Demystifying AI Adoption in Financial Services With Geoff Moore

Geoff Moore is the CIO of Valmark Financial Group, a financial services agency working exclusively with entrepreneurial wealth management firms to meet its clients' unique and ever-changing needs. He leads a team of experts who use technology to solve business challenges and manage the cybersecurity of Valmark and its member firms. Geoff has over 20 years of experience in technology and financial services and was selected to the Surge Ventures Industry Council for his expertise in cybersecurity for financial service firms in 2022.

Here’s a glimpse of what you’ll learn:

• [2:37] Geoff Moore discusses how Valmark Financial Group helps independent financial advisors run their practices

• [3:56] The reality of AI adoption among financial advisors

• [8:17] Tips for using AI to give financial advisors more capacity for client work

• [10:07] Connectors, skills, model selection, and the technical side of AI workflows

• [12:55] Geoff talks about AI guidance, safety guardrails, and support for financial advisors

• [15:33] The hidden costs and risks of building custom AI-powered software

• [18:13] Valmark’s AI-powered compliance review skill for advertising

• [21:24] How to treat organization-wide AI skills like software with source control

• [29:46] Geoff’s vision for AI as a personal orchestrator across connected systems
  

In this episode…

AI is rapidly transforming financial services, but many advisors and organizations still struggle to move beyond basic tools and experiments into reliable, secure automation. With questions around compliance, human judgment, technical complexity, and whether to build or buy AI solutions, how can firms turn AI’s potential into practical business value?

Geoff Moore, an expert in financial services technology, cybersecurity, and AI implementation, recommends focusing on practical use cases that increase capacity rather than on replacing human expertise. He emphasizes keeping humans involved when judgment matters, validating AI outputs, establishing safety and compliance guardrails, and getting technical assistance to move promising automations across the 15-yard line to completion. Geoff also suggests treating organization-wide AI skills like software, using practices such as source control, while exploring advanced tools, connectors, and workflows to deliver greater productivity.

In this episode of The Customer Wins, Richard Walker interviews Geoff Moore, Chief Information Officer at Valmark Financial Group, about practical AI adoption in financial services. Geoff also discusses the 15-yard line problem, AI compliance and security, and the future of connected AI-powered systems.

Resources Mentioned in this episode

• Quik!

• Richard Walker on LinkedIn

• Geoff Moore on LinkedIn

• Valmark Financial Group

• Box

• Aaron Levie on LinkedIn

• "Customer Experience, AI, and Cloud Security With Mikhail Ruchkin" on The Customer Wins

• "Transforming IT Buying: Authentic Matches With Tonya Turrell" on The Customer Wins

• "Building Trust-Driven Tech Partnerships With Brian Hyman" on The Customer Wins

  
  

Quotable Moments:

• “I think generally speaking, they're very open to it. I mean, we saw the note takers.”

• “Actually give them more capacity to do the things that they're best at.”

• “But that's where all the good stuff happens. That's what happens really fast.”

• “I really tried to challenge myself to try and leverage it, and create my own personal skills.”

• “Does it work or not? Yeah. Does it solve my problem? Yeah.”

  
  

Action Steps:

1. Focus AI on increasing human capacity: Use automation to handle repetitive administrative work so advisors and employees can spend more time on higher-value activities while keeping AI focused on supporting people rather than replacing human judgment.

2. Keep humans involved in judgment-based decisions: Build AI workflows with clear validation and testing processes so human oversight can ensure outputs are appropriate, consistent, and aligned with the intended process.

3. Help AI projects cross the “15-yard line”: Bring in someone with greater technical knowledge or AI experience to turn promising automation experiments into functional workflows.

4. Establish safety and compliance guardrails: Define clear policies for AI use so employees and advisors can explore useful applications without overlooking security and compliance responsibilities.

5. Treat organization-wide AI skills like software: Store shared AI skills in source control and maintain their versions to create greater accountability and structure as they become part of organization-wide processes.

   
   

Sponsor for this episode...

This is brought to you by Quik!

At Quik!, we provide forms automation and management solutions for companies seeking to maximize their potential productivity.

Using our FormXtract API, you can submit your completed forms and get clean, context-rich data that is 99.9% accurate.

Our vision is to become the leading forms automation company by making paperwork the easiest part of every transaction.

Meanwhile, our mission is to help the top firms in the financial industry raise their bottom line by streamlining the customer experience with automated, convenient solutions.

Go to www.quickforms.com to learn more, or contact us with questions at support@quikforms.com.

