managers on how they’re using [00:26:00] ai, to make sure that.

Matt Patrick: Any advice or information that they’re getting out of those tools on the backend is reliable and something you could depend on.

John Faustino: That’s a great question. we’ve heard this with record keepers that a lot of them with RFPs now they’re getting questions about cybersecurity certification. So that is something that I would ask about. I would ask about,their processes for encrypting data.

John Faustino: There’s standard sets of. Criteria that the Department of Labor has put out, associated with record keepers, things you should be thinking about, encryption of the data at rest, encryption of it in transit. and that’s something that I’m very happy to provide you. we do some cybersecurity audits.

John Faustino: For, record keepers, and I’m very happy to make that available. Let people know, Hey, these are the things that we do for record keepers and for other financial services firms, if you wanna share that with any of your clients. it’s not behind a wall. we’re very public about these are the things you should be doing, and the regulators are very public about it.

John Faustino: So if [00:27:00] you’d like a checklist, that’s something that we can absolutely provide.

Jennifer Doss: we’ve been talking a little bit about this. mentioned the. time it took to, get away from that voice recognition, innovation that we had. So how do you think AI is gonna shorten that time of innovation? Somebody comes out with a great new idea and then, two years later it’s obsolete now we just seems to be moving faster.

Jennifer Doss: and I could say that about a lot of things in the world, but, how do you think AI has shortened that time to new products, new innovation, that plan sponsors should think about?

John Faustino: the area, I’m gonna go back to drug development, where it’s been really powerful is when there’s huge number of combinations that different chemical elements can have together. And I actually interned for Pfizer back in 1998 after I did a degree.

John Faustino: and they were just starting to use machine learning, which is a subset of artificial. Intelligence and they were able to test all of these potential combinations, and figure out, okay, this 95% won’t work. Absolutely. So then we only have to test this 5%. So there’s been a huge decrease [00:28:00] in the time that it takes to develop new drug compounds.

John Faustino: And for me anyway, it’s easier to understand how that’s the case, From a retirement plan sponsor perspective, I would think about the advisor that you’re working with and the asset managers that you’re working with, and the, related ways that they’re thinking about new, identification of investments.

John Faustino: We’re seeing a lot of interest in alternatives now, and, it’s very reasonable that you can think about artificial intelligence to. Advance some of the scenario analysis to get people comfortable with using alternatives and what the likely impacts are. we can model large data sets in a much shorter amount of time than we did before.

John Faustino: And then one other area that’s very much related to, plan sponsors and to innovation that’s likely gonna have some acceleration from this is retirement income. So 30 years ago, 40 years ago, most folks had a [00:29:00] defined benefit plan instead of a defined contribution plan. there’s, pros and cons of each, but increasingly we’re seeing employees have an interest and some kind of certainty with their income.

John Faustino: But the problem with that is, personalization and, I think ai, as you can bring in more data on an individual participant, you’re gonna increase the likelihood that you’re aligning them with a retirement income solution that’s best for that individual employee. So those are a couple of areas where I see it, likely to make an impact,for your plan sponsored clients in the not so distant future.

Matt Patrick: you jumping into retirement income. I was just thinking that as you were talking about, areas that it could improve. ’cause I do think there’s a challenge with that now where it’s like people. In general, they like the idea, they hear it in concept, but trying to pick a solution that fits for the entire plan, it feels like an impossible task.

Matt Patrick: And so you end up doing nothing. So that idea of could we help personalize it or build in some flexibility there based on individual situation, I could see that being super [00:30:00] impactful.

John Faustino: I sit on the board of a FinTech called Resa, and what they’ve done is they’ve identified,a style box. You know, Morningstar is a style box with, value and blend and growth and large cap and mid cap and small cap, and they’ve done something similar. For retirement income, and they have people take these surveys and one of the things that they’re working on is, predictive retirement income style so that you can identify this is what this individual’s, style is.

John Faustino: But to your point, Matt, there’s been some trepidation with retirement income because inevitably when those analyses are run on an entire pool of participants. Maybe 70% fall into one group or 60%, but it’s not a hundred percent. so the challenges with getting things exactly right is slowing down that adoption for sure.

Jennifer Doss: one of my favorite sayings is, when you solve for averages, you solve for nobody. you solve for absolutely nobody.

Jennifer Doss: I think the more that we can personalize, investments, advice, all these things we’ve been talking [00:31:00] about, the more likely we are to get people to engage, the more likely we are to give good advice and more likely we are to give impactful advice.

Jennifer Doss: And I, I think that, AI is certainly gonna aid in that push.

