Execution to Strategy: Redefining Work in the Age of AI

[Music] Hi everyone, welcome to third Thursdays. We're so happy that you joined us. My name is Kelly and I'm going to be facilitating today's discussion. And we're excited. We've got a great topic and a great speaker. At Women in Retirement, we focus on five key areas, leadership, marketing, practice management, personal growth and advocacy. This is a place where we all come together to learn and grow. We're a community of actuaries, service providers, human resource, and financial professionals, just to name a few. There's a wealth of knowledge in this virtual room and we'll be opening the floor so we can hear from you. We encourage you to use the chat feature to engage with our other attendees or to ask a question. To ask a live question, we just ask that you raise your virtual hand and come off mute at the signal time. I want to remind everyone in order to receive your continuing education credit, you must stay on for the duration of the event. Please make sure you join with your first and last name so that we can properly assign your CE credit. If your first and last name are not visible on your screen right now, please make sure you change that. After today's presentation, we're going to move to some smaller group discussions and some breakout rooms. We encourage you to stay on so you can connect with your fellow professionals and you never know, find your next mentor. Today's topic is from Execution to Strategy. Redefining work in the age of AI. This session will explore the return on investment of AI for retirement professionals and how it's redefining task management and client focus. Again, we're so glad you're here and we want to start first by setting the stage with a couple of poll questions for you all our audience. The first question is on the screen. Has your firm started using AI for anything? You all just take a minute and answer yes or no? We'll give it a few minutes for our results. Okay, 90% yes and 10% no. That's awesome. Fantastic. All right, we've got another one. Are you concerned about AI replacing your job? Okay, 26% yes and 74% no. That's encouraging. All right, awesome. Well, I would like to introduce our speaker. Her name is Rachel Long. Rachel leads practice management at Zox and that is an AI assistant for financial provide, excuse me, for financial advisors. She specializes in helping large firms navigate the transition to AI driven operations at Zox. She partners with home office executives to design scalable one to many programs that turn AI from a buzz word into a core firmwide strategy with over 10 years in financial services, including wealth management and lending. She provides critical insights into advisor processes and client life cycles. Please join me in giving a warm welcome to Rachel. Rachel the floor is yours. Awesome, thank you so much for having me. I'm hoping everyone can see the slide deck. Okay. But yes, that was a fantastic introduction. Thank you so much, Kelly. So what Kelly said, I am leading the practice management team at Zox today, where we do focus on building AI specifically for financial advisors. This topic is something I definitely live and breathe every day. I was super excited and honored to be here. So let's dive into it. I'm going to do that. All right, so at this point in time before we really dive further into tactics and tools, I do want to start with some context because I think a lot of the reason many people may be underestimate what's happening with AI is that they honestly are measuring it on the wrong timeline. And what I mean by that is we're not talking about a software update or productivity trend. We're at an actual inflection point. The kind that really only comes along a few times in a generation. The internet changed how clients access to information, right? Mobile changed how advisors began to communicate and AI is changing what gets done by humans at all. So taking a step back further, even to what I'll call our middle school learnings, there have been several core transformations across humans historically. I think about the printing press during the Renaissance, literally going from handwriting, even hand painting books to allowing for mass distribution of ideas. And I think we all learned about the steam engine in around fifth or sixth grade, allowing for mass production of goods. The planet has seen groundbreaking innovations and have constantly adapted to the next flow of ideas that are generated because of it. So today's session, I'm really excited. I want to cover three things where AI actually is right now, where it's heading and what does that mean for any retirement plan professionals and those in the industry today. My goal is up by the end of this session. You can walk away with a clear mental model, not hype, not fear, but just grounded with what's in front of you. All right, sorry, I just check in the comments. So why does this moment feel different? Literally everything is AI right now. It's in the news. It's in your news feed. So what's actually changed? Thinking back another quick history lesson is AI is not new. We know this. What's different now is the speed of the actual progression in 2020 chat GPT 3 came out and it was super impressive in a party trick kind of way. It could finish your sentences, could write a decent paragraph. It could come up with some really funny limericks if you wanted it to. Think of it as like a capable elementary school age student, right? They're enthusiastic. They're sometimes right. They often need supervision. We were actually just talking about my daughter graduating kindergarten, so that makes complete sense there. But by 2023, chat GPT 4 and others began to cross their threshold that honestly most researchers didn't expect to hit so quickly. AI could start to reason through complex problems. It could pass exams. It could draft dynamic documents. A smart high schooler. Someone you'd actually trucks with real work to an extent. And then fast forward to today's models. They are now performing at an undergraduate level. They can handle multi-step workflows. They can reason across context and increasingly take action across systems and not just answer questions. But here's the part that matters for you when you're planning ahead for your firm. The progression alone is not slowing down. The gap between each generation of AI is getting shorter and not longer. What feels like advanced AI today honestly feels like a baseline 18 months from now. The firms that are building habits and workflows around AI now will honestly have a compounding advantage over those waiting for things to settle down. So, want to give a quick, I'm going to say 60 second explanation might be two minutes of what AI actually is. Because when most people say AI in this context, they typically mean a large language model. Think of your chat