Powered by Rise25 Podcast Production Company

Episode Transcript:

Intro: 00:02 

Welcome to The Customer Wins podcast, where business leaders discuss their secrets and techniques for helping their customers succeed and in turn, grow their business.

Richard Walker: 00:16 

Hi, I'm Rich Walker, the host of The Customer Wins, where I talk to business leaders about how they help their customers win and how their focus on customer experience leads to growth. Some of my past guests have included Mikhail Ruchkin of Innovative Digital Transformation, Tonya Turrell of TechnologyMatch and Brian Hyman of High Meadow Solutions. Today, I'm speaking with a prior guest and a friend, Geoff Moore, chief information officer at Valmark Financial Group. Today's episode is brought to you by Quik, the leader in enterprise forms processing. When your business relies upon processing forms, don't waste your team's valuable time manually reviewing the forms.

Instead, get Quik! Using Quik!, you'll be able to generate completed forms and get back clean, context rich data that reduces manual reviews to only one out of 1000 submissions. Visit quickforms.com to get started. All right. Geoff and his team of technology experts focus on delivering secure systems that enable operational excellence for member offices and business units. His team was recognized as a finalist for the Northeast Ohio Tech Team of the Year by the Greater Cleveland Partnership.

 And Geoff, you may have seen him. He's very active in industry initiatives and groups including Surge Ventures Industry Council, Barron's 100 Financial Services Institute, Industry Roundtables, and box.com works. Geoff was appointed to the Advisory Board of the Northeast Ohio CIO Forum, where he contributes to strategic discussions on regional IT leadership and collaboration. In September of 2025, Mark was awarded the Small and Medium Business (SMB) AI Disruptor Award from box.com for innovative use of AI features inside of box.com. Geoff is a frequent contributor on fintech, cybersecurity, and AI on LinkedIn, and has been a guest on dozens of podcasts, including this one.

 So I'm really excited to have him back. Geoff, welcome back to The Customer Wins.

Geoff Moore: 02:15 

Oh, thanks, Rich. Boy, that was a laundry list. Yeah.

Richard Walker: 02:19 

You've been out there doing a lot of things. So for those who haven't heard my podcast before, I love to talk to business leaders about what they're doing to help their customers win, how they build and deliver a great customer experience and the challenges of growing their own company. So Geoff, let's understand your business a lot better. How does your company help people?

Geoff Moore: 02:37 

Yeah. So we help independent financial advisors help run their practice. So some of our niche specialty is we work in both the RA independent broker dealer and life insurance to help create holistic estate plans for people and just really help advisors run their practices better.

Richard Walker: 02:56 

All right. So how does technology fit into that?

Geoff Moore: 02:59 

Well, yeah, that's the wheelhouse. Really trying to just make it easier for our offices to work with our back office teams and us to be as efficient as possible. So the last, really, year and a half, we've really been focused on AI and how we can use AI with our offices. How can we use AI internally? And I'll say I've been really focused on this question of trying to figure it out.

And it seems like almost every day there's some new news story about how different people are using AI or new features or capabilities that are available. So it's been a really exciting, I'd say, couple of years. I haven't really felt this excited since the early days of the web with all the innovation that's been happening.

Richard Walker: 03:38 

All right. Can I ask you to get real? And I'm not saying you weren't, but I mean, like, this is a great opportunity to hear really what's going on with the advisory world because a lot of us are hearing, oh my gosh, they're doing this, and they're doing that, and they're so far advanced, etc. we all think everybody's ahead of us.

Geoff Moore: 03:54 

That's right.

Richard Walker: 03:55 

Where are we really?

Geoff Moore: 03:56 

Yeah, that's a great question. And I think it's easy to go look on LinkedIn or read some news story about some innovation. But the reality is most of us are still in that mode of just maybe, maybe not using AI or just using it to answer basic questions. Schwab just came out with their benchmarking study, and I think the data was collected around March. And in the case of where advisors were using it for sort of automation, I think their stats were not exactly precise, but around like the 20% mark.

So for advisors that are doing that, that's still a relatively low benchmark. And I think if you would look in some of the other, even more advanced note taker systems that offer additional integration functions, if you would look and see what percentage of people are even using some of those advanced integrations, I think it's definitely less than 100% right. So even though you see maybe the vendor using or making some of these features available, you know, not everyone is going to be able to leverage and use all of those. So it's kind of a long answer, but you know, to keep it real. Yeah, I think we're still learning and growing for sure.

Richard Walker: 04:58 

Do you think a lot of these benchmarks are the kind of people you have? If you're not using ChatGPT specifically, if you're not building your own things directly, you're not using AI because if you bought a software product that's using AI, aren't you using AI?

