New call-to-action    New call-to-action    Spotify

Download

Read the Transcript

James Robert Lay: 

On November 30 2022, AI finally reached the mass consciousness of humanity with the launch of chat GPT, reaching more than 100 million users in just two months. And since then, a apps and tools and integrations that continue to grow at an exponential pace as financial brand leaders try to cut through the confusion with a mix of both excitement as well as caution. That's why on today's episode of Banking on Digital Growth, we're going to explore the opportunities for growth as well as potential pitfalls to avoid as we continue to journey forward together in this new age of AI.

 

Greetings and Hello, my name is James Robert Lay and I'd like to welcome you to another episode of the banking on digital growth podcast. Today's episode is part of the exponential insight series. And I'm excited to welcome my friend Connor Heaton to the show. He is the VP of Intelligent Automation and AI as strategic resource management and today, we're going to gain some awareness of opportunities for transformation. opportunities for growth when it comes to new AI tools, commonly termed large language models are MLMs that are spreading throughout the financial services space, while also established technology vendors scrambled to integrate them into their current offerings. We're going to also dive into and discuss potential pitfalls, pitfalls that you need to watch out for pitfalls that you need to avoid that can trip you up on your own journey of growth in this age of AI. Welcome to the show, Connor, it is good to share time with you today.

 

Connor Heaton:

Thanks so much, James. Really appreciate it. Glad to be here.

 

James Robert Lay:

Absolutely. And before we get into talking about AI and MLMs what is good in your world right now personally or professionally? It's your pick to get started.

 

Connor Heaton: 

Sure. I honestly I'll go with just the the developments in AI I mean, the space is moving tremendously fast. Yeah, there's there's too much to keep up with, which is kind of a good problem to have. And seeing how things are unfolding is is tremendously exciting. Also just having a new tools launch every other day, it seems that solve problems that I've wanted to solve for a long time. way my first project responsibility ever really was meeting minutes for federal clients. And with AI tools. That's that's basically a thing of the past. Yes, it's kind of cool to seeing all these things realized so quickly.

 

James Robert Lay: 

I like the way that you frame this. This is about solving problems that you've been wanting to solve for a very long time. And if we go back in kind of just recent history, it's not not too far too long ago, it was November 30 2022, that will say AI reached the mass consciousness of humanity with the launch of chat GPT. And I'm curious to get your take on this. Why is this an important date to consider? Because it wasn't that long ago?

 

Connor Heaton:

Sure. So I think you put it well, when you said this is when AI kind of entered the mass consciousness because AI technologies have been around in some form or fashion depending on on what we're classifying as technology since 1960s or so when the some of the first algorithms were really being developed. And it's had a number of kind of summers and winters of investment and excitement. And really November 22 with the sort of explosion of chat GPT at marks, I would say a new era of awareness and democratization of AI tools. So before AI was embedded in a lot of the things that we do there was some level of awareness that when Netflix recommended shows to us or Amazon recommended products or social media feed showed us something from a particular friend that there was machine learning working in the background to decide and feed those things to us. But it wasn't really something that people outside of academia or tech or engineering really spent a lot of time for the most part working with. And so having having the launch of Chechi btw and immediately useful, interesting and delightful tool just for free with all of these capabilities that are that are shockingly refined compared to sort of past iterations of the same thing because you know, there was there's a GP two there was an instruct GPT the the transformer architecture has been around for a while and actually We've played around with it a little bit, some years prior. But it just wasn't to a point where it had the immediate and obvious usability and it wasn't something that was put out there and able to kind of go viral and impress and fuel the news cycle. Yeah, it was, it was really, I think, just a, a coincidence of the capabilities, hitting a new and important benchmark. And that being accessible and, and testable by just about everyone. And so that was a perfect fuel, perfect timing for it to go viral, and really get a ton of attention and start to fuel this massive rise of AI investment and development and product realization that we've been seeing ever since.

 

James Robert Lay:

Yes, and you know, one of the things that you're discussing is, it's AI has been around, but it wasn't accessible to the quote unquote, masses, you touched on machine learning, or M ml. And I think it's important to just pause for a minute to help those watching and listening to gain some clarity about the relationship between AI and LLM or large language models. This is an area you spend a lot of time thinking about, what's the context here for someone watching and listening to be aware of?

