How CFOs Can Reduce Manual Reporting and Increase Strategic Impact Using AI with Brian McGowan
In this episode of FP&A Unlocked, Paul Barnhurst talks with Bryan McGowan, founder of Alpyne, about what great FP&A looks like, how technical skills can improve finance work, and where AI is actually useful for CFOs and finance teams. Bryan shares how his background in engineering shaped his approach to finance, why he built Alpyne to solve real workflow problems, and how AI can reduce manual work so finance professionals can spend more time on analysis and business partnering.
Bryan McGowan is the founder of Alpyne and a CFO with a rare mix of finance and technical skill. After spending a decade leading finance in tech, he launched Alpyne in 2021 to solve a problem he had lived himself: great CFOs were spending too much time on manual work and not enough time on strategic thinking.
Expect to Learn:
What great FP&A looks like in practice
Why accounting fundamentals still matter
How coding and data knowledge help finance professionals
Why Bryan built Alpyne to solve his own CFO workflow problems
How AI can help save time in finance roles
Here are a few relevant quotes from the episode:
“Models are never right, but sometimes they're useful.” - Bryan McGowan
“If you can create a structure, if you can provide a framework that allows for supports effective decision making across the organization, I feel like you're doing your job.” - Bryan McGowan
Bryan explains that FP&A is not about predicting the future perfectly. It is about giving the business a useful framework to make better decisions. He also shares that AI works best when it is used on process-heavy work like cleanup, mapping, reporting, and system workflow.
Follow Bryan:
Primary Alpyne website: https://getalpyne.com/
LinkedIn: https://www.linkedin.com/in/bryan-mcgowan-8572644/
EarnYour CPE Credit
For CPE credit, please go to earmarkcpe.com, listen to the episode, download the app, answer a few questions, and earn your CPE certification. To earn education credits for the FPAC Certificate, take the quiz on earmark and contact Paul Barnhurst for further details.
In Today’s Episode
[02:55] – What Great FP&A Looks Like
[04:03] – How Bryan Uses AI Today
[08:26] – Why Technical Skills Matter in Finance
[10:47] – Why Bryan Built Alpyne
[14:24] – Using AI in Finance Workflows
[17:49] – Building Client Context Into AI Tools
[21:44] – Checks, Controls, and Trust in Finance
[24:26] – A Real Finance Problem Solved With AI
[29:04] – Advice for Finance Professionals Starting With AI
[33:12] – What Holds Finance Teams Back From AI
[35:31] – The Most Important Skills in FP&A
[40:43] – Better Business Partnering and Closing Thoughts
Full Show Transcript
Host: Paul Barnhurst (00:00):
Welcome to another episode of FP&A Unlocked. Are you tired of being seen as just a spreadsheet person? Will others get a seat at the table? Well then welcome to FP&A Unlocked where finance meets strategy. I'm your host, Paul Barnhurst, the FP&A Guide. Each week we bring you conversations and practical advice from thought leaders, industry experts, and practitioners who are reshaping the role of FP&A in today's business world. Together we'll uncover the strategies and experiences that separate good FP&A professionals from great ones helping you elevate your career and drive strategic impact. Today's guest is Bryan McGowan. Bryan, welcome to the show.
Guest: Bryan McGowan (00:42):
Thanks, Paul. Good to see you.
Host: Paul Barnhurst (00:43):
Yeah, excited to see you. So we get a little bit of background about Bryan and then we'll jump into things. Bryan McGowan is the rare CFO who can read a balance sheet and build an API integration after a decade running finance and tech. He founded Alpyne in 2021 to solve a problem. He'd lived great. CFOs were drowning in manual work instead of doing strategic thinking finance might be in his blood. His grandfather was commissioner of the IRS. I don't know if I should congratulate him or feel sorry for him
Guest: Bryan McGowan (01:16):
And IRS is better days, I guess.
Host: Paul Barnhurst (01:18):
Yeah, but his approach is decidedly modern with degrees in electrical engineering and an MBA from Cornell plus four years at Intel. He combines financial rigour with technical chops in the AI era. That means building platforms that handle the repetitive work so CFOs can focus on what matters. Love the background and welcome again to the show. Excited to get to chat with you.
Guest: Bryan McGowan (01:43):
Yeah, likewise.
Host: Paul Barnhurst (01:45):
So we start pretty much every episode with this question. Always love to see how different people answer it. From your view, what does great fp and a look like? How would you define it? So
Guest: Bryan McGowan (01:55):
I don't know who this quote is actually attributed to. I've seen it associated with different people, but there's a mantra that I like to live by, which is that models are never right, but sometimes they're useful and it's something that I try to apply across the board, whether technology or CFO work. Basically what that means to me is I see the CFO's role or an FBA person's role as providing an analytical framework for the entire organization. It's not just with CEO, it can be sales leaders, marketing leaders, operations, et cetera. And so if you can create a structure, if you can provide a framework that allows for supports effective decision making across the organization, I feel like you're doing your job.
