AI Strategy for CFOs Is a Wild West Without Governance Turn AI Into a Portfolio System – Dave Trier
In this episode of Future Finance, Paul Barnhurst and Glenn Hopper sit down with Dave Trier, CEO of ModelOp, to explore the challenges and opportunities of implementing AI at scale in enterprises. Dave shares how organizations can manage AI responsibly, measure ROI, and move from scattered pilots to a disciplined, industrialized approach. He also discusses the critical role of CFOs in AI oversight, change management, and creating measurable business value from AI initiatives.
Dave Trier is CEO of ModelOp, leading the company with a focus on customer value, product innovation, and enterprise execution. With over 20 years of experience across AI, data science, analytics, cloud, and enterprise software, Dave is a patent-holder and trusted partner to CIOs, CTOs, and AI leaders. Prior to becoming CEO, he shaped ModelOp’s product strategy and held senior roles at Think Big Analytics, Powered by Action, and Accenture Technology Labs. He holds a BS in Electrical Engineering from the University of Notre Dame.
In this episode, you will discover:
How to industrialize AI delivery across an enterprise
Managing risk, governance, and compliance for AI implementations
Measuring AI ROI using financial, feedback, and usage metrics
The CFO’s role in AI oversight and rationalizing AI investments
Key lessons for change management and process discipline in AI adoption
Dave Trier highlights how enterprises can move from scattered AI pilots to a disciplined, industrialized approach that delivers measurable business value. He emphasizes the importance of governance, change management, and cross-functional collaboration to ensure AI initiatives succeed. CFOs play a key role in oversight, setting financial parameters, and rationalizing AI investments.
Follow Dave:
Website: https://www.modelop.com/
LinkedIn: https://www.linkedin.com/in/davidetrier/
Follow Glenn:
LinkedIn: https://www.linkedin.com/in/gbhopperiii
Follow Paul: LinkedIn: https://www.linkedin.com/in/thefpandaguy
Follow QFlow.AI:
Website - https://bit.ly/4i1Ekjg
Future Finance is sponsored by QFlow.ai, the strategic finance platform solving the toughest part of planning and analysis: B2B revenue. Align sales, marketing, and finance, speed up decision-making, and lock in accountability with QFlow.ai. Stay tuned for a deeper understanding of how AI is shaping the future of finance and what it means for businesses and individuals alike.
In Today’s Episode:
[00:00] – Trailer
[02:07] – Meet Dave Trier, CEO of ModelOp
[04:57] – ModelOp & AI Governance Explained
[06:21] – AI vs Data Governance
[08:11] – Evaluating AI ROI for CFOs
[13:24] – AI as a Managed Investment Portfolio
[16:43] – Change Management & Process Discipline
[20:48] – CFO’s Role in AI Oversight
[27:38] – Tips to Maximize AI ROI
[30:16] – Enterprise AI Complexity & Coordination
[32:13] – Dave’s Journey: Electrical Engineer to AI CEO
[35:12] – Closing Thoughts
Full Show Transcript: Host: Paul Barnhurst (00:00):
Welcome to the Future Finance Show where we talk about treasury management. Future Finance is brought to you by Q flow.ai, the strategic finance platform, solving the toughest part of planning and analysis, B2B revenue, align sales, marketing and finance seamlessly speed up decision-making, and lock in accountability with Q flow.ai. Welcome to another episode of Future Finance. I'm one of your hosts, Paul Barnhurst, and once again, I'm joined by my illustrious co-host, Mr. AI himself, Glenn Hopper. Glenn, how are you doing?
Co-Host: Glenn Hopper (00:55):
Good, Paul. Good to see you.
Host: Paul Barnhurst (00:57):
Good to see you as well. And then we also have a great guest with us today. We're really excited for this conversation. Kind enough to join us is Dave Trier. Dave, welcome to the show.
Guest: Dave Trier (01:07):
Thank you. I appreciate being here.