Geoff Moore: 05:11 

Yeah. Spot on. So that's the other thing, right? It's embedded in so many of our tools now, right? You might not have a separate subscription, but you might be provided by the software tool you use every single day embedded right in it.

Which is, which is also interesting from a security compliance perspective, because, you know, we want to check, we want to make sure that all our tools are using these things. So when they add them, it's an additional burden to make sure that it's working and complying with everything, even though they didn't have that feature before.

Richard Walker: 05:41 

Yeah. I look, I'm going to give you a premise that I have because I've been developing software to automate engineering and software development. And so I've learned a lot about what AI does well and what it doesn't do well. So I've come to the conclusion that the future of AI is actually software that makes judgment. I look at it that way because if you let the AI make all the judgments and therefore do all the work, you don't get consistent work.

It'll make different decisions every time on the same thing and do it in slightly different ways. But what we actually want in operations is a reliable, consistent process where it can make decisions as it needs to make decisions. So I'm curious about your view on that. If you think I'm right or wrong, I'm open to whatever.

Geoff Moore: 06:22 

Yeah. How to get it to make better judgment, more consistent judgment.

Richard Walker: 06:28 

How to.

Geoff Moore: 06:29 

Get.

Richard Walker: 06:29 

It. How do you, how do you ensure that it's actually performing the work you're supposed to do? A lot of people are saying, oh, I have AI. It can now do all these things for me. It can take over my workflow.

It can figure out these answers.

Geoff Moore: 06:41 

The back end, the validation, the testing. So let the AI do the thing, but then have some sort of rigorous way to verify that it's doing what you think it's doing. Yeah. Sort of like in software development, like the test-driven development methodology, except instead of a developer, it's AI doing it and seeing if it outputs to some spec.

Richard Walker: 07:02 

So a lot of my guests have echoed the same kind of sentiment around. You can't imagine a client who lost their spouse talking to an AI bot, right? It has to be with a human. So we can't replace humans with AI when you need human judgment. And that's the part I'm also thinking about.

If you have this great note-taking system and it then says, hey, you need to open up this type of account for your client, should it just open the account for the client? No. You want to make a judgment about that. You want to know what the actual conversation was. And so I look at this and say, well, really, where does AI fit?

 It's the moment you say, hey, let's open the account. If there's any kind of judgment decision processing, it's going to be like what forms to get or what information to pull. Let it do that, but don't let it do anything besides that. That's kind of where I'm headed in my idea of where software and judgment belong.

Geoff Moore: 07:50 

Where the delineation is. So this whole debate about is AI going to replace advisors? You would clearly then be in the camp of. No, it's not. It might help assist and get the forms ready or some of that stuff, but it's not going to push it in between you and the client with the example that you just gave.

Richard Walker: 08:09 

Well, and so let's reverse it. How many of your advisors are using AI to replace themselves versus doing what they're really, really good at more of?

Geoff Moore: 08:17 

Yeah. Yeah. Actually, give them more capacity to do the things that they're best at.

Richard Walker: 08:22 

Yeah. So are you seeing that then? Are they actually gravitating towards tools that help them do more of what they want to do and love what they do, or is there a resistance to it?

Geoff Moore: 08:33 

I think generally speaking, they're very open to it. I mean, we saw the note takers; it was the first big use case that was used by advisors. I had a story about an advisor. A young guy had three kids under five. And he was actually spending from like more like that 5 to 8 hour p m like actually typing up his notes, not spending time with his kids.

So when he moved to using the note taker, then he was using that extra time to spend with his kids. I mean, it's kind of a more heartwarming, touching story, but I think they're definitely trying to do more of the automation. They're still questioning like, how do I do this? How can I take some more of this type of work and automate it for me? You know, one of the things that I've started to observe, though, is I'm calling it the 15-yard line, where I think a lot of people have an idea that automation can help them or a vague idea of how to do it.

 And they might, and they'll get into like the 15-yard line using a football analogy, but taking it that the rest of the way to score a touchdown is where it's not quite there. And usually I think somebody's a little bit more technical needs to come in and give an assist. That's just what I've observed. It seems to be working well. So get the idea, get it most of the way there.

 But then you still, even with. I might need somebody with a little bit more of a technical brain or maybe just a little bit more experience. Somebody that's been using it a little bit longer to help you get it across the finish line.

Richard Walker: 09:54 

Okay. I want to explore that. So does that mean the software they're using, the product they bought just doesn't take it all the way? Or does that mean they're trying to build their own systems and workflows with AI, and they just don't know how to make it go all the way or something else?