Host: Paul Barnhurst (02:36):
I like that. So kind of providing that framework and support across the org and the quote, the least one I always think of. There's one that was by George Box who was a physicist. He said, all models are wrong, some are useful. That's probably some genesis of that. I've seen different versions of that float around, but that's one that's always like, I joke, it's like the same thing. All data is messy. Some actually has value.
Guest: Bryan McGowan (03:03):
Yes. If you can align around the notion that you're not trying to predict the future, you're trying to imagine versions of the future and is there a way to feed that back into contemporaneous decision making? That's the joke. Yeah.
Host: Paul Barnhurst (03:16):
We don't predict the future. You're not a soothsayer?
Guest: Bryan McGowan (03:19):
Not yet. I mean, there's all this news about anthropics new model. We'll see what that can do.
Host: Paul Barnhurst (03:24):
When does that drop?
Guest: Bryan McGowan (03:25):
Well, they're saying they're not going to release it because it's too powerful and it's going to hack into government systems. And so they're trying to put together this consortium of companies to figure out how to regulate this thing, which, how much of that is marketing hype? I don't really know.
Host: Paul Barnhurst (03:40):
Interesting. I only saw one link that article about data security concerns with the new model. But either way, it's amazing how quick it's all moving. And that's what we're going to talk quite a bit about today. So our audience knows we're going to dive into AI and talk about some of the things that Bryan's doing in that front. And what I'd love to know, just start off, what's the number one thing in your job you're using AI for today?
Guest: Bryan McGowan (04:03):
My job is a lot of different things. So it's software development, continuing to build and iterate platform and that was my path into ai. It was kind of a natural first application for language models. And so I've been using various models and tools for software development for 18 months now. Going back the original cursor release, so that side of the house a tonne. And then over the last six to nine months, I've started to embed integrations throughout platform that I use daily as a fractional CFO and that other CFOs use. And so have really tried to do it at ground level around a lot of different finance operational functions.
Host: Paul Barnhurst (04:46):
Got it. So obviously you're using Cursor a lot on the development side for your software.
Guest: Bryan McGowan (04:50):
Not anymore. So it started with Cursor. That was kind of really the first massive option tool on the development side. I started using Claude Code, I dunno, maybe a year ago. And it really took a step function at the end of last year. Really incredible what you can do with Claude Code specifically. I haven't tried some of the other tools and that's kind of my primary driver in terms of software development.
Host: Paul Barnhurst (05:13):
Yeah, I know four six took a huge leap a few months back. It definitely, it feels like all we've heard now is Claude the last few months. Everywhere I turn is something about Claude.
Guest: Bryan McGowan (05:22):
Yeah, yeah, it's amazing what Philanthropics been able to do just in the last six months. I've
Host: Paul Barnhurst (05:26):
Always liked philanthropic. It's always been one of the models I've used the most, but I can see the change over the last few months, so.
Guest: Bryan McGowan (05:33):
So everything in my platform that necessitates an AI integration, whether it's analytics, transaction classification, driving actions within the platform, that's all the backend is an integration with philanthropic over their API.
Host: Paul Barnhurst (05:48):
Sure. Now I know a lot of people definitely use their API. So I'm curious, stepping back before we get a little deeper in ai, you started your career in engineering before going back to school and getting an MBA and switching to finance. So what motivated the switch?
Guest: Bryan McGowan (06:03):
I was a Lego kid, math and science, and I started my career at Intel, which recent struggles withstanding was a really fascinating company to be at in the early two thousands. I got really interested in their history. If you read Andy Grove's biography, there's a couple other books written about Intel. They were obviously one of the original Silicon Valley companies and Paragon of American Innovation and Manufacturing. And they made some really interesting strategic decisions in the eighties to get out of memory and focus on microprocessors. And the whole initial Intel Inside campaign was one of the most brilliant campaigns of all time. And I remember those Super Bowl commercials with dancing in the funny suits and everything.
Host: Paul Barnhurst (06:45):
I just remember seeing every computer slapped on Intel inside. I mean, when you're just a part inside something and you can get them to slap that on every single one somebody's buying, you've done something right.
Guest: Bryan McGowan (06:57):
And so I got really interested in the intersection of technology and strategy and decided that's kind of where I wanted to be. Not even necessarily finance, just how to position technology, how to grow technology businesses. So went back and did an MBA after working for Intel. Really some of my favourite classes in business school were accounting, like you mentioned, it must be in Metallo somewhere. And I ended up after business school working for a number of, I'm from the New England area. I was living in Boston, worked for a number of companies at come out of MIT and Harvard and it was sort of organic. It was kind of well, you're the only person in the company with an MBA. You can run the books too and build the forecast our investors and our board and everything. And so it's kind of how I fell into it. So
Host: Paul Barnhurst (07:43):
You basically and told because you had an MBA and made sense for you to do the finance stuff and fortunately I ended
Guest: Bryan McGowan (07:50):
Up enjoying it quite a bit and I've spent the last 15 hours at CFO in various capacities.