GPT's, clawed, copilot, Gemini. Those are all public AI's. That's most of the technology platforms you see across the marketplace today, especially when you're evaluating different vendors for your firms. But here's the simple version of what's happening under the hood. All AI is math, specifically statistics. Think of a massive amount of information, essentially the entire internet, broken down into chunks that we call tokens. These tokens go into what's called an encoder. And the encoder is essentially pattern matching at an enormous scale. When someone asks for X in this context, the answer tends to look like why all of those go through what we call the neural network, the brain in the middle. And it comes out the other side as a response that we call the decoder. These large language models are essentially predicting. They're very sophisticated, statistically informed, but predicting. They're predicting what you want to hear based on patterns and data they've already consumed. And so what makes them better over time is the data. These models want to consume as much data as possible. They've already crawled through the internet. They're now reviewing videos, audios, self-driving car sensors, anything they can learn from patterns. So when you type into chat GPT and say, write me a bedtime story about a unicorn, the model isn't guessing. It's producing what it statistically believes a bedtime story about a unicorn looks like based on every bedtime story in every unicorn reference it's ever seen. So two things I want you to take away from this slide. When we say it's AI, it always means in this context large language model. Not necessarily like if you've seen the Will Smith movie about the robots taking over the world, we're going to talk about the language models today. And then second, these models run on two things, data and energy. And that is why compliance teams evaluate, you know, overall concerns about which AI tools your firm should use. If a vendor is using your client data to actually train their models, that data is now part of their system. It's being consumed, it's being pattern matched, and potentially surfaced in other responses. That's the risk. And it's exactly why choosing the right AI for a wealth management firm isn't just a feature decision. It's an entire trust decision. So now that we know what AI is, I do want to spend a moment focusing on the difference between AI and automation. Because the distinction can trip some people up, and if you get it wrong, you could be evaluating the wrong tools. You could be measuring the wrong outcomes. So automation, this is, you know, something most of your firms already are using today. It's essentially a rules-based workflow. If this happens, do that. So a couple examples. If it's the client's birthday, send an email, send a quarterly report. If it's the first day of the month, trigger a re-enrollment reminder, maybe when auto-enrollment lapses, it's valuable. But it has a hard ceiling. It can only do what you explicitly tell it to do. And the second you try to change the conditions, it breaks. Examples like this could be, maybe you need your report on a different day of the week. Or maybe you need to utilize different data to pull that report. You have to make changes to the automation to allow new scenarios to come through. And this is what essentially can add to your firm's tech debt. Tech debt is the maintenance required to keep technology processes moving and operating as expected. AI is fundamentally different than that. It reads the context. It understands what you're trying to accomplish. And it's not just understanding the button you pressed. The same AI that drafts a plan review for maybe a large manufacturing company with a very complex history, will write a completely different summary for a small nonprofit with a 403B and a high turnover. You don't have to build two separate workflows. AI is going to allow for the dynamics to change. So think of automation as more of an efficient conveyor belt, where AI is going to be a thoughtful partner who is continuously reading the room. And for retirement plan professionals specifically, your plans are never identical. Every sponsor has different demographics, investment menus, fiduciary histories, rules-based automation may not be able to handle that nuance, but AI can. So when people ask about the ROI of AI, we usually think about time savings. And yes, that's 1,000% real. I love talking about time savings. But I'd also argue it's actually kind of the least interesting part of the story. The entry point is definitely efficiency. A plan review report that used to take hours to assemble can now be drafted in minutes. Committee meeting minutes that used to sit for days can get summarized and sent the same day. You can multiply that across 40, 50 different plans. And the hours recovered are significant, but we're just getting started. The more interesting layer is when you start considering your capacity. The recovered time means that you can either manage more plans, maybe without adding headcount or with adding strategic headcount. Or maybe it's even something like your weekends back, the trips you haven't taken, the lunch breaks you didn't get to. And then finally the third layer, the leverage. This is where the real transformation begins. When AI can flag that a plan participation rate maybe has dropped a year over a year, it can identify the demographic segment, driving it. You're no longer just reporting on it. You're being proactive with your advising. That's the difference between being an vendor and being a full fiduciary partner. Your team can be proactive. There's a huge layer of value that just doesn't get talked about enough with AI. And that's what AI surfaces within your own practice. When every meeting is documented consistently, when every follow-up is logged, pattern starts to emerge. And so what that means is you now know who on your team has strong sponsor relationships, where our handoffs may be breaking down, which questions are the best questions to ask in the meeting? AI can essentially give practice leaders a mirror that they've never had before. Not to micromanage, but to actually coach, and to replicate what's working, to finally have the data to back up what your team was already assuming was happening in the first place. And it works both ways. You actually start to see what your clients are talking about. Not what you assume they want, but what's actually being brought up in their overall sentiment. Which topics generate the most follow-up questions? What are clients saying about the market in the industry today? Which sponsors are quietly disengaged versus deeply invested in your conversations? That kind of intelligence used to require, honestly sending out a survey and waiting for it to come back and doing some analysis. But now it lives in your meeting data and in your