The Human Financial Modeler: AI Insights From 60+ Global Experts
In this episode of Financial Modeler's Corner, host Paul Barnhurst is joined by Ian Schnoor, Executive Director of the Financial Modeling Institute. They discuss the future of financial modeling in the age of AI, with a focus on how modelers are using AI today, where human judgment still matters, and what governance, ethics, and skill development look like going forward.
Expect to Learn:
Why communication skills are critical for financial modelers
How to build models as flexible systems instead of one-time solutions
How Excel Esports helps students develop real-world modeling skills
How competitions can improve financial modeling education
Here are a few quotes from the episode:
"86% of the council members are using AI for modeling tasks" – Ian Schnoor
"The shift, there will be a shift, of course, from construction to judgement" – Ian Schnoor
Ian Schnoor shares why AI should enhance, not replace, human expertise in financial modeling, emphasizing judgment, governance, and strong foundational skills. He also offers practical career advice for aspiring modelers and highlights why mastering the fundamentals remains essential in an AI-driven future.
Follow Ian:
LinkedIn - https://www.linkedin.com/in/ianschnoor/
Follow Financial Modeler's Corner:
LinkedIn:https://www.linkedin.com/company/financial-modeler-s-corner/
Newsletter - Subscribe on LinkedIn: https://www.linkedin.com/build-relation/newsletter-follow?entityUrn=7079020077076905984Sign up for the Advanced Financial Modeler Accreditation today and get 15% off using the code “Podcast” at https://bit.ly/497oAqW
In Today’s Episode:
[00:00] –Trailer
[03:25] – Ian’s Modeling Journey
[06:20] – Building Modeling Skills
[09:59] – Global Leadership Council Report
[14:40] – AI Use in Financial Modeling
[21:31] – Human Judgment in Modeling
[26:09] – Trust and AI Governance
[30:11] – Ethics and the Future of Modeling
[44:14] – Final Thoughts on Modeling Careers
Full Transcript
Host: Paul Barnhurst (00:00):
Financial Modelers Corner is the world's premier modeling podcast. It is brought to you by Financial Modeling Institute, the world's leading financial modeling accreditation organisation. Welcome to another episode of Financial Modelers Corner. I'm your host, Paul Barnhurst. This is a podcast where we talk all about the art and science of financial modeling with distinguished financial modelers from around the globe. The Financial Modelers Corner podcast is brought to you by the Financial Modeling Institute. They offer the most respected accreditations in financial modeling, and that's why I completed the Advanced Financial Modeler. This week, I'm thrilled to welcome on the show the executive director of FMI in Schnorian. Welcome back to the show.
Guest: Ian Schnoor (00:49):
It is really great to be back here, Paul. So glad to be back and chatting with you again. I
Host: Paul Barnhurst (00:54):
Think this is what, third time we've done it for this show probably. And then count Mod Squad we've done what, 20 episodes together
Guest: Ian Schnoor (01:00):
Now? Yeah, probably something like that. I mean, it feels a little bit weird to be doing this just on your other show. We've had so much time together testing tools on the Mod Squad, but nice to be back on this show as well.
Host: Paul Barnhurst (01:11):
No, I think you may have the most episodes someone's been on across my podcast network outside of myself. Congratulations. It must be something like that. Well, thank you. I'm not sure
Guest: Ian Schnoor (01:18):
If that's a good thing or a bad thing for your audience,
Host: Paul Barnhurst (01:20):
But - I'll let you take that wherever you want. So why don't we start, just go ahead and take a minute and give the audience your background. Tell them a little bit about yourself.
Guest: Ian Schnoor (01:28):
Oh gosh, my background. Well, my background is that I've been effectively professional financial modeler for the better part of 30 years now, my entire career. From graduating with a finance degree, I went on to work in investment banking for six years. BMO, a Canadian bank. I'm in Toronto and then at Citibank as well. In those days, it was a bit of the wild west. This was the late '90s, early 2000s. Modeling was a relatively new discipline at that time and I was doing a lot of financial modeling for M&A transactions and I didn't love the way models were built and I felt like there was opportunity for improvement. So I always took it upon myself not to accept the status quo, but rather to see how we could make them... I always felt that they needed to be beautiful and clean and professional and to tell a story to make my job easier when I was presenting to clients and to make them comfortable.
(02:21):
So I did that for six years. I left and then started a training company called the Marquee Group, which was a financial modeling training business. Over 20 years, I spent about 2000 full days in a classroom. I think I taught 20 or 30,000 bankers around the world. I built up that platform and I sold that company a few years ago to a US private equity firm. And during that time had also was involved with a number of people globally in starting FMI as an accreditation body, a credential body where the difference is people learn financial modeling but through any organisation they want, but they can learn it through our materials, self-study. And then they work really hard to prepare for modeling exams to prove to themself that they've actually acquired the skills. So modeling has been front and centre in my professional career for a long, long time.
Host: Paul Barnhurst (03:09):
Yes, I know it has. And speaking of career, so we have a new section called the Career Corner. We're going to start that before we get into the survey. I got three questions for you today. Great. First one is when did you realise financial modeling could actually be a career path? How far into your career were you?
Guest: Ian Schnoor (03:25):
Well, that's funny you say that. Back in the early days, I never actually thought about this. I don't think that I ever thought of myself as a financial modeler for a long time. That identity, and in fact, I'd love to ask that question to people that are around my vintage. I'm not sure if many financial modelers self-identified as a financial modeler until more recently. I was an investment banker doing investment banking work, which required modeling. And then I was a financial modeling trainer teaching young investment bankers and private equity bankers about modeling. I don't know that anyone in the earlier days was identifying as a financial model. That feels to me like a newer phenomenon. In fact, it's only in the last 10, 15 years where job postings started saying, we want someone that has a finance degree or an MBA or a CFA. And it's only the last 10, 15 years where they say, and strong financial modeling skills.