Matt Patrick: one more question, John, on some of the challenges working with ai. So you mentioned earlier I. Hallucinations, in there, which I don’t know that we, exactly defined. So could you define that and then how do you personally think about trying to tackle some of the challenges that can come from taking information from some of these AI tools that are out there?

John Faustino: So those are real challenges. And just for everyone listening, if you’re not familiar, the hallucinations are when you make a request for something to come out of AI and what comes out is wrong. and it can come out wrong for a few reasons. You can have a. Bad data. And if you think about AI going out to the internet, if there’s bad data out on the internet, it might pull that bad piece of information.

John Faustino: I actually did that,when Google came out, I’d Google myself and I, chat gtd myself and said, Hey, what’s John Tino’s role? And it said a wrong role. So there can [00:32:00] be bad. Data that’s referenced.

John Faustino: You can also do something called over training the models. you need a set of data to train the AI on how to learn. And if you use data that’s very specific, it can always point to a certain answer that might be wrong. So that is a real challenge, and that’s where,the guardrails and having a human sit in between those requests and the output, especially in the early phases of working with it, are really critical.

John Faustino: So we need to get to a point where we feel very confident in the output. and I would say similar to what, Jennifer mentioned with a stratified approach to giving advice to individual. Plan participants. So those that have very generalized needs, you can send an email to those that need more personalized information, have an advisor talk to them.

John Faustino: If there’s something that’s fairly benign,How do I log into this website? that’s something you might be okay, letting people access and just on an automated basis [00:33:00] get on their own. But if it’s, what’s the right asset allocation for me? Which investments should I buy?

John Faustino: Those things you better be pretty darn certain about. that’s one of the things we think about is stratifying the human intervention that we put in place. Based on the nature of the request being made based on the nature of harm, if it’s a wrong answer. the concern about hallucinations,bad responses from AI is real, but there’s ways that you can address that so that it’s not, overwhelming.

John Faustino: For me anyway, it always comes back to that human, subject matter expertise. So having that trusted, advisor from CAPTRUST or another firm that understands these are the implications to you as a plan sponsor from a fiduciary liability perspective, from a participant, outcome perspective.

John Faustino: That’s why it’s so important that you have these humans that sit in between and help you arbitrate where you apply things and where you let ’em fly.

Matt Patrick: the hard part is when the tool’s wrong, it’s very confidently wrong and is very convincing in the way that it is. if I [00:34:00] chat gt myself about what my job as, and it was like, yeah, I guess I do make ice cream. That’s what it says right here.

Matt Patrick: I can’t, convince it otherwise.

John Faustino: I think that’s an important piece too, is What kind of data is it? Is it open to everything? Where is it getting its sources? How is it prioritizing those sources as well? is it giving prioritization to, matt patrick.com or is it giving a priority to, a government website or something like that?

Jennifer Doss: So I think those are all really interesting things to dig into. Um,John, we’re gonna move on from the AI topic onto the really hard topic, and that is. The topic of, retirement for you. we like to ask all our guests, what does retirement look like to you, John Fino.

John Faustino: thank you for asking and I’ll say I really love my job. I love supporting financial advisors, and I’ve been doing this for, most of my 30 plus year career. So I don’t have any designs on finishing up soon, and I’m very happy with Broadridge the place that I work right now. I did mention that I’m in school right now.

John Faustino: So I’m a nerd. I’ve gone back to school four times. my wife has told me this is my [00:35:00] last degree. and one of the things that I’m contemplating when I’m done with the corporate world is teaching a little bit, maybe writing a little bit. and my focus is on fiduciary advising and, retirement plans and IRA rollover.

John Faustino: I’m an awful golfer. I got a lot of buddies that are really good at golf, but, I’m interested in continuing to do something to stay active,with my mind when I’m done working. And I love southwest Michigan. where I’m at today. So being outdoors maybe a little bit more, once I’m done working is gonna be a big focus of mine and my wife’s as well.

Matt Patrick: It sounds like we’ve got some AI tools coming in play to draft some teaching agendas for you.

John Faustino: There we go.

Matt Patrick: so look forward to that. Awesome. we are, we’re gonna wrap up. John, thank you so much for joining us today. Really enjoyed the conversation. thank you to everyone that tuned into this episode.

Matt Patrick: if you’re enjoying the show, we hope that you will subscribe wherever you are getting your podcasts, and we will see you around next time. Thank you.

The discussions and opinions expressed in this podcast are those of the speaker and are [00:36:00] subject to change without notice. This podcast is intended to be informational only. Nothing in this podcast constitutes a solicitation, investment advice, or recommendation to invest in any securities. CAPTRUST Financial Advisors is an investment advisor registered under the Investment Advisors Act of 1940. This presentation does not contain legal, investment, or tax advice.