Geoff Moore: 10:07 

I guess I'll get real technical for a second and use just like a general AI system. So inside a general AI system, you've got a connector; you can connect to some other system, and you can write something called a skill, right? That's just a set of instructions on how to use that connector and interact with it. And there's a lot of different ways to do all of these things and to stitch them together. And I think that's the part where somebody with a little bit of guidance and expertise can, can be helpful in how to do it, even something as simple as which model to use.

I think that's still confusing to a lot of people. Now that these systems let you pick which model to use, you know, the, you know, Anthropic released their fable model. Should you use that for everything? Probably not. It's going to get really, really expensive.

 And then the other thing is just even being like. The discussion around tokens, like everybody's really unclear what that means. And it's hard to describe. And it's almost like something you have to experience until you see the bill and you go, okay, well, I guess that's maybe don't do it that way or just have a little bit more awareness. I think I've been spending a little bit more time just giving people education and kind of showing them, you know, okay, here's A, here's B, this is what the tokens look like with A, this is the cost, what it looks like with B.

Richard Walker: 11:24 

Look, I might be A because I'm in software; I might have a little bit of a bias on this. Yeah. But this feels a lot like 20 years ago when we were all talking about integrations, like, should we be building our own API calls or integrations with other systems, or should the systems provide that capability? And now AI is making it more compressed because it's so easy. Everybody can do it.

Geoff Moore: 11:47 

Yeah.

Richard Walker: 11:48 

But should they? I mean, really, should they be spending their time that way versus just wait for the software to evolve and then use it?

Geoff Moore: 11:54 

You could, you could be right. Rich. Just a little bit more patience. So I think there's some of us that are going to be a little bit more bleeding edge that aren't going to be as patient to wait. But, I think over time, more and more vendors will build out those integrations.

And we've started to see that, right? If you go in like even in a general AI tool, going to connectors, the amount of connectors available today are so much more robust. And then even the tooling within the connections has gotten a lot better. Like just thinking of a couple vendors off the top of my head, you know, the version of one of their products, it might have had like five tools in it. And now maybe they've got 25, 30 more plus tools available to them as they've had more time to build them out.

Richard Walker: 12:35 

Yeah for sure. How's valve positioning themselves and maybe you and your organization? I mean, do you think of yourselves as being AI consultants for your advisors so they can do business faster and make more money? Or is it like, well, hold on here. We're just going to evaluate the software you're getting and help you choose the right one.

Or like, how do you guys navigate this?

Geoff Moore: 12:55 

That's a good question. I think it's a combination of. Well, it's taking a positive approach to it, right? With safety guardrails in place, right? Having policy for safety guardrails, like operating within these guardrails, I think it's definitely something we're exploring is like, how do we, how do we do a better job of giving guidance both for safety and kind of what's available?

So that's something I'm noodling on quite a bit right now, actually, is how to do a better job of that for advisors, how to help more of them get to from the 15-yard line to the goal line. And I think that's going to definitely evolve over the next 6 to 9 months, for sure.

Richard Walker: 13:33 

Well, and I think I'm seeing this across all industries where I saw somebody say, I just saved $600,000 on Salesforce because I built my own version of Salesforce over a two-week period, coding it myself. What they're not saying is how much they're going to spend in maintenance and build and ongoing versions, etc., over time. I mean, that's 600,000 might be tiny in hindsight.

Geoff Moore: 13:57 

So I'd also say.

Richard Walker: 13:59 

You invest.

Geoff Moore: 14:00 

Well. Okay, so two, two points there. So one, they didn't build Salesforce, right? They did not build Salesforce. They built a small subset of the features that were available in Salesforce that they actually cared about.

So that's the first thing. The second is, okay, so you can write this, this coded app thing or whatever, but some of the harder parts are like, how do you, how do you handle the infrastructure? How do you handle the administration? How do you handle the safe storage of data? Like I've seen a number of examples where people have put up their version one prototype, but then when they had to get it production ready.

 I mean, it would take much, much longer. So I feel like some of those are not. You're not hearing the full story, right? Or it's not coming from somebody like you that's built like enterprise grade software. And, you know, of all of the like, they might not even be aware of all those extra things that are needed versus when you're designing it.

 Yes, you build out that feature, but you also have to spend a lot of time putting all the safety infrastructure in place, the backups, the disaster recovery, like all of these other things that are not necessarily code things or like administration things that are part of the overall software solution.

Richard Walker: 15:07 

Yeah, it is the hardest part. I mean, look, it took us, took us three years of development to get to a point where we could get a SoC audit. And our first SoC audit cost us over a quarter million dollars. You know, if you're building a product for financial services for your business and you're never going to get a SoC audit, how do you know it's secure? How do you know you've done all the right things, and can you afford that kind of risk or expense to make sure you did what you needed to do?