processes with AI. The question I'd encourage firms to honestly ask, isn't how much time do we save? It's what becomes possible when you have that capacity back. Now, efficiency, capacity, leverage, they're all part of the ROI. But that doesn't mean that everything is fixed and performing at an optimum functionality. This is probably one of the most important conceptual shifts that I've honestly dove into much further the last couple of months, and I just want to talk through it overall. So for decades, eliminating factor in wealth management, retirement practices, the industry in general, was execution capacity. There's only so many hours in the day. And if you wanted to grow, you needed to hire, if you served more plans, you needed more staff. AI is dismantling that constraint. So now execution can become abundant when AI can draft, summarize, prep, document, communicate at scale. Capacity isn't the bottleneck. But just because you remove an old constraint does not mean there isn't a new one waiting for you. And that new constraint could honestly be clarity, because with all this capacity freed up, what is your team doing with it? This is where it's critical to just reassess your current strategy as a whole. Roles become more important, not less with AI. AI amplifies your existing strategy. So if the strategy is fuzzy, AI is going to amplify that. If there are gaps in your current process, AI is going to take notice. It could be handoffs, it could be communication, role definition. The firms that will win with AI are not the ones who just have the most AI tools. They're the ones with the clearest sense of how to deploy them and how to ensure the team can work with them. So what does that look like? What does progression overall look like with AI and your team's overall role growth? Think about everything in your office and everything you do in a given week. You can plot almost any task on this graph. How routine versus how urgent it is and how much creativity and judgment it should require. The bottom left is your low energy, low creativity work. Your polling may be performance metrics, formatting reports, updating documents, scheduling. This is where a lot of that staff time quietly disappears. AI can handle this best. The bottom right, it's urgent, but it's still low creativity. Maybe it's templated emails, compliance checklists, same-day documentation. AI can draft these in seconds. The top left, it's creative, but not time sensitive. Maybe it's your service models or your plan design or growing your firm and your overall firm development. AI can be a thinking partner here. But the top right is the quadrant that honestly matters the most. It's the high creativity and the high urgency. A plan sponsor whose maybe CFO is questioning the fund lineup after a rough quarter. A participant who's escalating their escalation that could become more of a fiduciary issue. A committee meeting where you need to be sharp and deeply knowledgeable about that specific plans history. This is where the valuable work is. This is 1,000% human work. And this is where your business will focus with AI. The goal of any AI implementation, the decision should ask this question. Are we moving human time towards the top right? And so when people hear AI and the potential that it could remove the need for headcount, there is a natural reaction for concern. I absolutely love that. That was one of the questions at the beginning of our conversation today. So let's dive deeper and reframe this a little bit. The firms were seen when with AI aren't shrinking their teams. They're growing their books and maybe it's not without direct performance. proportional head count growth, but it's with strategic head count growth. There's a difference between replacing people versus not meeting to hire as many, or maybe you're able to take on junior advisers and, you know, think there's accession planning. We like to usually say you can do more with more. What can you do with AI with the same number of people? What if you hired based on that human energy discussion? We just talked about on the last slide. Before AI scaling a practice had a pretty predictable formula. More clients, met more staff, more staff, met more overhead, more overhead meant maybe your margins got squeezed as you grew. AI can break that formula. AI handles that execution layer. So advisors, retirement plan professionals can do more strategy. The support staff can oversee workflows rather than assembling reports. And for the smaller plans in your book today, the ones that historically got less attention because, you know, the economics and the ROI just didn't justify the time. AI can help level that playing field. Every sponsor could get a thoughtful, quarterly narrative. Every committee could get, you know, properly documented minutes with AI. It's no longer a, could we do this or how do we manage our time? It allows your firm to now make the decision is should we change the protocol? And maybe it is like I said, thinking through succession planning and how your team builds for the future. There's so many possibilities. And let's talk through maybe what some of those could look like with existing roles. So with this slide, the first thing I want to call out is you know your jobs better than any technology platform does. The point of the slide is not to tell you what to do. It's to show you potentially where a ceiling could go. Now that the, you know, we've raised the burden in general, every single role is shifting with with AI, but it's shifting upwards. It's not shifting outwards. AI is removing like we said, that low value execution work from every position. And it's surfacing that higher value thinking. So for plan admins and support, this role has essentially been what I would call like the engine room, right? It's the paperwork, the data entry, just making it go. AI can handle the repetitive turns. So now they can focus on maybe its exceptions, you know, specific governance workflows that require judgment for the retirement plan consultant. Every hour spent on post meaning documentation is an hour and not spent. That could be deepening relationships to grow your book. And for retirement plan specialists, if you know, the specialists are closest to the plan health overall, you're going to see the problems first. AI can give you the bandwidth to identify issues and become proactive on the next steps. And then for the practice leaders, the firm leaders, right, they're responsible for running the whole machine. AI is finally what lets them work on the machine, not necessarily inside the machine. But one practical note, firms that have these types of framing conversations with 13s, see a much better AI adoption. The resistance drops significantly when people start to see where AI is actually taking tasks they don't like and what this looks like for their overall professional