Host: Paul Barnhurst (07:57):
Well, that's a good thing. You enjoyed it obviously, and you can see how it's shaped your career, but it's funny sometimes how we get into things. I've told this before, I got into fp and a basically because I gave me a promotion, oh, you're going to pay me more. It's a finance role. Okay, I'll take it. You have a technical degree, I know you like to code. So how has coding helped you in your finance career? So not so much in the software platform, but just as a CFO, either being able to code from a thinking standpoint or in work. How have you seen that help yet?
Guest: Bryan McGowan (08:26):
Yeah, so even going back to those days with startups back in Boston, I've always been building automations and API integrations and trying to make my job a little bit easier. And that's evolved a lot over the last 10. APIs were a relatively new thing 10 years ago, and various of the accounting platforms have supported them to differing degrees, say what you want about Intuit. They've really supported the kind of open ecosystem. And so I've always kind of played around the edges and specific workflows, make my job easier, getting data in and out more efficiently, building custom integrations, things of that nature. But honestly, one of the biggest advantages has been understanding the backends of accounting systems and other kind of business systems. As you may or not know, accounting was invented in the 13th century in Italy and it's what was laid out at that time that has driven the object models and schema behind QuickBooks and NetSuite and Intacct. And so they're not really conventional softwares in a lot of ways. And understanding how they operate under the hood has really helped me as A CFO.
Host: Paul Barnhurst (09:36):
Got it. Yeah, I can see we're definitely helping with getting at data and understanding how the data flows and things like that. Yeah, like you said, double entry, bookkeeping and recording, that is not typically how you would see most databases record transactions, both software.
Guest: Bryan McGowan (09:52):
Correct. And it is really exciting what's happening now with some of the new entrants in that space and how they're reconceiving of how an accounting system can and should be structured.
Host: Paul Barnhurst (10:01):
No, we're seeing a tonne of activity. I've been tracking some of it and I think I'm up to the tools about a list of about 10 newer tools, kind of ERPs accounting software. And between all of 'em, they've raised, I don't know, 750, 800 million. It's a big opportunity. You don't need a huge amount of market share when ERPs and accounting software are not cheap.
Guest: Bryan McGowan (10:22):
Well, yeah, when the addressable market is the entire world, yeah, it's a big sample. Every
Host: Paul Barnhurst (10:26):
Business in the world needs it and it's a core software, not a bad business. If you can get a little percent. I know a few years ago you started Alpyne, you're a fractional CFO platform. So what really motivated you to start the platform? Was it to meet your own needs, you felt like the market needed it or how did that journey happen?
Guest: Bryan McGowan (10:47):
It very explicitly to solve my own problem,
Host: Paul Barnhurst (10:50):
That's kind of what I figured.
Guest: Bryan McGowan (10:51):
That's
Host: Paul Barnhurst (10:52):
How it works.
Guest: Bryan McGowan (10:53):
Even initially with the notion that offer it to other CFOs, it was really just I was practising and practising practical CFO. You've got somewhere between six and 12 clients and a given time, and inevitably they're all on some combination of back office systems. We're principally interested in the ledger, but CFOs, we want to know what's going on in the CRM and in Stripe and they got an AP system and what have you. And so in part just kind of out of my own laziness, I just got tired of, I do not want to log in and export another trial balance and paste it into a model and make sure that are working and it is not how I want to be spending my time. And so at the time, most of these systems had APIs, some better than others like I mentioned. And so I started there, started just building integrations, building uniform data warehouse across my client portfolio and then just different widgets and applications on top of that. And it was eventually turned into this, you can think of it as kind of an operating system for fractional CFOs and really accelerated in the last few months with my ability to develop quicker all of these new interesting integrations and plugins and everything. But when it comes down to it, I still like being a fractional CFO, working with entrepreneurs and business owners and understanding it in business models. And I'm always going to be the first customer.
Host: Paul Barnhurst (12:22):
I love that. And some of the best products are ones that we develop to solve our own needs and then it's just scaled from there. We understand the pain point and we see the benefit, otherwise you wouldn't have built it.
Guest: Bryan McGowan (12:32):
I've been involved with projects where you're trying to solve something where you're not the user and you're just sitting in a room with a laptop on a whiteboard and trying to imagine what people's issues are and it's a lot less fun.
Host: Paul Barnhurst (12:43):
It's fun when you're solving your own problems. I imagine the first time you had the APIs all working, you get the report, you're like, okay, that was just all automated. Now all I got to do is this, this, and this. Instead of 20 steps, it now becomes five or whatever.
Guest: Bryan McGowan (12:57):
To be quite honest, whether you're a fractional CFO or your real value is strategic partner, you have an intuitive sense of business and finance and accounting and the more time you're spending moving data around and manipulating it, the less you're spending in those other areas.