Host: Paul Barnhurst (01:09):
Excited to have you. So Dave Trier serves as CEO of ModelOp, where he leads the company with a clear focus on customer value, product innovation, and enterprise execution. He has over 20 years of experience in data science, AI, analytics, cloud, and enterprise software. He also holds multiple patents. He brings deep technical expertise and a pragmatic, transparent leadership style. He's a trusted partner to CIOs, CTOs, and AI leaders. Dave previously served as ModelOp’s SVP of Product, shaping product vision and strategy. He's also held senior roles at several companies, including Think Big Analytics, Powered by Action, and Accenture Technology Labs. He holds a BS in Electrical Engineering from Notre Dame. So again, thanks for joining us. Love the background, and we're really excited to chat. I think the subject of the decade is about right? AI
Co-Host: Glenn Hopper (02:12):
Is about right. Dave, I'm going to jump in and we were talking right before we started recording here. We have more and more of these it seems. But one of these podcast episodes where I feel like I could be talking, this could either be a podcast or I could be asking you about something I need to implement with a client right now. So with that in mind, I'm going to ask this as sort of a client-centric question, but I think for our listeners, it'll hit home as well. So right now I'm working with quite a few public companies on their AI implementations, and for all of them, I mean there are three main areas that they're concerned about: compliance, data security, and trusting the ai. But the biggest one we have to address is compliance and governance. And a lot of people I talk to, I'm coming in through the office of the CFO normally, and they're not yet comfortable addressing this. I think the timing of this recording is interesting too, because I'm sure you've taken probably a deeper look at it than I have, but the treasury just released it's long, what do they call it? The AI F. It's the Financial Services AI Risk Management Framework. And they've set out some guidelines now, try
Host: Paul Barnhurst (03:21):
Saying that fast three times.
Co-Host: Glenn Hopper (03:24):
Yeah, they've got an acronym that I think is hard to say as the full phrase itself. And I know you don't focus exclusively on compliance and the CFO's office, but thinking about what ModelOps does and the role that you play in helping companies roll out and manage their AI implementations. I guess two questions there. First, are you familiar with the fs, AI RMF, and is that aligned with what you guys do? And maybe even before that though, explain to our listeners what ModelOp does.
Guest: Dave Trier (03:57):
Sure, happy to. So ModelOp, we've been focused for over seven years on one thing only, and that's helping large organizations and enterprises be able to govern, manage, and operate AI at scale across the enterprise. That's all we've ever done. We want to make sure that they can use AI reliably, rapidly, and responsibly as they're starting to think about where and how they can use AI to transform their businesses. So, naturally, to your question, yes, absolutely, very attuned to all the different regulatory frameworks, whether it's the O-C-C-S-R 11 seven, which was the grandmother of all risk management frameworks and financials, or the EU AI Act, NIST put out ai RMF. Canada's got one with E 23. So yes, familiar with all of those. And at the end of the day, and I'm sure you'll ask is they're just about how can you put the right level of oversight and risk management and mitigation in place so that you can again, responsibly use AI throughout your organization?
Co-Host: Glenn Hopper (04:54):
One of the issues right now, I mean outside of the trust issue and all that, but is the governance. And it's got to be, it starts with data governance, right? Well, maybe you tell me, do you see AI governance just as an extension of data governance or is it, or is it a lot more complex than just data governance, which a lot of companies still have a hard enough time with that alone?
Guest: Dave Trier (05:18):
Yeah, no, it's a common question. They are distinct overall, right? Data governance is focused on all of your enterprise data and making sure you have the right security, understanding of taxonomy, oversight, and management of your data. AI is different and distinct because it's not just data. It's actually a combination of data and software with a probabilistic approach to using data and software. So there's not always a deterministic answer. There's not always the, if we put in this, we know we're going to get that, right? So that's something that is very, very distinct from anything that we've done in the past, right? It's not like traditional software. It's not like traditional data governance, it's not traditional supply chain, if you will, because you have that intermingling, that mix of data plus software technology, but also the inherent risks that are involved because of the probabilistic nature of these overall algorithms at the end of the day that are being used to make these decisions. So there's an inherent risk factor that is not apparent in data governance as much as others.