 

Connor Heaton:

Sure. So AI, artificial intelligence means a lot of things. And that makes talking about it somewhat difficult. Right? You know, it's it's the buzzword, we don't hearing it in different spaces, different ways for a long time. But when it comes down to it, it's it refers to a whole spectrum of technologies that are in some way, fueled by, by machine learning, where instead of a program of software being explicitly coded to take a series of steps to a particular thing, it instead is coded to learn from a set of examples in order to guide its ultimate behaviors, right. And so that's, that's used in any number of things. Right? It's in image recognition, it's in speech recognition and processing, it's in sentiment analysis, it's in document extraction, optical character recognition, all different kinds of analytics. It's in digital assistants, it's in natural language generation, it's in image generation now. You know, it's in knowledge representation. There's any any number of areas, right, so the way that I tend to think about AI, and what I tell clients is that you should think about AI as infrastructure. It's, it's how something is done, it's not a what, really, there are AI products that are products built on top of AI, there are products that are, you know, 95% just regular code with AI bolt on for some piece of it. There are tools that have trained up an algorithm like like loan decisioning, for example, or customer segmentation, which would be trained up on their algorithm that on a set of data, and that's machine learning, that's AI, but in terms of how it executes it, it's not learning in an ongoing fashion, right? It's it's a bolt on that provides one piece of a broader technology solution. And so that very solution to solution how much sort of ML is in it, but usually, it's, it's a piece, it's a it's a, how it's not what? And so, you know, when we talk about doing AI, or investing in AI, usually we're talking about solutions that are built on top of, of AI in some form or fashion.

 

James Robert Lay:

What surprised you most, since the launch of of chat GPT thinking about the pace of change, it's not just the pace of change, it's really the pace of exponential change that we have experienced since then, in all different areas. If you think back, you know, to 1994 you had the launch of the internet, essentially, once again, reaching the mass consciousness of humanity. It had been around in academia before that, but 9094 I think it's it's fun to go back and watch how the Internet was being talked about on the like the Today Show, what is email and things that we take for granted. It's almost History is repeating itself now where there's some confusion. There's some hope there's some optimism there's some some people are scared about this, but what surprised you most Oh, just thinking back, you know, to November of 2022, with the launch of GPT.

 

Connor Heaton:

Sure, so I have something that specific that I'll Get to but in terms of like the explosion of investment, I'm maybe a little surprised by its magnitude. But less surprised by kind of the course that things have taken. So I was working in automation, specifically robotic process automation. Back in my days before, the current kind of AI, Renaissance, and the way that a lot of these things tend to work with technologies that primarily automate tasks is there's a paradigm shift, there's a new type of technologies, there's an advancement in new architecture. And that's used to, there's sort of like that seminal moment where it comes out in the market realizes that it's valuable and can be productized, outside of the research space. And then the next kind of five to 25 years, depending on on what technology we're looking at, are spent then building that into the business environment. Taking that core capability and productizing it in hundreds of 1000s of different ways to go from that fundamental sort of new finding new capability that lets new things be possible and then making those easier and more accessible for industry use, and also adapting business processes and adapting industries and roles and structures and services entire industry sometimes, around that automation. And that's that's just the history of how this sort of thing works, right when when we moved from from switchboards to to centralize telephone routing, and we moved to cell phones, you know, the introduction of spreadsheets, spreadsheet software, and ultimately Excel, right it, you know, accounting used to be done very differently. And, you know, now it's just kind of second nature. To us, this is just how we do things. But there was a whole period after the introduction of those tools that were early adopters, they were late adopters. And a lot of old business processes didn't match perfectly with the new capabilities. And so there's this adjustment on both ends on the vendor side, on the solution side, you have the tools getting more usable. And on the client side, you have the business processes being re engineered to work better and take better advantage of the tools to maximize the value you're getting out of them. So I'm not I'm not actually too surprised by the direction things have taken with AI investment, because I think we're seeing very much the same thing right now. What I didn't expect, perhaps about this, this. This newest architecture is the level of emergent capabilities, which is, you know, kind of like a add on the imprecise word in this field, because it can refer to so much it's very vague and sometimes used as, like a blanket, just the model did something we don't really understand how it's emerging. But but it's the case, it's the case where seeing just the massive spectrum of domains, where large language models have had a huge impact and rapidly, right, we could not just in a research capacity or an A, this is something that concerns data scientists, but in the way that this substantially transforms different industries, and how we, how we understand and do things, right. So it's, it's not just a bot that can produce text and extrapolate from text to to try to finish writing a chapter of a novel, right? Or to write marketing copy. It's a virtual assistant that's shockingly capable, from everything from summarizing lots of data to analyzing tone to evaluating new content to copy editing, to code generation, and code review. You know, and then you hook that up to the right things. You can code websites, basically, from from a prompt, you can generate images, if you add in, like the particular framework that's being used to generate images from static and you add in some of that semantic understanding. It's just been a kind of puzzle piece that has slotted in in so many different places. And particularly the the ability to sort of reason that that has come out of this from just feeding these architectures, enough data, the right data, doing the right training and fine tuning to let them basically draw on their training sets to say what's a sensible next course of action given XYZ input? And that's it enabled some truly insane things like, what surprised me most I think when I saw it some months back was some of the impacts to robotics, which I know isn't really a financial industry sort of area. But it was, it was crazy to me to see a demo of, I think Google's palm II, that was able to plan and execute actions are basically you have have the large language model acting in some senses, the brain reasoning to interpret the visual input to say, and what am I looking at? Then what do I need to do? What's the what's the plan and sequence of actions to take care, and then that being able to be translated automatically, in real time into instructions for the physical robot body to to flexibly complete tasks, you know, to pick up objects, clean things up, get something from a drawer, and to have that be resilient to interference? You know, I mean, we've all we've all seen the Boston Dynamics, those robots adjusting, right, but that's that a lot of that is kind of the is more hardcoded. But like the the reasoning, or the appearance of reasoning that you get out of these and hooking that into agentic models and things that can act as agents in the physical or virtual world. That is something that that they're really thrilling.