Host: Paul Barnhurst (13:17):
Well, that's just it. When I was a director of fp and a, I worked at a company that we had systems that were very dated, one of my director of fp and a roles, and I spent a tonne of time cleaning up data because it was the only way I could get to decent analysis and it was painful. I mean, I spent a tonne of time in Power Query more than I'd ever want to admit we made a tonne of progress, but it's like I'd much rather be figuring out this commission plan or the strategy for this customer or talking with the sales person and seeing how we can grow this instead. I'm trying to figure out why we have 10 different versions of the same product with different names that don't match. I think every fp and a professional's been there, and that's some of the exciting stuff of how AI can help with that. You still got to have a decent data foundation, but it can help you with cleaning and matching data. I'd love a little bit of your AI journey. I'd love to kind of talk a little bit about that and more how you're using it in your fractional CFO business versus decoding side. I know you've done a lot of coding. I may ask a question there, but would love a little bit of your journey, how you started and what you're doing today?
Guest: Bryan McGowan (14:24):
Yeah, there's a lot there, quite a bit there. I mentioned starting on the technology side, but with respect to finance and accounting, my approach has been pretty organic As of middle of last year, I had a pretty good sense of what AI can do and limitations and hallucinations and all of that just from having experience on the development side. I'm like, all right, so how do I start integrating this both as a user platform in a scalable way and things like analyse this data and give me trends that's table stakes. I was much more interested in what are processes, procedures, workflows, things like that where I'm spending time that's not really creating value to me or my client. And I'll give you an example. So when I get involved with a client, one of the first things I do is get the charter accounts organised.
(15:14):
And as I'm sure you'll come into a business, maybe they started with the default charter accounts and QuickBooks and there's been kind of chart creep over the years and it's not well organised and maybe they use account numbers, maybe they use other kind of dimensions, but it's never the same. And it's always challenge. If you're going to stand up a budget, you want a degree of an organization. And the traditional way of doing this is you have a lookup table in your model workbook where you say, I'm going to group these six accounts into a t and e line. I'm going to set a budget that group rather than, I don't want to set a budget against airfare and cars and hotels, and maybe you do in most cases. I don't professional services, there's some convention to this. And so that's always the first thing that I do.
(16:01):
This is something that AI is quite good at, is it can parse natural language, it can figure out groupings and organizations and things like that. So one of the first AI tools I built into my application was a chart of accounts mapping an organiser. And as with most AI tools, it's not perfect, but it'll get you 90% of the way there. Like, all right, it did a pretty good job. I've got these kind of nominal logical account groupings. I tweak a few things, now I have this foundation for the company. I can go and do my reporting, do my model, everything else, and I'm not having to manage the Z lookup table. And inevitably somebody adds an account and didn't tell me and Oh crap, I've got to fix that. So I've tried to find kind of point areas in the sort of natural progression of A CFO engagement where I can start hooking things in. So
Host: Paul Barnhurst (16:56):
As I'm hearing is kind of the workflows. And the first one is obviously every company you go in the chart of accounts is typically a mess. How can I organise that in a consistent way that makes it easier for customer after customer? Take the 90% of take the standard chart of accounts from QuickBooks and say, okay, oh, that's going to get mapped here. Here's some other ones I see a lot, map 'em here, use AI to make assumptions for that, that it can't map. And then you review the final table
Guest: Bryan McGowan (17:26):
And that's a process that maybe used to take me three or four hours and I only need to do it once, but if you're onboarding a new client once a month or whatever it might be, now I can do that and three or four minutes and then I'm onto the next task.
Host: Paul Barnhurst (17:41):
What are some of the other kind of workflow things? You mentioned that one. Where'd you go from there on the workflow front? What are the other areas you're looking at for ai?
Guest: Bryan McGowan (17:49):
One of the things I've been working on lately and hoping to have in production next week is, so the genesis of Alpyne was wrangling data is you have all of these different systems, it's not conformed. And so my initial task with platform was like, let's get everything structured in uniform. So I said I can see it into whatever kind of workflows that I want to build. The other challenge I have as a fracture CFO is you're managing all of these relationships. Maybe you have six or eight or 10 clients and different personalities, different needs, different schedules, different requirements. Typically you're managing communications across a lot of different platforms, email, slack, text, in-person meetings, virtual meetings, meeting notes. They have all these channels of interaction with your client. And the problem with using AI point solutions is they don't have all of that context. You can see the snapshot of a p and L into Claw Excel.
(18:46):
It'll build a really nice looking model for you, but it's not getting the data updates, but it's also not really getting all of that client context. So I built this whole kind of client context engine where it's pulling in your emails with a client, it's pulling in meeting notes if you're managing tasks, all that kind of stuff. So in addition to, hey, the model can see a realtime p and l, it can also see your interactions with the clients. So hey, I am worrying on the model. I as a human, forget a lot of stuff. I forgot I had this conversation with the CE O2 months ago and he mentioned this was happening and that's going to have an impact on cashflow or revenue or whatever else. That's really the next layer for me and for my users is, okay, how do I pull really everything about that relationship into whatever the current issue or conversation is?
Host: Paul Barnhurst (19:41):
That's incredibly handy, right? The more context you can give AI one, the more time it saves you, but even more important than the time is the better the answer you get.
Guest: Bryan McGowan (19:52):
Yes, that's what it's all about. And I mean this is one of the limitations of language models is it will very confidently tell you something. You don't have a great sense of confidence level. And so the more that you can layer onto it and give it structure and give it background, the better the result's going to be.