Host: Paul Barnhurst (06:21):
That makes a lot of sense to me of that inherent risk factor, right? Data, you know what data is, you can look at it, as long as you write the query the same way and nothing's changing the database, you're getting the exact same data back. Not like when you ask gen AI something or if you can't tell the SQL query, no, you're wrong. I have the right answer, and it's this, and it won't come back and say, yes, you are right. Sorry, I'm wrong. When it was right all along
Guest: Dave Trier (06:49):
Right? Or it's so confidently answers it and so you believe it, right? So SQL is ever going to try to embellish anything. It's just going to give you the answer, right?
Host: Paul Barnhurst (06:59):
Yeah, exactly. Our favourite was when we were testing one of the AI tools. Glenn wasn't on this, but I was testing it, and the balance sheet wasn't balancing it. It built a financial model for us, and we asked it to find out why, and it kept coming back, and all of 'em will confidently tell you the answer, and it's always wrong. And it came back, and it was 1.3 million out of balance and said, " That's an acceptable variance. That's only 0.3%. That's not how a balance sheet works. There isn't an acceptable variance. So I would love to get your answer to this. I know CFOs are struggling, right? Just like everybody, we're all trying to figure it out. It's moving so fast. I can't think of anything that moved this fast. You think back to computers and yes, it was quick and you started using them, but there wasn't a 30 x improvement that quick with computers. At least it didn't seem like it. I was also a teenager when computers came out, so it might just be my memory, but I would love to know if A CFO asked you how to determine whether their AI programme is creating actual value for the business, what would you tell them? I know a lot of people are struggling. This is all moving so fast; they're trying to keep up, but they're wondering, " Am I really getting an ROI here? Am I getting value?
Guest: Dave Trier (08:11):
That's right. Yeah, no, I get that question a lot and I like to think of it in three layers or tiers, if you will. The top tier, which is obviously what every CFO wants to see is is there direct financial correlation between the AI that I'm using and some benefit, whether it's cost takeout, RevGen or just in general operational efficiency. And yes, all the buzz has been around headcount reduction. They have large enterprises laying off a large percentage of their workforce overall, but there are actually other direct financial metrics. For example, there's reducing the time to market for your product, right? Whether that's a physical product, whether that's a service or software or we work with a lot of large enterprises, especially in the regulated spaces, financial services, pharmaceuticals, can you speed up drug discovery, right? Using ai. And there was actually just a recent publication from the FDA where they are going to allow them to use analysis and let's just say forego clinical trials in certain situations because the data and the AI is so good now that the F FDA is allowing that.
(09:19):
So you have to think about it as not just what you see in the headlines as headcount reduction. There actually are financial metrics around time to market, around sales, and helping to grow your pipeline by X percentage, which obviously leads to tangible impact in terms of your top line or automating back office tasks where in the past you had to go and pay an outside contractor to do, and this is more efficient, cheaper, et cetera. So again, that's the top tier, which everybody wants to get. The second tier is, okay, well we don't have a direct financial correlation, but we know that it's actually providing a lot of value to our internal users or our customers. And that's where you bring in the feedback layer. So you start to collect feedback from your customers, your users, or in the agentic world, who are the overseers of the agents to see if it's meeting the need, did it solve the problem?
(10:09):
Did an agent do its job without having a human intervene? And how often did it do that? How often did I have to intervene? So that's kind of that second layer that the CFOs can start to think about, how do I judge the value is just getting the feedback. Now, the most fundamental layer is just usage. And I know that's not the most exciting thing from a CFO perspective, but if you're dumping a lot of money into it and spending either licenced money or putting resources and infrastructure into it, you just want to make sure the thing is used. So with all of our customers, it starts with just making sure you have the usage understood. Are all the users using it? Are you looking at it? If it's more of a systematic approach, how many transactions is it actually processing? And how is that better than what we're doing today? So if you think about those three layers, everybody should be tracking usage, you should be tracking that feedback across your customers, users, and overseers of agents, but then also at the top level, is absolutely in many cases, you can at that direct financial correlation. So that's the three-tier approach that I typically talk about with executives and CFOs regarding the value that AI brings. And tracking it
Co-Host: Glenn Hopper (11:15):
You said is so interesting to me because right now there's a million studies out there and there's a new one every week or every day maybe it seems like. But it goes back a little bit to the ROI thing, but also just in the brass tacks of talking to companies who are trying to roll out ai, the typical use case I give is a company says, okay, we're going to do some ai. So they get enterprise licences, just throw 'em over the fence to their employees and say, go use AI. I argue all the time that if you're just buying software licences and talking about an ROI, it's like, what's your ROI on Excel if you're just buying? So it's a software expense, but one thing that I'm seeing, and this is if you do just throw a bunch, you get a team account or an enterprise account or whatever and throw it out to your employees when you're trying to measure ROI, the employees may be getting much more productive.