 

James Robert Lay:

And this is I want to stay on this point for just a bit longer about exponential change, because what you're speaking to, is not just impacting financial services, it's impacting all different areas of life, the robotics is a great example. But we also can look at healthcare, we can look at education, we can look at, you know, what is this going to do to transportation even. And it's, it's almost like the paradox of choice. There's just so much for our little linear brains to process that it becomes a little complex and overwhelming. And if technology is moving on an exponential curve, and we're linear thinkers moving this way, it's creating this gap. And I know there are a number of financial brand leaders who feel like they've been hit by an AI avalanche, particularly coming out of the COVID chaos. And I've had some conversations with CEOs with executives, I've been on the road speaking. And a good number of them, they're tired, they're tired, they're overwhelmed from all of the changes that they've had to deal with going back to early 2020. I said this, you know, back then early podcast that this is going to be a decade of exponential change that even really kind of thinking about AI at the moment, it was more of a COVID perspective. And many are just trying to keep their heads above water. Their perspective, though, I think of the future is directly and will directly impact and influence how they think and make decisions now here in the present moment. So thinking about this pace of exponential change in some of the perspective that you just shared. That Yeah, might not be financial, financial services related directly, it will indirectly impact this vertical. How can financial brand leaders keep up with all of the change going on right now in this age of AI?

 

Connor Heaton:

Sure. So I guess I will, I'll take that question and break it into two pieces. One is how do you keep up informationally? And then two is how do you actually keep up keep up in terms of ability to cope as an organization and to adapt? The the answer to the first one, I think hasn't changed substantially AI is a wonderful set of tools for Data Summary. I have lots of articles that cross my email or desk in some form or fashion. And the ones that are longer and don't need precise detail, I absolutely feed through large language models to get a summary. You know, there's free tools to do that with YouTube videos. You know, there are many, many options, many of them free even to help process more data more quickly. And that's you know, I SRM is a relatively small company, I'm sure that this is a much more refined pipeline at other places where it's, you know, automatic data, scraping of news articles, improved summarization, and like those those sort of pipeline tools have existed for a long time longer than these AI, sort of new capabilities. But now, you know, like, like I was talking about earlier, it's more accessible, it's cheaper, it's easier, it's better. And so, so there are absolutely tools to help sort of ingest more information sooner. And I think as as always with this, you need to understand your priorities. So having a strong strategy and knowing how are you planning to win in your space, what matters to your customers or members? How are you going to compete into the future it helps to guide and bring that At a firehose of information down to a more manageable level. It is important, I think, to still sometimes go outside of that typical lane to, that's where you see some really exciting innovations or where you you see surprising solutions to problems sometimes. But, you know, for the day to day keeping up, I think that's that's what I would recommend. As far as adaptation. The other half of this, something I talked about a lot with executives is change agility. So, as you said, changes, exponential things are getting faster and faster. You know, there's the classic kind of Moore's Law of computing that's been taken and apply to a whole bunch of other places. As far as like just the pace of change. And even looking at AI particularly, you know, we in computing, in general, we went from mechanical computing, to programming, and that was 1900, to 1950. But 50 years or so, we got into more of the digital era. And like 30 or so years, we entered the early cognitive era in the 2000s, and 2020. We're having this like aI explosion, right, going from mechanical tools to digital and like increasing levels of automation, and more and more kind of builds on top of each other and making new things possible, and moving faster and faster. And so, you know, this is just a fact of life. Now, there are going to continue to be lots of new innovations, lots of things coming down the pipeline, and you need really strong strategy, like I said, to help decide what's important, what's not, what do you need to focus on, and what's the threat, having solid horizon scanning is important to keep an eye on that, and understand where and when, and how things are hitting your industry. And really structuring your organization to be able to deliver on change, right. So if you don't have a good learning and development, group and program to help upskill people to help them adjust to changes to policies, if you don't have solid internal communications, right to let people know, hey, there are these large language models out here. But by the way, if you put any proprietary data can end up in a public training data set for this set of tools. You know, you hear hear things you need to know about new types of fraud. And that's not just a technological capability, that's a culture thing, you have to balance the the overall sort of workload of people so that they have enough room, you have the right volume of communication, your right channels, you don't have too much that duplicative people aren't so overloaded, that they don't pay attention to any of it. There's a lot of like fine tuning to get that right. And then having a solid business transformation function, right to actually do it run projects. So projects are more and more important. We often see clients not paying enough attention to sort of enterprise transformation in terms of where does it sit? What power does it have? how closely do we curate, and, and trim our portfolio of projects? Is it all stuff that's pushed by vendors were our dependencies, it goes pretty deep, I won't, I won't go down to the bottom level. I'm sure folks don't don't need all of that. But there's a lot of work that can be done to improve how good your organization is in managing change. And as our environment speeds up, that becomes one of the most important things competitively. institutions that can adapt will survive, and ones that can't will slowly become irrelevant.

 

James Robert Lay:

That's a that's a great point. And one that I've been thinking a lot about, I've been working on my second book banking on change. And the subtitle is how to achieve exponential growth in the age of AI. The book has nothing to do about AI, per se, it's how do we focus on human transformation in the context of AI? Because you think about the studies of digital transformation historically, it depends upon who you're reading, but they're all saying the same thing. 60 to 85% of digital transformation projects fail or failed to meet expectations. And it's not because of the technology, it's the human element. And so when we look at at AQ, is something that we can measure AQ being adaptability quotient, and like EQ AQ can be measured, it can be trained, it can be expanded, I've often worked with leaders and said, you want to increase your AQ right now. They're like, Yeah, take a take a cold shower for seven days, and see how your mind not reacts but responds to environmental changes. And it's funny because many don't take up the challenge because they're comfortable. They're stuck in the cave of complacency. They're comfortable with what they know. And and I think that's the other thing too perspective shapes reality when working with financial brand leaders, and we've been having these conversations about AI, particularly chat GPT, because it's just kind of a hot topic. And I look at this very objectively, and I have a whole, no judgment. And I said, Well, what's your perspective of chat? GPT? Like, oh, well, that's what kids are gonna use to cheat on at school. And I go, Really, why do you why do you feel that way? Why do you believe that way? Well, that's what I've been hearing. That's what I've been seeing on the news. And so I think back to the other point you're making education is critical here. And we have done some research studies that have found that around 70 to 80% of financial brand leaders invest one to two hours or less per week and ongoing learning and development. But when we look at some of the secondary market research coming out of like Microsoft and Google, they're like, No, we need to be investing five hours a week, 10 hours a week, and ongoing learning. And I think there's just this this conflict here. So I'm curious when it comes to AI, and in this world of large language models, and you share some really good examples of how you're feeding articles and even YouTube videos to get summaries. I think of what Peter Diamandis is doing with a platform called Future loop that I'm subscribed to, I get summaries of what's going on. And it gives me the kind of the horizon without having to get too deep, but to kind of get the larger picture. When it comes to AI and large language models. What's the misconception, a common misconception that you're hearing financial brand leaders talk about, but you have a different perspective on based upon your knowledge of the world?