Host: Paul Barnhurst (20:11):
One of my favourite, when you talk about confidently with the answers is we had an LLM that try to get the balance sheet to balance. It wouldn't balance. We were testing it in Excel and Excel agent, we asked it to keep looking and said, I found this. And it got a little closer, got a little closer and it was like, I dunno, 1.3 million off on a billion dollars. And it said that's only a 0.3% variance. That's not how a balance sheet works. You know what the word balance means.
Guest: Bryan McGowan (20:35):
So I have a whole kind of system prompt layer that's very finance and accounting specific and you literally have to include things like a balance sheet must balance, provide the source whenever you're quoting a number.
Host: Paul Barnhurst (20:48):
Yeah, we were talking, well a good example of that is on a webinar yesterday and Glenn Hopper shared, if you ask an LLM to build an amortisation table, not use code, but just build it. It'll build it and at first glance it'll make sense. The problem is it's just building it off probabilities. It's not doing any math. It's been trained on 50 amortisation tables. So it's just trying to make 'em look right. As you start looking at the numbers, you're like, oh wait, that one doesn't make sense in every single one half this whole thing's a mess. How do you manage that with ai? I know you're doing a lot of coding, so I imagine anything that needs to be deterministic or you're writing your own Python code, having the AI generate some code. I know there's a balance, right? There's certain things, probabilistic is fine. That chart of accounts make an assumption, your monthly commentary, whatever, there's lots of areas where I don't care what word they used, I get the context, if the math is wrong, you got a problem.
Guest: Bryan McGowan (21:44):
I learned in the pre AI world as a CFO that I'll say, I don't need help from AI to make a balance sheet that doesn't balance. I can mess that on my own quite easily. And so I really got into the habit of building parody checks into these workflows, into this output. So very simple example is you've got a three statement model. Is the cashflow statement footing to the balance sheet and my building cashflow statement and does the change in cash match the actual difference on the balance sheet? And so in spreadsheet world, the way to do that is you put in a line with a check in some conditional formatting and it turns red and it doesn't. So I mean I still do this. I have my agent building a model, but it's still putting the parody checks visually into the workbook or into the dashboard. So when I'm looking at it, I know what something's wrong, but there's something analogous going on in the backend as well. If it's building a balance sheet, there is a literal parody check in the code to make sure things like balance sheets balances or we have the right valence based on account type debit credit, et cetera. So CFOs and accountants have these analogues kind of on the technical side that I think are really important to develop trust in these systems. Yeah,
Host: Paul Barnhurst (23:13):
I know Claude release skills, some of those things. How has that helped now that it's much easier to give instructions to give detailed even to code, some of it references all those things. How has that helped in the work you're doing with ai?
Guest: Bryan McGowan (23:29):
Maybe this is a politically incorrect analogy. It's kind of like I have three kids, it can be working with toddlers. You got to go with boundaries and rules and kids feel safer when they have boundaries. And I think the AI does as well. And so whether it's it's on the development side or the finance side and anytime it messes something up, which it still does, I mean it's gotten a lot better, but it still does. You can create a rule, you can put that into the context. And on the finance side, that's kind of like the tuning element of this is as me as a user and as other users are getting into it, we stuff is going to happen. It still does hallucinate. It still does. You're just incrementing these kind of guardrails around the agent and the model.
Host: Paul Barnhurst (24:13):
Yeah, I mean I think it's a good way to look at it. I like the toddler analogy, we'll call it politically incorrect, but that's okay. We'll go with it. What's the thing you're most proud of that you've kind of automated a workflow with AI thing you're most proud of that you've done so far?
Guest: Bryan McGowan (24:26):
Oh man. I had a project just a few weeks ago. You brought up amortisation, so it's a different kind of amortisation, but it was a real estate investment firm that was managing, I dunno, 50 or 60 properties and they needed to issue K ones, which is a couple of weeks ago. And nothing about what was coming out of their accounting was correct and very tight timeline. So I think I got engaged business tax is due March 15th. Investors wanted K ones a week ahead of that. I think I got engaged on March 1st, something like that. We eat p and ls by property because we don't know how to do the allocations. We're basically at square one here. And so what I was able to do is first get connected to their account system. So all of a sudden I have all of the data in my warehouse, it's structured so I know what's actually in there.
(25:20):
I know it's wrong, I'm looking at the principal interest on these properties. It's like, huh, why is interest same every month or why are they booking everything to principal for the property or all the mortgages were standard AMD station, but the accounting was just completely wrong. And then I could feed in the PDFs of all of the mortgage agreements and like you were saying, it generated the principle and interest tables so that I could match that against what was actually in the accounting system, generate all the proforma p and ls and then feed the adjustments backing to the ledger. Obviously that was highly spoke project, but it involved some heavy duty finance, a lot of custom code. I could spin up custom interfaces so they could see these kind of bridge p and ls by property and it's pretty cool. I felt pretty good about that. How
Host: Paul Barnhurst (26:11):
Much time do you think you saved being able to use ai, being able to code? Say you had to do that all yourself, would you have hit your deadline?