(12:12):
And it depends on the environment that you could be scared to say what they're doing with ai. So these Ethan Moric calls 'em, which are secret cyborgs doing shadow AI in the distance and they're doing what used to take 'em eight hours. I see this with coders a lot. It now takes them an hour and a half and then they can hit the golf course or whatever the rest of the day, or maybe they're doing more work and reporting on it or whatever, but the shadow usage and sort of throwing a bunch of pilots out there, I see so many companies in that right now it's like, yes, we know we have to do some ai, but they don't really know what that means. And I'm wondering from your standpoint, because I think for ModelOp to really be effective or for companies to be effective with ai, it can't just be a bunch of pilots. So, how do they move from just whatever we're doing right now, some pilots and pseudo-organised just tests out there to shifting to where AI is treated like a managed portfolio of investments is the way I like to think of it. And I'm wondering, and again, going back to the CFO perspective, how do you ensure that the portfolio is optimised across the company, not just for their organization? What is everybody in the company doing?
Guest: Dave Trier (13:24):
Yeah, absolutely. And I love that analogy that you have to consider AI just like your portfolio of investments, because it is AI cannot be thought of as a series of experiments. It can't, there's too much cost and risk involved. Obviously, the benefits there as well, but there's too much cost and risk involved to think of it as just an experiment that, just like you would with your investment, needs to be managed with the same discipline that you would manage your portfolio of investments. Yeah, some will fail, some will succeed, some might be a surprise, but that's okay. But you still have the management discipline around how you approach them. And what that means is that at the start of it, it's not just go throw spaghetti at a wall and get a licence for everybody, just hope, right? Build it, and they will come, right?
(14:10):
No, you actually, again, with the proper discipline, you go through the process of just analysing, okay, where are the areas that we potentially could use AI that'll drive the most value through an organization? Which particular processes is AI really suited to improve upon? And then what's that business benefit look like? But then you actually have to, well, cross-reference it just like you would in your portfolio of stocks, your cross-reference with what's the risk involved, right? Is it going to use sensitive data? Is it going to be opening up by firewalls to external parties that could potentially have security exposures? What's the complexity of managing it? Right? And are we prepared for that complexity also? What's the ability to deliver? Especially if it's an internally developed type project where you're creating a custom agent or a custom GPT, if you will. Are we able to deliver that?
(15:02):
And then the last aspect you got to look at, again, analysing it like the discipline rigour of your portfolio of stocks is what the change management is involved, right? Because that's actually one of the often overlooked parts of AI within the enterprise, is, well, what changes do we need to put in place? Are we actually changing some of our business processes? Which is not a bad thing, but it means you have to have the right structure in place to go through that change with your employees, et cetera. So again, to your point around, well, how do I treat AI like a portfolio of investments? It's going through that process, it's going through an understanding of where I can see the most benefit and then cross-referencing that with the risk involved, the complexity, my ability to deliver, and then are my people ready to change? Which is something that, again, the enterprises really need to focus on overall.
Co-Host: Glenn Hopper (15:48):
Oh yeah, a hundred percent. And I mean that's what I'm saying is the change management piece because there's the early adopters out there that this is life-changing. I'm doing everything. I'm running my open claw and I'm doing all these crazy things and I've got agents I've got
Host: Paul Barnhurst (16:05):
That he's talking about himself. Just so you know. Dave, how much do you spend a month on software? Again, AI software. Are we up to 500 yet?