growth. The financial professionals who've lean into this early aren't necessarily the most tech savvy. They're usually the ones that we've heard from who just said, you know, I got honest about where our time was going and I decided I wanted that time back. Oh, I love this slide. So knowing where AI is today and the changes you're seeing overall in your role, maybe it's your firm, the industry in general, it is a lot to digest. Some of the questions you could be asking are, how the heck do I even get started? This sounds great. What does that even look like? So one of the benefits of learning how to adopt AI is that this technology can literally be utilized anywhere and everywhere, right? It's not a planning tool that's specific to building a plan, whether you're at work or you're at home, you get to test it out and see what works best for you. So I do want to take a quick detour here before we dive into some more use cases because I think the fastest way to get comfortable with AI is just to start using it every day in your life. So I just want to share three personal examples of how I've utilized AI. First, planning the trip to Disney World. I'm just going to call it out. I am the most type A OCD wife when it comes to vacations. We always plan on taking our daughters to Disney World and I've been nervous for years about this because it's like the high pressure, right? It's supposed to be the most magical trip possible. And yes, that's the exact personality that runs those magical trips. I completely understand that. And because I understood that, I knew I had to do something different this time around. I wanted to enjoy the trip. I didn't want to be the mom who had a plan every single minute. I also never got to go to Disney World as a kid. And so selfishly, this was also my trip and not just for my kids. I decided to offload this. I asked AI to build a full itinerary for my family, factored in ages, which rides have height restrictions. The best times to hit certain parks to avoid crowds. How does sequence dining, dining reservations? What would have taken me days of research across four different parks to one conversation? And then as my husband and I began to adapt our plans, our AI generated trip, adapted as well. And I have to be completely honest. I had a chance to breathe on this trip. I enjoyed every single second my mom, my kids got their mom. And I got, I want to say, I got a little piece of my childhood with it. Second scenario, my husband and I have been finishing our basement. It's been months now. And we've been completely stuck on when I quote, how to pull a room together. I didn't know that was a thing. Pretty, also pretty presentations, cake walked to me. Pretty basement that, you know, feels like our personalities brings the warmth to the surroundings. Not a chance to know creative bone in my body at all. But we knew what we liked in our overall vibes. I described the lighting situation, the paint colors, the initial aesthetics and the general feeling we wanted. AI gave me a full palette with specific color schemes, textures, and how to make the basement warm and inviting. But also the buzzwords I use for millennials, minimalistic, modern and moody. It explained why each color worked well together. It even flagged which combinations might feel too dark, just giving the lighting choices. We just actually picked out our couch last week. And we're finally feeling like we're almost done with it. And then third, and this one may get a laugh, but I read a lot of books, typically fiction, you know, escapism, when you're stressed with work or life or whatever it may be. And I have serious anxiety about any books that end on a cliffhanger. I now ask AI to provide some non spoilers, but also just give me a, you know, mental health win. Is this book series I'm reading going to like destroy my soul, the entire series? Is there positive closure at the end? No details, no spoilers, just yes or no. And it has saved me from lots of emotional disasters with my book traces. These are great stories and all, but why am I telling you all this? The reason I wanted to explain these different scenarios is that the skill you're building when you plan a trip to Disney is the exact same skill you use when you ask AI to prep a review for you. You're learning how to describe what you need, how to ask good follow-up questions, you know, how to evaluate the output. It's easy to start personal, get familiar with it, and then you can bring it into the workplace when you're ready to do some really cool prompts at work. So let's get into some actual professional use cases. This is just again examples of real applications where we're seeing this today in the industry in general. Plan sponsor commute. I think I lost my share. Can you guys hear me okay, still? Yeah, that's my fault. I click. Yeah, I can hear you. No, no worries, no worries. Okay, you guys are seeing it again. Perfect. Okay. So yes, where were real use cases? So plan sponsor communications to me. This is honestly one of the most immediate wins, you know, whether it's a quarterly plan review, that's just saving time, helping you pull the reports, participation data, formatting and narrative. Those can be drafted very quickly. Committee meeting prep, those are also, you know, maybe used to mean a morning of more manual report polling. You can get that narrowed down into minutes as well with a really nice brief participant engagement. This is where, you know, retirement plan advisors have maybe always wanted to do more, but maybe didn't have capacity. Thinking about like personalized enrollment messaging, is it based on, you know, age, band tenure, contribution levels. Maybe you would have to rely on a marketing team or expensive record keeper campaign. AI can make the successful to any practice. All the documentation, I'm also a big fan of, especially for compliance purposes, that's critical. AI that can auto generate, you know, the meeting minutes, build a file will help remove the friction. It'll reduce your exposure overall. And then one thing I'll always be obsessed with is any of the deliverables, these are awesome use cases. So how you present to your plan sponsors, your clients, what that looks like, you probably have a client profile today, right? It's dynamically branded, maybe it's a slide deck. It's usually static, you normally need to manually update it. Imagine having a dynamic one that just changes automatically, whether you've had a meeting, a phone call, you know, a quarterly check-in, whatever that may be, and having the ability to generate those different types of deliverables you can think about like mind maps, different summaries, progress updates, just allowing your team to provide more personal engagements and less time. One thing I do wanna call out with these examples, none of these use cases should require your team to learn how to write