Guest: Bryan McGowan (26:20):
Oh, no way. In a pre AI world that would've, and without my existing foundation of a platform, that would've taken multiple people weeks to months to do pretty
Host: Paul Barnhurst (26:31):
Amazing. And that's the thing, sometimes people don't realise, yes, AI gets things wrong. That doesn't mean it still can't be really helpful in finance. And that's part of why I want to bring people like you and others that are seeing those huge time savings on is to help people realise that there is a big opportunity here. How much would your analysis work you do now? How much would you say runs through Claude?
Guest: Bryan McGowan (26:55):
I don't think of it as a replacement. I think of it as a support. My philosophy is in an increasingly AI-based world, things like business acumen and strategic expertise and experience are more and more important. And so if I can build systems and workflows that support my ability to be more focused on that. And when you say analysis, it can mean what was the top
Host: Paul Barnhurst (27:25):
Line? That's a very loose, let's say this, how much time do you think in a month on average you're now saving with ai?
Guest: Bryan McGowan (27:33):
I mean again, for me it's two sides of the house on the development side.
Host: Paul Barnhurst (27:36):
Let's talk the fractional, let's ignore the software for a minute. I know you're getting huge benefit on the coding side, but let's just talk kind of that fp and a fractional CFO, the finance work.
Guest: Bryan McGowan (27:46):
I think in terms of fractional CFOs and the amount of time, so if you're spending 30 hours a month on a given client previously with AI and data normalisation and integrations and all of that, I'm probably cutting half of that. So not having to do model updates and produce the analysis and particularly kind of in the initial part of an engagement when you're doing all of that setup. And that's not just ai, that's also just the whole platform, the whole substrate.
Host: Paul Barnhurst (28:15):
Sure, there's other automations behind that and technology. Sometimes I think we think everything's AI and sometimes good power query or a good API or just a good rule-based process is all you need
Guest: Bryan McGowan (28:29):
Across all those things. It's maybe half of the time that I would formally spend on a given client.
Host: Paul Barnhurst (28:35):
So material, I mean it's a substantial savings obviously. So let's talk, I mean obviously coding, you do a lot of coding, you have your own platform. Let's take a minute and talk a little bit about the average fp and a person or kind of average finance. Assume someone who has never coded before, they know Excel, they might know a little SQL or Power Query, but they don't consider themselves a coder at all. What advice would you give them as they're kind of trying to figure out this whole AI journey?
Guest: Bryan McGowan (29:04):
I would say the first thing to kind of develop some intuition around is the limitations and that personality. And this doesn't even necessarily have to be finance specific. I use AI in a tonne of different ways. And so pick one of the models, get a $20 a month subscription and just start engaging with it in, I mean it could be everyday sorts of aspects of your life and your work. And what you'll start to get a sense of is what do people mean by context? So if you had a long running conversation with Chachi BT or Claude or Gemini, what you'll notice is it's gotten better but it still degrades over time. And so you'll start to develop an intuition of how and why that happens and what the implication is. And then also other kinds of limitations, whether it's hallucinations or like you were saying with composing emails or jumping in or just filling in gaps that you didn't really want it to fill in.
(30:02):
And so you'll start to develop this kind of intuition before you can get into finance workflows and things like that. And then beyond that, if you want to start getting more specific, I know you've posted a lot about call for Excel and milli models and things like that. Get practical. That's always the solution for me is pick a problem in your client or if you're in-house somewhere, a thing that you're trying to solve and make sure it's not mission critical. And I just find it's always more realistic if I'm working with real data with a real problem in the real world. And so pick one of these tools and try to solve one thing. And then there's a very meta element to it. AI can be very helpful in helping you understand how to use ai.
Host: Paul Barnhurst (30:48):
Yes, I definitely have noticed that I know what you're talking about on the whole and
Guest: Bryan McGowan (30:52):
Not to be an anthropic homer, but they have really good documentation. There's so many tutorials and stuff out there. I just think Anthropic has done such a great job of supporting the community and the ecosystem
Host: Paul Barnhurst (31:05):
And you're totally fine. Feel free to be a homer. I think right now everybody sees anthropic in the lead if they're using other tools. I mean at least the majority. They've done a really good job the last few months. I'm like you. I've jumped more and more on anthropics even two years ago. I like them best mostly for writing, for writing posts, summarising my podcasts and now I like 'em for a lot more of everything.
Guest: Bryan McGowan (31:31):
And I'd add, even if you don't have a background in software development, it can still be incredibly helpful to have basic intuition understanding around data strips. And you can do this with the chat ai, you can just talk to it about SQL and if you can get access to a database somewhere and just start getting your head around how data is structured, how tabled relate to each other, basic object schema, I think that's really critical foundational knowledge. You don't need to be able to write code, but having this sort of just data facility can be really helpful.