Geoff Moore: 15:33 

Yeah. And I tell people still like it's the end of the day, if you're going to build that, somebody has to sit in front of the terminal, the interface and type chat stuff, right? Like, like that still has to happen. So even though it's sped up, even though it's faster, it's easier. It looks, you know, better than something you create on your own.

That's still somebody's job. And, you know, in a financial services org, is that still the highest and best use of that person's time? Or is it better served, you know, talking to clients, servicing clients, doing marketing activities, you know.

Richard Walker: 16:07 

Do you get a sense from any of your advisors or what percentage of advisors are pushing back on this idea that they need to be AI first versus just let me do what I do? Let me focus on that, and we'll get automation over time. Do you see that kind of difference in people's mindsets?

Geoff Moore: 16:26 

Well, I think I probably have a selection bias in that the people who are talking to me are the ones that are probably most interested in doing this kind of activity. So I've probably got a. Yeah. So I got a little bit of a bias. But if I look at it just like that.

Like the percentage of people that are on our enterprise plans for different software packages, that gives me a pretty good indication of, you know, what people are thinking and where their focus is. I'll say, though, that at least with the note takers, that was probably one of the fastest adoptions we had of any software that people wanted for sure.

Richard Walker: 17:00 

Yeah, I know, I was astounded at how fast they've gone through the marketplace, and they've grown. And I'm excited for it because I think it's adding a ton of value to the market and to the advisor.

Geoff Moore: 17:11 

I do think, I guess two things. One, I think for advisors specifically, some of them are still like, there's so much noise in the news. Like some people want to wait until the dust settles. So I do think we still have some dust settling to do, so to speak, before people feel really comfortable. Like I still, you know, we've had ChatGPT as an enterprise option available for over a year now.

And I'm still getting some advisors going, like, okay, well, I think I'm ready. You know, and so it's taken even a year, even with all the new cycles and everything.

Richard Walker: 17:39 

Yeah. I find myself using ChatGPT less and less these days.

Geoff Moore: 17:44 

Yeah, I know, I know we're kind of anthropic as well.

Richard Walker: 17:48 

But it's not just that I like. I've answered so many questions I wanted answers to, and now my team's using it more, so I'm using it less, I don't know. I see we've gone through this kind of cycle of I was the leader driving people to use it. So I was showing all the examples, building the skills or the custom Gpts. Now my team does that and I do less of it. So maybe that's why I feel that way.

I still use it almost daily. It's just less daily.

Geoff Moore: 18:13 

Yeah. I think the other thing that'll start to drive it too, like is just more of these advanced use cases come online. More people are going to want to use it. So like one of the drivers we just had, we just released a compliance review skill. Funny that the first enterprise skill we launched was from the compliance department, but it just does a first check of their advertising review using some Finra rules, and it is still reviewed by compliance at the end of the process.

But it's just meant to be a first pass. Just check it through, check it through here. Put in your advertising request and it'll check the rules. And then once it goes through that, send it to compliance. They'll do their final check.

 But it's just meant to be a little bit quicker. So I think some firms or offices are starting to go, okay, well now maybe that's a good reason for me to go ahead and, you know, pay whatever a couple dollars a month.

Richard Walker: 19:04 

All right. So let me ask you then why did you guys build that? There's tools in the marketplace that do that already. Why did you build it?

Geoff Moore: 19:11 

That's true. That is true. I, I guess I'm, I'm a little bit more, I, I've kind of grown up in the more like enterprise software and customizing enterprise tools as opposed to buying individual niche solutions. And so, you know, for the cost of the license. And then once you build the skill, you have it.

And this is for a simple example, like, right. I think there are other pieces of software more complex that would probably be better suited to, to purchase some of those things. But there's definitely a number of like smaller kinds of small automations that I think makes sense to roll up into a skill and then share org wide.

Richard Walker: 19:48 

I think that makes sense because when we're doing that here at Quik!, we're building skills and sharing them. One thing I'm struggling with is how do I know if they're used? And maybe I can go look at the usage. How do I know people are knowing when to use it and actually invoking that skill? How do you guys try to make sure people actually take advantage of what you built?

Geoff Moore: 20:07 

This is all new to us too. So I think looking at the analytics, although I will say, I feel like at least in the anthropic ecosystem, it does a very good job of detecting when a skill should be used.

Richard Walker: 20:19 

That's true. That's true. It does. It does pick that up. Yeah.