code, and we're gonna go through a little bit demo of that at the end of this. So how do you know which AI to choose? Given how many AI tools are hitting the market right now, and there are a lot, I just wanted to give a few fillers, filters for evaluation, a couple additional call outs also for personal usage as well, but data privacy and security is non-negotiable in our industry and also at home. The first question to any AI vendor is, what happens to the data I put into your system? There are some general purpose tools that use inputs to train your models like we talked about earlier. In a world where you're pasting client names, account details, financial situations into a prompt, that's a heavy risk. Review the different pricing tiers. I have chat GPT just for my personal usage at home, while it's done planning for my trips and done these incredible things for our basement, I have never once provided PII information to it. I keep it very general. I have two kids, they're around these ages, this is what we're looking to do. Focus on the enterprise levels, though it's where you start to see the more robust compliance control where PII can safely be used. So with that being said, compliance alignment is definitely the second gate. Your supervision obligations definitely do not pause because AI can now draft an email for you. Having proactive conversations with your compliance team before deployment, not after, save a massive ton of headaches, and finally just general adoption overall, the most sophisticated AI tool that literally nobody uses has zero ROI. Understanding what the AI platform can do and what your team's processes are for adopting new technology is always needed. Everyone has different learning curves and change takes time. Creating a cohesive process takes time, but my goodness, once it all comes together, the payouts fantastic. AI should be used to make your process stronger, not introduce new gaps. So before we get into what I'll call the Fundamo portion, I do wanna close with some thoughts on where we see AI going in the future. So two of the big phrases you'll continue to hear, I'd say, definitely over the next year, maybe next one to two years is agentic AI and embedded intelligence. Because these are going to be the next major shifts and we're already in it right now. So previously, AI has been pretty reactive. You ask it something, it responds. A gentic AI acts autonomously on your behalf. So it's going to monitor your client lists. It could identify who needs a touch point, drafts the outreach, flags it for review, sends it when approved, and honestly within the next 18, 24 months, this could easily be table stakes when you think about competitive practices. Now embedded intelligence, this means that AI stops being a separate tab when you open your browser. Instead, it's going to become native to the tools you already use. Maybe it changes how your team starts the day. Maybe it's now within your team's proprietary system. You're no longer going to have to separate tools to do certain parts of your process. Embedded AI, this means your technology platforms silos are gone. Your data sinks across all platforms and the intelligence layer can now create new opportunities, identify next steps. AI should be embedded within that tech stack, no longer running parallel, but the systems working together. So what does this mean for your firm? Right now, as we've mentioned, it firms that use AI have an edge. In two to three years, though, everyone's going to have access roughly to the same tools. The differentiator, and this is the, I'd say the most crucial part, is the fluency of it. So how well does your team utilize the AI? How well do you prompt? And once you get that data from the AI tool, what are you doing to evaluate it? How are you applying that data and moving your firm forward? How cohesively do you operate across your book of business and within the roles in your firm? The human relationship, the trust, the judgment, especially if it's a volatile market, those things matter more, not less. The firms who learn to work with AI as a true partner are going to be more effective. And here's what excites me about this group here. And I mean this sincerely. I had previous experience in the industry working with retirement plan professionals. And so another reason I was very excited to be on this call today. But the retirement plan split, the retirement plan space has always attracted people who care deeply about outcomes for people who you may never actually meet, right? The participants were accounting on someone to make sure their plan is ran well. AI does not change that mission at all. It just removes the friction between the people who care and the work that matters. And to me, that feels like a good thing. So now let's dive into some of the fun practical applications. I get asked all the time, which AI should I use, which platform? I'm going to keep this completely unbiased and only focus on the public AI's. And honestly, the before I have listed here, chat GPT, Claw, Gem and I, co-pilot, they're a lot more similar than maybe some people think. You can think of them as like four really great co-workers. They probably all went to great schools. They have similar skill sets. They have a slightly different personality, slightly different strengths, but none of them are going to be so different or so much better than the other that you should just hold up and really evaluate which one you pick before you even start. So here's what I actually want to think about it as well. Don't necessarily start with the platform, right? When we're looking at these four listed here, start with what problem you're looking to solve. What is one thing in your week that you just dread doing because it takes too long, maybe it requires too much formatting. You don't actually, like you're not doing critical thinking on it, you're just doing it. Start there and then pick the platform. If your team already logs into one of these most of the time, so let's say that your firm already uses Google to start with Gem and I. If you're a Microsoft shop, co-pilot, if you're starting fresh with zero loyalty to any of these, my personal opinion, Claude is fantastic for anything that requires what I'll call nuance, longer documentation. Chat GPT though is still my go-to at home, one, 'cause it was all the craze when I first started working with AI. I don't use it for any exporting. You can definitely utilize those features. All of these systems are also going to have the ability to connect to other systems. And so you'll want to keep that in mind as well. For co-pilot and Gem and I, you do want to make sure you're aware, typically they're going to be more structured on how they can connect with other systems because they're going to want to stay in that same, I'll call it hemisphere of Google products, Microsoft products. The