 

Connor Heaton:

Sure. I think that it is, it's largely a lot of what you said around kind of, they've seen it for one domain, or maybe they spent, you know, 10 minutes back in November playing around with chat GPT and have a sense of, of maybe what it can do, or they've tried a few things with it. But they really don't have a good sense of holistically exactly how much it underpins, in terms of new capabilities, and all the different domains, which this is driving forward. As well as having, I think there's also misconceptions around thinking that it's just a chatbot, for example, or, you know, it's more like the fuel generation of intent based chat bots that folks may be familiar with that replaced a been replacing IVR is kind of slowly over time. And so I think it's important to kind of connect the dots here to see that, yeah, language models are driving changes to content generation of basically every type. You know, it's video, it's text, it's audio, it's video editing and replacement of, of doing like video effects editing, having virtual actors, it's voice synthesis, both for customer service, and on the negative side for fraud and cyber threat. It's, it's accelerating development. It's, it's fueling a disruption to art and content creation. And so there's this common sort of, I would say, there's there's level of understanding that you get from seeing how these things connect, that helps you to make better decisions make better use of these tools, and that's something that that we actually advocate for, pretty often, it's hard to know what to do with the space with as much as is going on. But given how much of our just existing technologies, you know, our our loan processing and origination systems, our our marketing apps, right? Anything that we're using for content review generation, scheduling most of those solutions, if they aren't using AI, and specifically in sort of large language model infrastructure in some form already, they're likely to be doing that in the future. And so there are common strengths, weaknesses, works, understanding to kind of get the most out of these tools, that that can help across a lot of different domains. And so we in the sort of noise that we see in the space right now, often what we recommend is get better at change, right, as is just something that that very often needs to happen. And particularly like get get your people starting to learn and experiment with large language models in a safe context. So have an AI policy do training on large language models specifically, because there's a lot of low hanging fruit today, for F eyes, and particularly for marketing and copywriting, I imagine a lot of folks are already using large language models for content generation, but doing so in a way that is compliant, that's safe, with regulations and with your brand standards to keep your reputation safe. And having having your employees starting to learn these tools and become experienced and understand. This is, you know, this is what they're good for. This is what they're bad for these pitfalls, will help a lot in the future, as so many more things have some level of this architecture baked into them. So even if you don't know what API product or products, you're ultimately going to be using a there's an absolute storm of them right now. And that's likely to continue for the next while. Having having some experience with this and making sure that that experience is being gained in a safe way can only really benefit you in the future.

 

James Robert Lay:

That's a great point. It's almost kind of like, you know, going back to being a curious kindergartener, or taking the Socratic approach to, I know, I know nothing. And then just getting out playing, experimenting and learning and looking at the world through a lens of curiosity, you're talking about some of the content. You know, there's Jasper, there's copy.ai, even I facilitate a workshop about building a website that sells and how that what I'm what I'm teaching, you know, exploring the use of, of mid journey as a creative capability for UI and for UX, and opens up a whole new world. But it was interesting, I was working with a group of marketers recently, and I asked everyone in the room and there's a couple 100, I said, Would you raise your hand, if you have at least just tried, we'll just say chat, GPT, or any type of generative AI capability, Everyone raises their hand for the most part, how many are using it on a weekly basis? I would say 70% of the hands go down, how many of you are using it on a daily basis? Maybe 5% of people had their hands left up. And I think and I'm like, Well, why. And maybe it's the fact that we have to rewire our own minds, to think through this new context that AI capabilities in large language models. It's another teammate, it's a, it's someone that I can collaborate with, to co create. And it's just a different way of thinking, which leads to a different way of doing. But what I've found is that sometimes to bridge the gap between the thinking and the doing comes down to the feeling. It's like, I gotta feel like, I'm confident in this capability. And so either the capability of the AI or the capability of myself, which were, you know, I think we can talk a lot about growth opportunities and potential here from different lenses. But But I want to, I'm gonna turn this to the other side, dangers and pitfalls, and one you touched on. And I hear people often bring this up, around voice and around video. And I know that you recently did a session and had Morgan Freeman, join you in on this conversation. Although it wasn't Morgan Freeman, it was a deep fake of Mormon, Morgan Freeman, what are the potential problems or pitfalls that we need to consider as financial brand leaders when it comes to AI and large language models that that could be an impediment to future growth?