Guest: Dave Trier (16:14):
No comment.
Co-Host: Glenn Hopper (16:15):
We're around that for mind this way. Then there's the other guys though who are still, I miss my paper ledger. I dunno, maybe those are few and far between, but the leadites in the group too. But that said, with the risk, and I mean I understand their position as well, but trying to wedge AI in between all those people, that's one of the bigger parts of the challenge. It's just that soft skill change management piece.
Guest: Dave Trier (16:43):
Yeah, that's right. And again, back to the original conversation we had about, well, how do I get the ROI out of it? You actually do get the most, there's the most opportunity for that transformational ROI that everybody's expecting when you do enhance your processes, transform your processes. So totally agree. But that does involve the change management. So as long as, again, you have the discipline to plan ahead and not just throw it out there and hit and hope as they say, right? But you have, again, that particular discipline, and we'll talk about this I'm sure in a little bit about how do we go from a one-off kind of cottage industry of let's have pockets of innovation, everybody's doing it different to something that's more industrialized, more industrialized delivery of AI, as I call it from, I've got an idea through all the way through those different rigorous processes, the financial understanding, the complexity, the technical, et cetera, to actually being using that AI itself. So that's part of that discipline that we'll talk about. I'm sure here
Host: Paul Barnhurst (17:39):
Ever feel like your go-to market teams and finance speak different languages? This misalignment is a breeding ground for failure in pairing the predictive power of forecasts and delaying decisions that drive efficient growth. It's not for lack of trying, but getting all the data in one place doesn't mean you've gotten everyone on the same page. Meet Q flow.ai, the strategic finance platform purpose-built built to solve the toughest part of planning and analysis, B2B revenue QOO quickly integrates key data from your go-to-market stack and accounting platform then handles all the data prep and normalisation under the hood. It automatically assembles your go-to market stats, make segmented scenario planning a breeze and closes the planning loop, create airtight alignment, improve decision latency, and ensure accountability across the team.
(18:47):
As you said that, it made me think when companies start right, they're going everywhere a mile a minute, they don't have good processes. It's almost like every company's had to start with AI and figure it all out. Yes, there's governance frameworks, there's data frameworks, but this is different. We haven't had a tool that can spin things up this quick and is probabilistic, it can be wrong, but yet has such tremendous value. We want to use it everywhere. And so nobody has, I shouldn't say nobody, but very few people have really deep established processes in this space. There are still figuring it all out and what it means. And so it's fascinating because you give that example the scaled company, and it makes me think almost everybody on AI to a certain level has kind of had to act like a startup and figure it out.
Guest: Dave Trier (19:34):
It has. And frankly, that's why we exist as a company or software we set out over eight years ago, and we knew that it was going to be the wild west of AI that different teams, different departments were going to start using ai. They were going to take different approaches and use different tools, but you needed to bring some rigour, some discipline to how we can use that at the enterprise level, how we can drive consistency, how do we make sure that the following the processes, how do we make sure that they're not just doing experiments and it's something that turns into tangible value overall. So that's again, a lot of what we help with our customers and our software drives is going from that one-off wild west of AI into something that is consistent, repeatable and make sure that the outcome of which is valuable, profitable, and of course responsible at the end of the day.
Host: Paul Barnhurst (20:24):
So one last question. What role does the CFO play in AI oversight? Obviously they need to make sure it makes financial sense, but as systems start influencing major business decisions, which they're starting to, we're seeing it, how does the CFO kind of keep an oversight there and be involved? What's their role? How should they think about these things?
Guest: Dave Trier (20:48):
Yeah, so there's a couple points. So first off, the CFO should be setting some of the financial parameters for going and using AI across the organization. Meaning that, hey, here is that, as I said, that upfront analysis that needs to be done, what's the business justification? Who are going to be using it? What do we expect as the potential enhancements or improvements to either our current process or to generating new revenue streams? So they should be setting those financial parameters by which you go and judge, Hey, this is where we want to put our effort and resources. Here's what we want to prioritise. Second, they should be helping to define the framework for how you're going and looking at the success. And I'll use that in air quotes, the success of a given AI project that you're working on. And that means going in and understanding, okay, well is this project actually on track or is it vectoring to failure?