goal is you're just starting out and you have not used AI before is not mastery. It's just getting familiar with it. It's getting comfortable. Having a conversation with it. If it responds to you, push back on it, ask it to reduce something and you'll start to understand how fast it learns and how fast you can start to build what you want to build. We're also going to focus on the bots. And so the bots are the agents, right? The AI bots that can actually do workflows for you. They can put together some really cool deliverables. All of these platforms have that feature. They're called different things. And so you'll notice under each four, I put the bot name. So if you are utilizing chat GPT, kind of a boring name they call their bots GPTs. Cloud calls them projects, Gemini calls them GEMs and Copilot agents. So if you ever hear bot or agent, just know that each system, whichever one you're choosing to work with, could have a different name before it. But this is where it gets really powerful for your practice specifically because when you start building with a bot, you're no longer just utilizing generic prompts, generic AI. You're starting to build something where it's going to know your team's voice, your clients, your format preferences. That's the version of AI that will really start to feel like an additional team member. And my recommendation, if your firm is still scoping out how to incorporate some of these into your tech stack, like I said, test them all out, put together a list of use cases, have some of your users spend time running through examples. And then similar to what we just talked about, check the compliance on it, make sure you get the right pricing tier. And then if you are a testing all of them, put the same prompt into each of the different platforms. You're more than likely going to get a little bit of a different result just because they have different learning models. So wanting to talk through a live example, I'm hoping this example hits the mark. You may be saying that's a crazy example. But I was thinking through what would maybe be fun to talk through and build today. And so what you're looking at is the template. And I'm actually going to stop sharing my slide tune and actually start sharing cloud soon. But I was thinking through what we could build and what a bot could look like, especially for retirement playing professionals. Fulture and transparency, this was done in, for sure, all of this, what I'm going to show you in the demo less than 15 minutes. And that includes setting up cloud for the first time. I wanted to start from scratch and to say, if I was starting today, how long would it actually take me? So it can move super fast. And I wanted to kind of walk you guys through that. Now I don't have all of my connections set up. So when you're thinking about pulling in client data and you're on an enterprise, you know, cloud plan, you can pull in your CRM data. If you're using platforms like FI360, if they don't have a direct integration, you can typically download the giant Excel spreadsheets from FI360, upload those. In my example, I am only using test data. I actually asked cloud to generate test data at the same time. So if you are looking for integrations and you're testing out, maybe you already use Copilot, maybe you already use cloud or chat GPT today. And you're not finding an example of what their connectors are. Just know that there's probably, you may have to manually upload some of the data, but most of them you have pretty well rounded connectors at this point in time. Each AI platform will be different though. So what you're looking at is I was thinking of like, what if we did a plan sponsor score card? So I put just C7 key retirement, I'm going to call them dimensions up there, participation rate, deferral adequacy, what their fund lineup is, their health, their fee reasonableness. That was a really long word of fiduciary process, stuff like that. And then each one I wanted to make sure it gets color coded with maybe a line explaining what it wears at. And these are just examples. There are probably more relevant criteria that you would want to include and you would want to customize. And then beyond the score card, I also thought about like, what if we had sponsored goals? What if we identified the risks? What if it said, these are the talking points. Next steps. Just putting together like a full deliverable of presenting where you're at with your plan sponsor in like a quarterly review. This example again is just preloaded, but it is something you can build today. So any time you start when you're building anything with AI, you're going to have a prompt and think of your prompt as an instruction set or like a baking with a recipe. So you're going to want to have it be super descriptive. And so what I have on the left here is the full recipe. And you're not going to get it perfect the first time. But you are going to want to assign it a title, right? You're going to want to give it a role. So it thinks like that role. And then you're going to be very specific with how you actually build it out. Typically, I'll start with a prompt of whatever I'm looking to build. And it's going to take me three or four times of just going back and forth to be like, oh, shoot, I forgot. I forgot to include disclosures. I forgot to include my dynamic branding. And so it just takes a little bit of time. But essentially, once you have this prompt built, you can copy and paste it directly into the tool you're utilizing today. And then we will work to create a bot with it. So we're going to test this out. I do want to double check because I know this was a longer presentation. Kelly, are we good if we keep going on like showing the demo within Cloud? Or is there anything-- I think we're good. I think we have until five. Is that right, Madison? Yeah. So we've got 15 minutes. Yep. OK. Perfect. Thank you. Yeah, I just want to make sure. I'll reshare my screen. OK. So you guys are now looking at Cloud. This, again, I just honestly started in brand new account. Today, I wanted to walk you guys through it. Cloud has multiple different functions. I purchased a higher tier. So you're going to see chat. You're going to see co-work and code. Chat is going to be just that simple back and forth. Have the conversations. It can still create visuals for you and provide really cool AI outputs for it. But if you're utilizing just the chat functionality of any of these different AI tools, you're going to have to repeat the process every time. And so when you think about a bot or an agent, that's where you want to scale. So if you build something, you don't want to have to rebuild every time. You can just click a button, maybe upload some new data to it, click Go, and it spits it out automatically. So I literally just took that prompt, I copy and paste it. I can put my branding in here. I