It's such a wild world right now because there's so much you can build on your own. And yet there's that last mile, that last 15 yard problem that keeps popping up over and over again for situations you find yourself in. You know, for example, we're building prototypes really fast on things. Well, some of the prototypes are so good that we're saying, how do we now port that into a production product? Internal product only, not an external product.

Geoff Moore: 20:48 

Right.

Richard Walker: 20:49 

And you still have to think, okay, do we get it on our own servers? Do we have all the source code? Do we have to modify that source code? Do we have security controls on it? Do we know if we have the security controls on it?

Geoff Moore: 21:03 

Yeah.

Richard Walker: 21:04 

And I, I presume that anybody in your role who's managing this has a similar fear as I do about security and the breaches and, and the challenges of stuff too. So when people come to you guys for either internal, you know, organization or external with your advisory firms, how do you vet a new product?

Geoff Moore: 21:24 

Yeah, well, we haven't, we haven't got into the whole like building product game yet for say, like with offices, for example, it's more limited to just skills, for example. And I'll say internally we've started treating at least org skills like software. So like checking them into source control, for example.

Richard Walker: 21:43 

Yeah.

Geoff Moore: 21:43 

I don't know if other organizations are doing that, but that's just we felt like if it's, if it's a process that's defined and someone might change it later, it would be good to have it checked in the different versions as people are going through and making changes to it.

Richard Walker: 21:58 

Yeah, I think that's a good idea. I'm not sure we've thought of skills that way yet.

Geoff Moore: 22:02 

I know it's weird because it's this weird place where it's like, it's almost like, well, if somebody wrote like an Excel macro, would you put it in source control? I don't know. But if we kind of decide, like, if somebody writes a personal skill, that's one thing. But if it's like an org skill that's going to be used across either all of internal employees or out to the field, or we're going to check it in.

Richard Walker: 22:23 

Yeah. So you personally, how much time do you put into building things with AI? Or do you.

Geoff Moore: 22:31 

A lot of.

Richard Walker: 22:33 

It was way too much.

Geoff Moore: 22:35 

Probably. Yeah, definitely my preference is cloud code in a terminal inside of VS Code, I have a whole folder of projects, even small things that we're doing just I'll say like right now we're starting our budget season. So like, just for a simple example, I made sure I loaded in all of my last year's, all the, you know, charges, invoices, everything, you know, here's my template, here's some things, extra invoices and had, you know, AI, take a first. This is something I've done for years, right? So I have a good idea.

I just had to take a first pass at it this year. I just want to see how it does it. I really tried to challenge myself to try and leverage it, create my own personal skills, that kind of thing. So I'd say probably a couple hours every day, just my own productivity. Or if I, you know, we hear something from somebody else to try to help somebody else get over the 15-yard line, for example.

Richard Walker: 23:30 

What do you think has been the biggest challenge to doing this for you that you might share with others?

Geoff Moore: 23:37 

Oh, man, that's a really good question. What is the biggest challenge? I, I for me, this is going to sound weird, but I think a lot of people are afraid of a terminal or some of these advanced use cases like Cowork or some of the new apps that ChatGPT and I would say just go for it. Like that's going to give you your highest leverage because I think there's, I, I'll give you a simple example. I had our CFO, she was working on a very advanced project.

She was loading Excel sheets into chat, right? The chat window. And she's like, oh, it's timing out. It's not working quite as well. And I'm like, oh, well, you actually need to put it in something like Co-work or like one of those tools or like a terminal, because then it's not going to necessarily fully process up inside, You know, the cloud, it's going to use some local processing.

 And then it worked for her. And so I think there's a lot of things people could unlock by kind of going from that next step, going from the chat window into like the cowork window, or maybe even all the way to the terminal. Even if you're not a developer, that's what you're going to get the, some of the, some of the best leverage and speed using AI. I'm sure you already know this way better than I do for sure, but well.

Richard Walker: 24:50 

Yeah, I mean, you're past that, but that's okay. Yeah.

Geoff Moore: 24:53 

Yeah, yeah, yeah, I.

Richard Walker: 24:54 

I just helped a friend who is not a software engineer who's building a fintech product go through this path. And you're right, it's this concept of, oh, I can do everything in this chat window on the web browser. And it's, it's super amazing. Yes. You can do so much.

And then you told me he was doing coding that way. I'm like, what? No, no, no, no. Yeah, take a step back here. And I do think you're right.

 There is a reticence to what's this black screen, this vs code world or this terminal. And why is it all just blocky text? It's like you've entered the matrix and now you see the matrix, and you don't want to be in that. You want to go back and eat steak.