 

Connor Heaton:

Sure, so I would say there are two really big ones. And it's more it's more challenges to implementation. And challenges to getting realizing the value that you hope to realize out of them. One is kind of the classic shiny object syndrome, right is there's all this hype, we don't know what to buy, we don't let's just find a tool, do something. And even within that, there's a lot of choices around do you build your own? Do you fine tune the model, and so on so forth. But you don't want to have a solution in search of a problem. You should start again, from your strategy from your priorities, to understand what are we trying to do? What are the use cases that we're trying to solve for? And then what can we do to apply large language models or more traditional AI or traditional software solutions to tackle those? Right? It still needs to be based on your strategy and your prioritization for what you're doing. There are there are institutions I recently read an article about Bloomberg and they're producing their own internal generative, bluebird GPT basically for but and they started back in 2020. And they didn't really have any particular use cases in mind for it. And they've learned a lot. It's the they're planning to invest even more in it. But, you know, it was it was a solution in search of a problem. So it's perhaps good for learning and for getting, building the skills within your organization to give yourself that kind of challenge, but it's not as good for solving business problems, right. So so start from from what you're trying to do. That's been true for, for AI projects for decades and remains remains true. The other thing is data. So SRM tends to work with organizations, mostly below the 10 million assets mark. And a lot of them even the larger ones, tend to not have their data environments in order, right, they have data in a bunch of systems, they're not integrated, they don't talk to each other, they don't have a centralized data warehouse that they can query. You know, they they don't have trust in their, their answers to different questions are asking we are data driven enterprises from a decision making standpoint. And so data is necessary for really any custom AI implementation you want to do, you know, there are uses, there are plenty of tools that you can buy off the shelf, that will provide a level of value. But if you are planning to do something that is using proprietary data that is customized to you, you're going to need the data to be able to train it. And you're going to need to have your technical environment set up in such a way that you can hook it into an automated flow. Right. So there's, there's some commercial off the shelf that you'll tend to have less configuration into and that you know, there's there's a middle ground right with some of these where they have some built in training either as part of the implementation or that you can do afterwards in a in a more business friendly way. But that still requires having your data together, right? So and that's just something that I think, you know, most CTOs we talk to, it's on their list, you know, they know that they need to be better at data. But without concrete sort of use cases to point to or that like a CEO that really understands the value of data and really drives towards it. Those those initiatives seem to end up deprioritize fairly frequently. And if you're if you're part of the turn, remember the statistic like 6060 to 90% of institutions that think that AI is going to give you a competitive edge of that AI is a really important area of investment for you, then you should really think about increasing the priority of the those data initiatives. Because that's that's going to be important foundational piece to it.

 

James Robert Lay: 

That's a great point on two fronts, data, yes, but also perspective at the C suite at the C level executive level CEO, because that will help drive prioritization. The other point you made to about potential roadblocks and pitfalls is the shiny object. Right now, it is like, I think you mentioned as we started this conversation, it's like drinking from a firehose separating what is hype? And then what is actual practical use case that can create value. But it does come back to what's the strategic objective and tying that strategic objective back to accountholder pain points, or member pain points, externally are looking internally, how can we be more efficient, and free up our people's times to do more important work that leverages AI capabilities? As we start to wrap up, I want to get really practical here and send someone away with you know what, yeah, this is a this is a big ocean to boil. We don't have to do that right now. I want to help people increase their courage to commit to move forward with confidence. And a lot of that is by just getting a simple win and making some progress getting some wind in the sails, if you will. What's one thing something small, that someone who is watching or listening can do next on their own journey of growth, to integrate AI or large language models into their daily operations? And maybe maybe just starting with their own personal daily workflow and not even thinking about organizational change at this point? What would be one small thing that they can do right now?