don't personally have my own website with all of this. So I said, you know what we're going to call a said Pfizer Family Wealth Management. And then I'm biased towards the purple color scheme. So I said, hey, can you mirror it like the Zox website? It reviewed everything. And it spit this out initially. And so this looks great. It's pulling in, again, fake data. If I did have real data, you can either source it from your connector. So if it hooks up to your CRM or your other platforms, or if you downloaded data from FI360, you could also upload that Excel spreadsheet here. And then I would say utilize the spreadsheet as your source. So it's going to start spitting out exactly what I asked for. But it's still not going to be super pretty. It's going to give me some talking points and next steps. It's looking great. We can pretend that I went back and forth a little bit. I actually asked it if it could create a well-formatted HTML download, because what it spits out here should probably get a better verb than spits I apologize. You can copy and paste it. But right now, I can't actually download it. It's stuck in cloud. And I don't want to copy and paste it because it's not going to look as pretty. And so when I asked for just a really well-formatted HTML download, it brought it back to me. And it actually brought me this. And so you can utilize the download to present on your screen. It even created a print and save as a PDF button. You can make this very dynamic. So I've seen it where if I only want to filter on needs attention, you could click it. It would filter down only showing ratings that need attention. But this is just an example of what it could look like. You can put your disclosures at the bottom. Right now, I just have my branding and what it was generated for. So overall, if that's looking pretty good, I want to present this. I'm going to say I went back and forth and did some editing. But now I want to make sure that I don't have to ask those questions every single time. So I'm going to create what's called a skill. It's called a skill in cloud. They all have different names again. But I said, can you actually create a skill for this so I can utilize it going forward? What I love about this is I just set it in like normal playing natural language. I don't know how to create a skill. I think there's buttons. You can push on the page. I'm not diving in that far. Hey, can you just create it for me? It's like, yes, perfect. I'll create it. So it actually walked through it. It told me exactly what the skill will do. It's going to say what triggers it. And then it's going to give me this nice little download here. And so this is the retirement plan scorecard. So this is essentially all of your baking instructions on your little recipe. So you can save this. You can download it if you want to use it somewhere else. But once I saved it, I actually came over here to co-work. And what's great is when you get set up with cloud for the first time, it's actually going to walk you through, like, hey, let's learn what type of role do you have. I said I was in finance. It offered me different skills. I could just automatically add it then ask which systems I would be interested in connecting with and took me through the whole process. And then, finally, it was like, do you want to test some skills? And I was like, well, actually, I have a skill to upload already. So it's like, great. Go, drag and drop it. And so I included it here. And it built my agent for me. So that's all it is. That when we say there's an agent built, it's ready to go. And now I know, for example, I came back in a couple weeks later. And I just said utilize. And any time you actually type the forward slash, it's going to pull up the skill. And then if blue harvest manufacturing-- again, I said pretend I uploaded data-- but I could say for blue harvest manufacturing, please utilize these sources. And it can be your sources from your CRM, your meeting notes. Zox today, we do have an MCP with cloud. So you can pull in actual meeting summary and client intelligence with cloud. Maybe it's the excels from FI360 or another platform like that. And it'll actually go through. And it'll kick it out exactly the same as before. So you actually don't have to rebuild this every time. You can download the file. So for instance, if this was a brand new session, and I just wanted to get started, I would literally push the forward slash. And these are all the skills. So some of them come automatically with Clawed. This is the one I built. So I'm literally just saying retirement scorecard build for, I'm going for the harvesting theme today. And then I would also be saying for the source, here are the uploads or here are the links to it. And then I'm just going to click the button and it's going to immediately start generating it. With any of these agents you build, you can have them timed. So if there's something you want, let's say every Monday morning, maybe there's something that you generate and you come back to it, you just kickstart your day because you're going to go meet with a plan sponsor, a client for the day. I actually did a full webinar yesterday on just client review deliverables and what that looks like. So there's lots of fun stuff you can do, utilizing these different tools. Just make sure again, as you review who you're using in your firm that you're looking at that PII, that you're checking this example here, I am using the Clawed Pro version. So Pro is one stop down from enterprise. And it asks me, point blank on the screen, can we utilize your data for learning? And I said, though, and I shut that down. So just go through your due diligence. I know there's, like I've been reading, obviously being in the industry myself for years on the strategy side at a home office. I always stayed up to date. I volunteered with Rock the Street Wall Street. So very invested in just women and finding it like in the financial services industry. And there's a lot out there today about, women being slower to adopt AI than men. And I know it's because we're going to look at it much, much more closely, I bet. But thinking about, typically it's also women who are more likely to already be having those types of EQ conversations with their clients, with their plan sponsors. We're already setting ourselves up for success that utilizing AI to remove these low hanging fruit, and the lower execution work to have those EQ conversations. It's just ready and waiting for you. All right, any questions? Is there anything you would like me to go through the demo itself, or just any, I guess, AI questions in general before I think we have breakouts? I know we didn't want to open up the floor for questions. Rachel, thank you so, so much for that presentation. That was amazing. As we start to use AI here in our office, I'm picking up on some buzzwords now, like the prompts