Geoff Moore: 25:29 

Yes, yes. But that's where all the good stuff happens. That's what happens really fast. And like, I know you're well beyond this. When you get all your tooling and connectors and knowledge bases all set up, then it can start to really cook really fast.

Richard Walker: 25:45 

Well, and look, you hit on one of the two things that I think are the hardest to work with AI. So one is context and how much it can store in its memory. So context actually means two things. One is what is the memory allocation, which you can think of like RAM, how much does it have? And two is what does it know about your need, your problem, your project, like you said, you're loading up invoices or Excel spreadsheets.

How much of that memory can it store? And did you give it enough? Like what if you didn't give it a whole series of, I don't know, Amazon bills and it didn't know about that and your budgets off because of that. Oops. Yeah.

Geoff Moore: 26:21 

Exactly.

Richard Walker: 26:22 

Yeah. And the second one is behavior. How do you get AI? I use cloud the most or ChatGPT or Gemini or any of them. How do you get it to actually do what you want it to do?

And I laugh almost every day, somebody complaining about it, not doing what you wanted it to do. Hey, go rename this. Okay, I've renamed it, but there's something else you should know that I found like, no, no, no, no, just rename it. That's all. Don't tell me all this other stuff.

 Have you found that to be a challenge?

Geoff Moore: 26:51 

I don't find it a challenge. I. I don't know, I actually think sometimes technical people have more challenges with it. Actually, as strange as that might sound, because they're so used to the precision of code. And sometimes I think the variability of some of the AI systems can be just unnerving almost to some of them.

Also, I was a liberal arts major, so for me it actually feels more natural to me. Like, this is the most natural interface that's ever felt. And I think some of that's just coming from more of a little bit of a different background or even being in management, right? Because management's a lot about just telling people what you want them to do and then giving them sort of like rough guardrails or not as precise and giving them a little bit more variability in the outcome. And so I'm very used to just giving people sort of more general instructions, letting them have some autonomy in how it gets delivered.

 And that's more what AI is. But if I were to give very, very specific specs, sometimes I think that can be hard. I don't know what you think. You've done a lot more with kind of narrowing some of these systems.

Richard Walker: 28:00 

Yeah. And look, I do come at it more from the software standpoint. But look, just saying a document is my biggest frustration with ChatGPT. I have a long document I want built, and I've talked to it about what I want. And so it starts drafting it.

And maybe it's going to be eight sections long. It'll do the first three sections and say sections four through eight follow the same. Just repeat. No no no, please, ChatGPT. Finish it.

 Okay, I'll do section four for you. And it does it separately. It doesn't add it to one through three or it deletes one through three, and it only puts it in section four. It's like, come on, do what I've asked. Just put it all together for once.

 That's the kind of behavioral stuff that I see that drives me crazy. But I also agree with what you said, that it's coming back with some very open-eyed, open-ended ideas about what could be done. And to me, it's one of the best brainstorming tools I've ever had. It's the most I've ever felt like I'm working with another person without being another person on the other side.

Geoff Moore: 28:55 

I actually think that's been one of the hardest things. When the models change or they switch, if it's almost like getting, it's like getting a new thought partner, and it's a little bit different. It interacts differently. And you're like, oh wait, I was used to talking to you and you're not the same person or whatever. I know I'm anthropomorphizing it, but yeah, I noticed it especially with the latest Opus Five release, that it definitely seemed very different from previous iterations.

A little less playful, very concerned with testing, safety, double checking, triple checking. It's like, oh, okay, you're a little different.

Richard Walker: 29:30 

It is, it is. It's kind of wild. Do you have any predictions about what's going to happen with your firm, your industry, AI? I mean, if you forecast three, four years from now, do you have any vision of what that looks like?

Geoff Moore: 29:46 

Yeah, I have this concept. I think it looks more like talking to systems naturally and a lot more systems connected and your AI kind of acts as your personal orchestrator. I don't know if that's three years off, five years off from just, you know, space kid left field, but it feels like you're going to be able to talk more naturally. A lot of your systems are going to be connected together. The AI is going to just help you kind of coordinate all that stuff for you.

That's how I envision it. That's kind of what I try to set up for myself. I don't know you.

Richard Walker: 30:27 

Well before I, before I answer, I mean, do you at Ballmer, are you guys thinking that way? Are you building the underpinnings for communicating amongst systems, or your advisors having communication with your direct systems?

Geoff Moore: 30:40 

I'm definitely thinking about the ways that AI can interact with our systems and making sure that we're able to participate in that conversation.