(21:44):
Is it going to look like it's going to have a bunch of cost overruns? Does it look like it's going to introduce additional risks that we didn't regionally think about? So having again, that cost benefit analysis, that risk reward analysis, that's the framework that CFOs can help to put into place both upfront from understanding the benefit and the potential costs, but also the framework to go and get the metrics, and I call it instrumenting the process, instrumenting the process so we can judge, hey, is it too much risk, too much cost? Is the benefit in line with what we expect? So those are the types of things that the CFO can help put in place. And then the last thing of course is just reviewing that on a regular basis, they should have an enterprise wide view of all the different AI projects with some of those inputs we talk about to help to rationalize the AI portfolio. We help business leaders to make decisions like, okay, I get it, this looks pretty good, but we got some trade-offs here from a risk perspective. Should we kill that project and instead focus in this area? So that's where the C FFO can be very helpful for business leaders of helping to rationalize their investments, rationalize which ones make the most sense to proceed or potentially fail fast, as I like to say.
Host: Paul Barnhurst (22:54):
Yeah, the rationalization point really hits home. We've had to do that with SaaS software and now we got to do it with AI and projects, otherwise we're going to have AI sprawl everywhere.
Guest: Dave Trier (23:05):
Yep, that's right. The wild west of AI was sprawling over the place and people swiping credit cards and just starting to use AI. And you find out later when you get the bill, right?
Co-Host: Glenn Hopper (23:15):
So Dave, this may sound crazy to you, but as a former CFO, in all of my implementations, I'm selling into the office of the CFO and there are two types of CFOs out there right now. One is the CFO that leaned into data early and from BI to data science and really understood the value of analytics and kind of expanded their FPNA beyond traditional FPA into using machine learning and being data forward and everything. And these CFOs over the years, several of them have clawed away sort of ownership of data. Now, CFOs and their domain expertise most don't have a data science background, certainly not a machine learning engineering background, certainly not a developer background, but because of their familiarity with analytics and because they've been using machine learning, some of these CFOs have taken ownership of AI and for the whole org, not just for the finance group, some of them know what they want, they may be leaning into CFO, but it's not their domain expertise.
(24:23):
They went to business school and got degrees in accounting in finance, and this is a bit of a stretch for 'em. I'm finding right now that sometimes I'll come into a company, we get all the way up to the CFO, they love the idea ready to go forward, and then CIO head of it, CTO, whatever the role is, we'll step in the 11th hour and say, whoa, we're not doing that. And then you have to kind of start all over. And I guess, so two questions here. The first one is, I don't know how many CFOs you've experienced like the former that I described, but beyond the question that Paul just asked you, if you see a universe where depending on the CFO, where A CFO could sort of own orgwide AI implementation in management, I might be trying to hole my own interest into what the CFO should do. And I'm wondering, as someone who does not just work with CFOs, your thoughts on that are?
Guest: Dave Trier (25:19):
Yeah, so what's interesting about this before I answer the question is that I've been doing this for over 22 years. I've never seen quite the political land grab ever. In my experience working with the, again, enterprises Fortune 500, it's insane. I've seen, again, everybody from C-O-O-C-I-O-C-T-O, data analytics, everybody, governance, they're all trying to get the land grab buyer. I want to own ai. So I'll just put that as a starting point, if you will. What I would say just in council, it depends on your organization. For some enterprises or even just some small medium businesses, you may not have that full gamut of the cs, the CIO and C, DO and C, all of those. And in that case then, yeah, absolutely it could make sense for the CFO to own it because you may not have a potentially dedicated CDO or Chief AI officer even at that.
(26:19):
So that point would make sense. However, what I would say is that it always is a partnership. It has to be a partnership with other executives, the CIO, the CIO, like securities always involved if there is a head of data that they're going to be involved as well. So it's always going to be a partnership. But again, I can see to your point for certain companies where you don't have that level of management executive level tier there that yeah, the CFO could own it because it is something that is one of the biggest investments and will be one of the biggest investments in your company going forward. So I could see that, but I would just always counsel, always start with the partnership. Otherwise, exactly what you said will happen. It's going to, you'll get to the very end, you get to the finish light and somebody pops up said, Nope, can't do that for security reasons or whatever it may be.