Host: Paul Barnhurst (32:09):
I'll speak to that a little bit. I a hundred percent agree with you. I did my master of science in information management, so there were some data table type classes in there and I did report writing for about a year and a half out of grad school before I moved into a more traditional fp a role. So a lot of SQL and reporting out of complex Excel files and things. But learning that data and okay, when do I write code? When do I build a table to do something has helped incredibly, whether it's Excel formulas, whether it's using Power Query, whether it's just being able to think about how the data needs to connect to get good input out of it. And so I'm a hundred percent with you. I found that kind of that base invaluable for me and I don't consider myself a coder. Yes, I know some S two L, yes, I know some power query, but I know I Python, I've never been one of those people that wants to code. I find the data side invaluable. So I'm a big believer in that. I agree with you there. What do you think is limiting most people from getting more out of ai? What do you think the big problem we need to overcome in finances?
Guest: Bryan McGowan (33:12):
Yeah, so I'd say one element of it is the limitations that we've talked about. There are still hallucinations, there are limitations in how it gets applied, but something that I felt that might resonate with people, it's kind of a generalised anxiety around it, especially the last nine months. It feels like things have changed so quickly and in the Bronx world, even just the last two to three months, I mean I just feel like it's exploded in 2026 so far. And I totally understand how that can be overwhelming and disheartening in a way. Shit, am I keeping up? Am just coming from my job. I feel that on a daily basis. And so I can understand kind of the anxiety and the resistance and what I'd say is all of these people who are posting hype on LinkedIn or I've created a one person, a hundred million dollar company, know these people are any smarter than you are.
(34:14):
And a lot of what they're saying probably it's not grounded in reality. And so don't be intimidated by a lot of what you see on social media or other types of media. A lot of that is not really the real world and with finance professionals in particular in this increasingly automated AI world, there's going to be more and more of a premium on experience and expertise and the ability to have insight. And if you can figure out tools and processes to deal with everything else and you can focus on that part of what you offer, I don't think that's going to go away anytime soon.
Host: Paul Barnhurst (34:54):
Great advice. I particularly like the whole reminder look, social media is curated, take all of it with a grain of salt. Some more than others for sure. But yeah, the LinkedIn stuff, use my prompts and save half a million dollars or whatever.
Guest: Bryan McGowan (35:11):
I replaced my entire finance team with Claude and Zapier.
Host: Paul Barnhurst (35:16):
Alright, so I want to move on. We have a fp and a section where I ask some pretty standard questions, then we have a little bit of a get to know you, some fun questions and then we'll wrap up here. So if I asked you what's the number one technical still that fp and a professionals should master, what would you say?
Guest: Bryan McGowan (35:31):
I'm going to sound sold saying this, but fundamentals and accounting. Honestly, how do financial statements relate to each other? Debits and credit, just the logic that above straight behind everything that we do. And maybe you would consider that technical, I don't know. And anything you do on top of that is going to ultimately relate to that. And so if you're using an AI tool to build a financial model, you need to be able to explain how balance sheet relates to the p and l, relates to the cashflow.
Host: Paul Barnhurst (36:00):
Appreciate that answer. What about softer human skill?
Guest: Bryan McGowan (36:03):
I think the best thing you can do, whether you're a fractional CFO or you're doing fp and a at a larger company, can you inhabit the mindset of the entrepreneur and the business owner or whoever the principles of the business are, what are they trying to optimise for? What are they worried about? What are they thinking about? What do they want visibility into? How do they like to consume information? Some people like visuals, some people like numbers. To what extent can you just sort of empathise with who your customer is, whether it's a client or a boss or a colleague or anything like that? And if you can use that to inform your processes, your work output, I think that's what's going to make you successful.
Host: Paul Barnhurst (36:47):
Which excel mistake that you've made over your career has taught you the most?
Guest: Bryan McGowan (36:52):
I have definitely put balance sheets in front of boards that don't balance. If you're looking at something that financial professionals put together and you see that you can't trust anything else, that person puts in front of you, then you're checking everything. And so for me it's been, and not that it ever got to a state of perfection on this, but it's what are the simple checks and balances and quality control bed into my processes that I'm not degrading that trust.
Host: Paul Barnhurst (37:23):
And I'm curious when that balance sheet doesn't balance, how often was it a formula mistake versus not understanding financials?
Guest: Bryan McGowan (37:31):
For me it was typically somebody added an account that I didn't catch. It is typically processed stuff and having multiple cooks in the kitchen, that's why that was one of the first features I built. It was like I needed better handle on the chart of accounts and I need to know something changes or something's off.
Host: Paul Barnhurst (37:51):
Alright, so we're going to move into the Get to Know You section. I have a few questions I just want to ask, what's your favourite hobby or passion? What are you doing in your free time?
Guest: Bryan McGowan (37:59):
Yeah, so I'm a musician, so this is a small part of the collection. So I'm typically playing in a band at any given time and then I look down the road from you in Utah. And so I really enjoy all the kind of recreation that is offered here. Back country skiing, there's no snow this year, not biking, those kinds of things. Yeah,
Host: Paul Barnhurst (38:18):
I was going to say skiing this year was lousy if you were a skier. Alright. I'm curious to see how you answer this one. If you had to listen to one song and only one song for the rest of your life, what are you picking and why?