 

Connor Heaton:   

Sure. Next time, you need to write a complicated email, or report or social media or blog post. As long as there's no proprietary data in it, run it through GPT, right make make an account with open AI, turn off the the data sharing. And then use that just ask it to copy edit, right to make it more professional or expand on it or make it more concise, whatever, whatever you think needs to happen. Or even try feeding in the input email and ask how should I respond this in order to accomplish whatever goal And just by the process of doing that, even if in the end it is, even if it takes you more time than just drafting the email would have, you'll have learned things about how the technology works, what it can do, how to work with it. And you may be really impressed by what it outputs, or maybe you won't be right. And you'll have discovered some of the limitations or ways that you need to work with it differently in order to get what you want out of it. But that's something that's that's small, easy, very low hanging fruit. And that, you know, once you see the possibility there, it very easily scales into, okay, you know, let's, let's consider using this for more of a marketing copy, or, you know, those board reports that we have to generate, or those worldwide communications, or intranet updates, you know, I know a few CMOS, who, for whatever reason, own the intranet or the website. And you can use content generation content review, you know, check things for consistency, integrate new data into an old summary. Right. So there's, there's lots of very easy things that you can do using free tools that that are very easily available. And I think seeing that possibility, then a communicates the value of this and shows you the sort of runway that exists, and makes you consider if this is what we can get just with something totally out of the box, a free tool. For the customized tools. You mentioned Jasper, for content creation, marketing, there's also Canva for presentations. There's tome AI and gamma AI. I mean, you know, any number, right. So, so those really have much more specialized, it's productizing, the capabilities that you're seeing when you tie that into specific use cases that have value for your enterprise at a more out of the box level, right. So it's, it's a, it's showing that value, and it's starting your own sort of learning journey. And you may you may find, you may find that you end up using it a lot more.

 

James Robert Lay:

And I think that's the key takeaway of all of this. It's to always be learning. And great point you mentioned, too, when you're trying to experiment with some of these large language models and GPT platforms might take you quote, unquote, longer to do what you're looking to do. But that's part of the experimentation. And I know that being mindful and aware of, hey, I'm going into this to learn it'd be faster if I did it the old way. But over time, I'm going to gain back time by using this as a collaboration tool, the more that I learn how to use it, in being able to quote unquote, hack time, this has been such a great conversation, Connor, what's the best way for someone to reach out and say hello to continue the conversation that we've started here today?

 

Connor Heaton:

Or you can find me on LinkedIn, or email me at C Heaton at SRM. corp.com But yeah, happy to chat more. We're building more and more offerings around AI advisory from a general you know, how do we get our arms around this? How do we educate on this to more specifics of how do we develop a compliant AI policy that can be that sort of north star as we move into this space? So there's a lot of exciting stuff going on and always happy to chat.

 

James Robert Lay: 

Connect with Connor learn with Connor grow with Connor Connor, thank you so much for joining me for another episode of banking on digital growth. This has been a lot of fun today, buddy. Thanks so much, James. Until next time, and as always Be well. Do good. Be the light

 

Brief Summary of Episode #323

On Nov 30, 2022, the introduction of Chat GPT  brought AI into the limelight, amassing a whopping 100M users in just 60 days.

Such a surge led to a myriad of innovative AI-based tools & applications.

Connor Heaton, VP, Intelligent Automation and AI at Strategic Resource Management, Inc. joined the Banking on Digital Growth Podcast to shed light on AI's incredible potential in finance.

Imagine revolutionizing customer service, honing risk assessment, bolstering fraud detection, and reshaping financial forecasting.

But, it's not all sunshine; challenges like potential biases and the enigmatic 'black box' dilemma remind us of the importance of human oversight.

The key takeaway?

Embracing AI isn't just about automation; it's about continuous learning, exploring beyond the usual, and ensuring both humans and AI work hand-in-hand for groundbreaking innovations.

To all financial brand leaders out there: Let's lean into this exciting future, harnessing AI's full spectrum for unparalleled growth & value.

 

Key Insights and Takeaways

  • How to keep up with exponential change? (16:20)
  • Misconceptions about AI and its benefits. (27:15)
  • Dangers and pitfalls of large language models. (31:00)