Host: Paul Barnhurst (27:09):
That never happens.
Guest: Dave Trier (27:10):
No, no,
Host: Paul Barnhurst (27:11):
Never, never mixing it. I've never had those projects die on the vine because of politics. Only everyone ever. I felt like it's a really good point. Appreciate that. Would love to get your thoughts. So for our audience listening, CFOs, finance professionals, what's the one thing that they should be doing to improve the ROI for their companies? If you were to offer 'em one piece of advice, what should they be thinking? What should they be doing?
Guest: Dave Trier (27:38):
Yeah, the one thing I would say is, as I said, get away from the wild west and turn something, go into more of a repeatable, as I say, industrialized AI delivery process. And even if you're a small company, it doesn't matter. You can put in the right process, the right rigour, the right discipline to make sure that you're upfront evaluating and prioritising what makes the most sense that you are making sure that each of the steps in the process are being followed. Again, what we just talked about, making sure that your data, your security, your it, your finance, architecture and security teams are involved. So making sure the processes are being followed such that when you get to the point where say, great, hey this looks really good, let's start using it, you're not going to get blocked. You're not going to told to be go back to go and don't collect $200, that you'll be able to go from idea to actual usage rapidly, effectively with the right level of rigour and oversight so that all parties across all different of the different teams are happy with what has been done.
(28:40):
They can sign off on it and you can start to use it. But by doing this, by just establishing that, like I said, industrialized process, you can be assured that every team and every user across the organization, whether it's a data scientist or a back office HR person, that they're going to go through this process and you know, can trust and you can sleep well at night, that this is the right thing to do for a company financially, organizationally, and from a risk perspective overall. So that's what I would say is counsel is it's not too early to go and put in that, again, process and discipline so that you can trust what's being put through the funnel, but through the factory, if you will.
Host: Paul Barnhurst (29:19):
As I think back on our conversation and mayor here, we'll go to our kind of fun AI section. I couldn't help but think I was listening to you talk. So much of this comes down to process and change management. And that's so true of so many software tools, right? Implementations often fail. Rarely does an implementation fail because the technologies can't do it. Most technology can do it. There are exceptions obviously, but most of the time it fails because we don't have a rigorous process. We didn't do a good job in selection or we didn't manage change at all those things. And it feels like a lot of what you're saying, it's kind of a repeat, but with ai, yes, there's some different governance, there's different things we need to consider, but just having good process to bring things in and thinking from a company and documenting and managing change management gives you a higher chance of success regardless of catch GPT, clawed, whatever the tool might be. Am I missing something or It
Guest: Dave Trier (30:16):
Is a lot of it. The only thing I would say is that AI has more complexity in not only the technology but also the number of parties that are involved. If you think about software delivery, normally you have your business stakeholder, you got your software development team and you got a production team, great, but that's kind of it. It's those three with maybe a data team, but with AI you have to pull in at least 10 different people. On average, it's 10 different teams that I see across it. Legal risk compliance, data security, architecture, project management, production support. So because of the nature of AI with inherent risk that's involved, there are more stakeholders involved. So your process has more people involved, but then also it often touches more systems too. On average, I see anywhere between eight and 12 different systems that one AI solution is touching. So there's just a level of additional rigour that has to be put into place because you're touching more teams, more systems as part of it. But yes, just in terms of use, good process, absolutely. We can learn from the days of DevOps if you will, but AI has again, that slight nuance you have to up your gain as it relates to risk and security and governance, et cetera. No,
Host: Paul Barnhurst (31:28):
That's a really good point. The whole probabilistic and the fact that the same tool can be used for everything across the company. That's right. In just so many crazy ways for all kinds of stuff, unlike we've seen before. So the complexity is definitely a good point. So thank you for that. Alright, we're going to move into our personal section. So how this works is we use different AI tools. This week I fed it into Claude. We take the questions from the episode your, your LinkedIn profile, whatever it can find on the internet, and we ask it to come up with 25 kind of personal, unique, a little quirky. I put that in there actually list of 25 questions. And so we we're going to ask two, I take one approach, blend takes another.