Guest: Bryan McGowan (38:30):
My favourite artist is a guy named Jason. He's been around for 10 or 15 years. It's in like the Americana, somewhat folky genre. He's won a few Grammys. He's got a tonne of incredible albums and songs. He wrote one when he was, he might been a teenager when he wrote this. It's called Decoration Day and it's like a Ner album. I mean it's just like legacy of the Civil War and family dynamics in the south and it's just a banger of a song and he played it. I saw him in Salt Lake here last year on the front row and he played it. It was awesome.
Host: Paul Barnhurst (39:04):
Nice. So that's going to be your song. Appreciate that one. What's your favourite food? I
Guest: Bryan McGowan (39:08):
Really like Mexican food. Been to Mexico a bunch of times. First on the list.
Host: Paul Barnhurst (39:13):
So what's your favourite Mexican restaurant for those listening in Utah might visit here.
Guest: Bryan McGowan (39:17):
Tiana or Tijuana two, I mean really can't go wrong and they're around the corner from each other.
Host: Paul Barnhurst (39:21):
That's mine as well. So we just put a plug in for Red Iguana. If you're listening, come sponsor an episode. Alright. What's one business belief you hold that you think most people would disagree with?
Guest: Bryan McGowan (39:32):
There's probably less disagreement on this than there used to be just because of how things are evolving. But I think there's so many paths to building a business. It's very easy if you're living in LinkedIn or Twitter or in certain parts of the country to think that the only way to build a business is to go raise venture capital money and get on this treadmill. And that's such a small slice of the world. And so I really support the notion of just different paths to being successful as an entrepreneur. And I've tried to live this, I'm not VC funded, I still have a services offering that is keeping me going. And I think there's just so many ways to be creative and to inject your personality into building business.
Host: Paul Barnhurst (40:19):
I'm not sure how controversial that is now, but definitely you do. A lot of people have the idea, you have to go that VC raise, capital router you can't build and I've never bought into that. Do I think it makes sense in certain situations? Sure. But for the vast majority of businesses you don't need it to be successful, I think. Alright, so as we wrap up here, last question I want to ask, what advice would you offer to our audience to be a better business partner?
Guest: Bryan McGowan (40:43):
I just re-watched Ted Lasso. Have you ever watched that show?
Host: Paul Barnhurst (40:48):
I'm familiar with it, but I haven't watched it. I know a little bit of the premise of the show.
Guest: Bryan McGowan (40:52):
It's one of my favourite shows and I think it's a great study in leadership and teamwork and it's also really funny and heartwarming, but there, there's a key scene. I think it's the first season where he's trying to impart a message, which is be curious. He meant it in terms of just interpersonal relationships and not being judgmental with people or what might be going on in their lives. But I think that applies here. Whether it's adopting new tools, maybe you're initially intimidated by whether it's getting to know a new client or a new colleague or a, I think the most important thing when you're trying to drive the direction or you're having a role in driving the direction of the business is to be curious as to what am I not thinking about? What am am I avoiding? What questions should I be asking? How can I be as much of a sponge as possible so that I can do my job the highest level?
Host: Paul Barnhurst (41:43):
That's an answer I get a lot. I love that answer. I think it's critical to fp and a, you got to understand what's going on around you and being genuinely curious is the best way to do that. Alright, last thing. If somebody wants to get in touch with you or maybe learn more about Alpyne or the things you're doing, how should they do that?
Guest: Bryan McGowan (42:01):
Got a website, it's get Alpyne.com. So Alpyne with a Y. You can email me Bryan@getAlpyne.com. Bryan with https://getalpyne.com/. I'm trying to presence on LinkedIn. I'm not nearly where you are Paul, but you can find me on there. I'll mention one quick other thing. I just launched another business with my brother in the mental health and recovery world. It's kind a space that's pretty important to us. My brother's an executive at a recovery centre in Connecticut and so we built a payments patient management platform for those kinds of residential facilities that solves a real problem in that world. And that business is called Care Ledger. So you're interested in that space as well. I'll be having some more content around that. Cool.
Host: Paul Barnhurst (42:44):
Well congratulations on starting that. I mean, that's a very important area to have good resources for. We need more of that. Alright, well thank you for joining me today, Bryan. I appreciate you taking some time and getting to chat and enjoy hearing a little bit more of how you're thinking about ai, how you're incorporating it in the work you're doing. Yeah,
Guest: Bryan McGowan (43:04):
Absolutely. Thanks very much, Paul.
Host: Paul Barnhurst (43:06):
That's it for today's episode of FP&A Unlocked. If you enjoy FP&A unlocked, please take a moment to leave a five-star rating and review. It's the best way to support the FP&A guy and help more FP&A professionals discover the show. Remember, you can earn CPE credit for this episode by visiting earmarkcpe.com. Downloading the app and completing the quiz. If you need continuing education credits for the FPAC certification, complete the quiz and reach out to me directly. Thanks for listening. I'm Paul Barnhurst, the FP&A guy, and I'll see you next time.