and the workflows. So I've got a lot to learn. That was great. But yes, I think we do want to open up the floor for questions. So I think they're probably coming in. Yes, I will. Kelly, I tweaked the slides a little bit with some of the words. So I'll send you over the latest slide deck. So I can distribute it out to everyone. Thank you. Thank you. Thank you. Yeah. Okay, we do have a question coming in. It says, how do we use AI to take meeting minutes? That's a good question. That, I appreciate that question since I work at Zox. But yes, Zox is an example of an AI. We started out as an AI note taker. It's a full financial assistant or AI assistant for the financial services industry today. And so getting set up with any type of platform like that, it can join your Zoom calls, your team calls. There are different platforms that record. Zox specifically is a non-recording transcription only service. But being able to utilize that in your sessions, you can do in-person, you can do it on a mobile app. And then you have all that data after your meetings. Fantastic. I've got another question here. I may know the little clarification. Has anyone used this for processing manual distributions? I might need a little more clarification on that one. I think they're speaking of from like a plan sponsor. One of you there. I'm going to come back to that one. Somebody's asking, what is the co-pilot version of a skill? Yeah, I think it's actually just called a skill for co-pilot as well. And so as you're building through it, it's still just actually teaching it those similar instructions. OK. Fantastic. Thank you, Barbara. About the back to the manual distributions using paper forms. So has anyone used it, I guess, for processing using paper forms? Oh, great question. So AI does offer a lot of those features of OCR. Like if you guys are familiar with Holis to plan that's OCR, it's going to scan the documents. Zoxel so has, I apologize. I was being so objective today. But yes, Zox has document intelligence, which can actually scan different forms. You can actually push structured data to like e-money or CRM. We also have a forms feature. So if you mean by processing forms, you just want to collect it in a structured way, we have that as well. Fantastic. I've got a lot of people coming in saying we have Zox. It's cool. And somebody is asking for the Zox link. So we'll get back to that. Let's see here. For meeting minutes, does legal and compliance have concerns about discoverable information and litigations? Yeah, that's a great question. So whenever you're evaluating any AI technology, you want to make sure that it's objective. So any of our templates, we don't put recommendations. We don't put any type of subjective insight into it. It is strictly, this is the transcript. This is what happened. We do have a lot of firms who are archiving their books and records. You want to make sure you get it over to your CRM. We can assist with automated workflows for filling out your client review, stuff like that. We have certain firms who want different data retention policies. It's all very much up to the firm overall and how you want the compliance and legal to be set up for that. Perfect. Another good question. Curious about this? Our Arissa Council still advises against certain AI tools for this reason. Because of the question before about the concerns about litigation and things like that. What are your recommendations to build strong initial prompts? Yeah, I love that. Be as specific as possible, but also know it's going to take a couple of times going back and forth with the AI. Always, always start out, who do-- like if this was a human doing this job, who was doing that job? So I am a financial advisor, say, you are a financial advisor focusing on x, y, z. Put in the very specifics, how do you want it to be formatted, whatever you're asking? And then making sure you'll notice, especially with the public AI's like chat GPT Cloud push back on it. They are very much going to most of the time to like, yes, you have a phenomenal idea. Let's keep rolling with this. And it's like, no, stay objective. And so just getting comfortable. If there's saying something in a certain way that you don't want it formatted, like push back, there's so much flexibility in how you communicate back and forth with the AI. Excellent. OK, perfect. I don't see any other questions coming in. Let's see here. What was Rachel's take on iRobot starring Will Smith? You know what? I have seen it. I can't talk about movies if I have not seen them. Full transparency is scared the bejesus out of me when I was child and I saw it. So yeah, they have much better special effects nowadays. Excellent. Well, thank you so much again, Rachel. We really appreciate your wealth of knowledge and just really appreciate you being here today. I'm going to pass it on to Cynthia. She's going to make some closing remarks. And then we want everybody to stay pleased for the breakout rooms. Thank you for that, Kelly. Thank you, everyone, for your time today. And a special thanks to Rachel for being our speaker. I thought this presentation was amazing. And I am very appreciative of her for providing this presentation. We hope that you also learned something valuable. And before we open up to the breakout rooms for small group discussions, which we do encourage everyone to stay for, here are a couple of reminders. Please allow two to three weeks for CES to be applied to your account. If you have years of experience in the retirement industry and an extra hour per month to spare, we want you. Sign up to become a mentor through the ARA's drive mentoring program. We have mentees in the queue who are looking for guidance on everything from navigating career changes to developing leadership skills. So you can apply today by scanning the QR code on the screen. July 14th, which we'll be here before we know it, will be the inaugural Women's Retirement Security Day. Observed annually on the 2nd Tuesday of July, this day brings together employers, financial professionals, policy makers, and community organizations to raise awareness, share resources, and advance practical solutions that support women's long-term financial security. At the heart of this effort is the saver. This is the woman building her retirement future. She may be just getting started or getting back on track or working to strengthen what she has already built. Like many women, she balances competing priorities, including career caregiving, family responsibilities and day-to-day expenses, all the while trying to plan for the future. We hope that women's retirement security day will become a day to educate and support the saver at every stage of her journey. You can scan the QR code to learn more about how you can get involved. Finally, save the date for our next third Thursday meeting, which will be August 20th. Thank you.