Guest: Dave Trier (32:13):
Okay,
Host: Paul Barnhurst (32:13):
You get one of two options. I can use my random number generator to pick a number between one and 25, or we can keep a human in the loop and you can pick a number between one and 25 and I'll ask you that question.
Guest: Dave Trier (32:24):
Oh, okay. Well I'm an engineer. Let's go with the math. Random, random number generator.
Host: Paul Barnhurst (32:31):
Alrighty, we get number two. So let's see what we have. So it says EE major, so electrical engineer major turned ai, CEO. Was there a specific moment when you realised you were never going to design circuits for a living and that your future was in software and data?
Guest: Dave Trier (32:52):
Yes. That's a great question. And there was a moment actually, and what it was is that out of college, I went and joined an RD lab, which was fantastic, but in the very first week of work, I was building this large scale multi-user touchscreen and we had to just do a transfer of data and I had to go and solder an RS 2 32 adapter in order to be able to transfer from the huge multi-user touchscreen into the actual, at the time it was a computer with Nvidia chips, which is really cool. I used Nvidia 20 something years ago. So that was the point where I said, okay, I don't think I want to do this electrical engineering. I'd like to go more down the route of developing product solutions that are more in the software space. But that was again, a couple months out of college.
Host: Paul Barnhurst (33:40):
So it sounds like it was early.
Guest: Dave Trier (33:41):
It
Co-Host: Glenn Hopper (33:42):
Was pretty early.
Guest: Dave Trier (33:42):
Alright.
Co-Host: Glenn Hopper (33:43):
I know we're pressed for time. So what I do on mine, I take the human completely out of the loop. I figure AI generated the questions and we'll let AI figure out which one to ask. So I just run clog to do that. Okay, so Paul, well you got number two. I got number one. This happened in our last week too, where the random numbers are getting to, they're skewing towards the front end here. Oh, you studied electrical engineering at Notre Dame. Are you a diehard fighting Irish football fan and does game day still rearrange your fall schedule?
Guest: Dave Trier (34:15):
Yes, and yes, absolutely. I've been a diehard Notre Dame fan pretty much my whole life and every Saturday that's what we do. It doesn't matter if we've got kids' stuff going on, it's like, okay, arrange the kid's stuff around it if there's really a kid's sports game or tournament whatnot. I've got my phone, which is the beauty of the mobile era, so
Host: Paul Barnhurst (34:35):
I only have one bone to pick. You didn't play us in the bowl game. I'm a BYU fan. Ah,
Guest: Dave Trier (34:40):
Okay. You guys back
Host: Paul Barnhurst (34:41):
Out. We do get to play next year though, but I was looking forward to that.
Guest: Dave Trier (34:45):
I know. I'm looking forward too. Yeah, I'll be there. I can't wait. I've never been to BYU stadium, so I'm looking forward to it.
Host: Paul Barnhurst (34:51):
Oh, you're going to the game in Provo?
Guest: Dave Trier (34:53):
I sure hope so. I think tickets are going to be hard to get.
Host: Paul Barnhurst (34:55):
I live here, so I'll have to grab lunch. Let me know when you come.
Guest: Dave Trier (34:58):
Let's do it. Yep.
Co-Host: Glenn Hopper (35:00):
Well, Dave, I know you've got a hard stop coming up. Really enjoyed having you on the programme and I'm definitely beyond the podcast. I'm definitely checking out Model Lab because I see a lot of applications for that. So thanks again for coming on.
Guest: Dave Trier (35:12):
Thank you. Love the conversation. Really appreciate you having me.
Host: Paul Barnhurst (35:15):
Thanks for listening to the Future Finance Show. And thanks to our sponsor, QFlow.ai. If you enjoyed this episode, please leave a rating and review on your podcast platform of choice, and may your robot overlords be with you