Richard Walker: 30:50 

Yeah. You know, for me, I, I've kind of said it in the sense that I think it's going to be software that has judgment. But what I really think is that more and more software companies are forming; they're finding new ways of employing AI. And I think over time, they're all going to converge on the concept that software should drive the process. And AI should be creative within that process and only within that process, not outside of it.

And so I think the technology tools that we're going to buy off the shelf are going to get way better and way more compliant, safe, easier to use, fun, etc., because these software companies are thinking that way. And while right now everybody's kind of building their own house on the prairie and figuring out how to, you know, carve out their land, etc., I think a lot of that will get replaced over time with really robust enterprise-grade software that's been tested and tuned and optimized.

Geoff Moore: 31:42 

So I think what I'm hearing, what you're saying is, yes, it's going to be great to help create more options, but it's still going to be the domain of experts to help build niche solutions that work really, really well to do all the things you just described. And while some of this stuff is around, you're still going to have your main systems that are really because it's still like, like we were talking about somebody still has to sit in front of the terminal or write the spec, or think through the problem or ask the right questions that you might not even know which questions to even ask the AI to work on.

Richard Walker: 32:15 

Yeah, see, I don't agree with the premise that many software people have, which is, hey, I've given you my whole source code. I gave you the spec, go figure it out and come back when you're done. I don't think that's the way. I think you have to really have humans in the loop to engage in the process, to make clear business decisions and product decisions and design decisions. But you know, that's me from a software background.

What if we're talking about medicine? What if we're talking about logistics? What if we're talking about all sorts of other parts of our world, and AI is being used in those ways? I still think the software makers in those facets or industries will continue to converge in this manner, too, and will just be the beneficiaries of this AI. So part of the question, and this is where this comes from, Geoff, is if you forecast 5 or 10 years from now and you had built a product that you had named AI, whatever, would you regret calling it AI?

 Would it still be called AI, or would it just be a product?

Geoff Moore: 33:09 

It's just the product that does the thing.

Richard Walker: 33:11 

Yeah. Right. Because in the end, I don't think we care. Does it work or not?

Geoff Moore: 33:15 

Does it work or not? Yeah. Does it solve my problem? Yeah.

Richard Walker: 33:19 

Yeah. Oh man. I, I'm realizing time. So I gotta wrap this up before I get to my last question. What is the best way for people to find and connect with you?

Geoff Moore: 33:29 

The best way is LinkedIn. So Geoff, with a G, GEOFF MOORE on LinkedIn, Valmark Financial Group.

Richard Walker: 33:38 

Nice. Awesome. Yeah. And you are out there; you are, I love it. You give these impressions of you at conferences and what's going on.

You're very vocal, I love it. All right. Here's my last question. Who are one or two people you're following closely to stay up to date with AI and all the things that are changing right now.

Geoff Moore: 33:55 

All right. I'll give you two people. The first one is Aaron Levy with box.com. And I'll tell you why, because you hear a lot of these benchmark studies that come out. A new model comes out, a benchmark comes out and it's like, well, what does that even mean?

Right. And often box when a new model comes out, they've got their data sets that they prepare for, like their customers and actual real life documents, and they're evaluating it against their data sets, which I think is more of like a real world benchmark representing like what, you know, businesses are actually doing with their data. So I think it's pretty good there is a good thought leader, but also their perspective is just, they're putting out some benchmark data that I think is realistic. The other one that I listened to all the time is Nathaniel Whittemore with AI Daily Brief. He posts something daily.

 I don't know how he has time to collect, even if he's using AI extensively. He's very thoughtful and just, he's a good thinker for people that are trying to actually use AI within their businesses and to help their customers. So he's a, he's another really good follower I have, I'm just trying to stay up to date with what's going on in the world of AI.

Richard Walker: 35:02 

And I'm glad we asked this question. I didn't know about Nathaniel. I didn't know Aaron was so prolific with AI stuff too. That's great. Yeah.

I'm glad you're not asking me this question. I couldn't name two people, so. All right. I'm going to give a big thank you to Geoff Moore, chief information officer at Valmark Financial Group, for being on this episode. The second time The Customer Wins.

 Go check out Geoff's company's website at valmarkfg.com. And don't forget to check out Quik! at quickforms.com, where we make processing forms easier. I hope you enjoyed this discussion. We'll click the like button, share this with someone, and subscribe to our channels for future episodes of The Customer Wins. Geoff, thank you so much for joining me today.

Geoff Moore: 35:45 

Thank you Rich. Take care.

Outro: 35:48 

Thanks for listening to The Customer Wins podcast. We'll see you again next time. And be sure to click subscribe to get future episodes.