(04:19):
That did not happen more than 10 or 15 years ago. Now you have firms that have people that are head of financial modeling or they work in the financial modeling department. There were no financial modeling departments 25 years ago. So I would say it's more recent actually where that was an identity element and proud of that identity, but that's evolved over the years.
Host: Paul Barnhurst (04:39):
Yeah. And I appreciate you sharing that and I figured that's what you're going to say. And I think it helps people get an appreciation for how new it is as an identity as it's becoming kind of its own career in some ways. It was just something you did as part of another job, which you still can do.
Guest: Ian Schnoor (04:56):
You still can do. It still can be a critical part of other jobs. I don't know how many private equity bankers or investment bankers or credit bankers today or corporate development professionals identify as a financial modeler. They might identify as their role, but they know. I mean, you can still identify similarly today as in the past. You can be a strong finance professional with excellent modeling skills, but now there's an opportunity to actually work in some organisations as a financial modeler and to have that on your... Well, nobody has business cards anymore, but if you did have a business card, it might actually say financial modeler on your business card. I promise you 20 years ago, nobody put financial modeler on their business card anywhere. But it's nice that that can be a department, a profession, and it should be because it's a critical discipline and a skillset.
Host: Paul Barnhurst (05:48):
Yeah, no, I agree. Yeah, like I said, nobody's using business cards anymore. Maybe it's on their LinkedIn profile. You see that financial modeler for sure.
Guest: Ian Schnoor (05:56):
And it is. And it is on people's financial model as their primary role. And that is also newer in the grand scheme of things.
Host: Paul Barnhurst (06:04):
Totally agree. So that leads to a question. If you were starting a career today, how would you go about getting into financial modeling? If you wanted a financial modeling role, like that title, you know you want to be a modeler. How do you think about that?
Guest: Ian Schnoor (06:20):
Well, I mean, financial modeling is a profession and a career if you, as I said, if you choose to be in a role where that is your self-identified title or if you're in a department and a financial modeling group, but you can just as easily be an exceptional financial modeler if you work in a traditional role. I mean, if you really want to learn modeling, and first of all, I think it's maybe the most important skill that a junior finance professional can become exceptional at is financial modeling. Because as you know, I'm going to say, once you are a strong financial modeler, it elevates you and projects you forward in so many other areas. Having strong modeling skills ends up doing so much because modeling is so multidisciplinary. It helps you in so many other ways throughout all aspects of your career. I'm also going to throw out a bit of a warning and some people will appreciate this and I'm sure some people will not appreciate or not necessarily enjoy what I'm going to say.
(07:22):
If you want to become a really strong financial modeler, the best time to do that and for many people, the only time is towards the beginning of your career because that's when you're given the time and the space to become truly exceptional. If you have been a finance professional or if you've been working for 10, 15 years already, it's not that you can't do it, but you need to carve time. You'll need to really create time. When you become a more mid-level or senior person in any organisation, you're expected to do more people type work, client management, managing teams of people. Financial modelers that have become exceptional often had many, many days, hours, months just to sit in the weeds with headphones, with headsets on either literal or just proverbial headsets, just quietly absorbing. And so if you didn't do that early in your career, you'll have to really find that time to do that later.
(08:25):
But anyway, I do think it's tremendous to have as a skillset.
Host: Paul Barnhurst (08:29):
And I agree with you, and I think that's true of most technical skills. You really want to master them earlier in your career because you start leadership, you just don't get the opportunity to use them as much. Not that you don't get to use them, but you don't have the time to use them like you will early in your career.
Guest: Ian Schnoor (08:45):
And it's like riding a bicycle or it's like playing golf. I mean, if anyone here is a golfer, or think of any sport. If you become, let's say, 40 years old and you've never played a sport, tennis or golf or a musical, you can certainly learn it, but it always feels like it's not necessarily your second nature. Whereas imagine if you're 40 years old and you go golfing with someone who's also 40 and they were a child golfer, they golfed years and years and then they didn't do it for 10 years. But I promise you, when they pick up golf clubs after a multi-year hiatus, they're just going to look natural. They're just going to feel like it's just part of them. And there's something that happens with learning a skill earlier. So the message to your listeners is if you are in the earlier stages of your career, take advantage elevating the skillset.
(09:36):
This is the time.
Host: Paul Barnhurst (09:37):
Great advice there. All right, so let's jump into the purpose of today's episode. The main thing I want to cover is you guys are releasing your finance survey from the Global Leadership Council. So why don't we start with, tell the audience what this survey is, a little bit of how you went about it. Give us the rundown on this.
Guest: Ian Schnoor (09:59):
Sure. Well, maybe I'll even start two seconds before that into what the Global Leaders Council is. As FMI, we are the only credentialing or accreditation body in the world for financial modeling. People learn modeling either with us or from anyone that they want, but they enter the FMI community. They put a date in the calendar to sit a challenging exam and there's no magic to the exam. The magic is that it forces them to practise and practise and practise that quiet time. And by the way, if you're a little bit more advanced in your career, it's a great way to really pick up those skills because you're forced to prepare for an exam. And we view ourselves as trying to be at the forefront of helping to think about the future of the profession. And it's been a wonderful opportunity for me over the years to get to know financial modelers all over the world.
(10:54):
And we thought it would be nice to create a group, a representative group of financial modelers that would help collectively put out thought pieces, leadership though pieces on what do we think the future of the profession looks like. So we assembled a group of 63 financial modelers all over the world. They are in 26 countries. They have a wide range of day-to-day jobs. These are not the best modelers in the world necessarily. They're not the only modelers. It's just a representative sample set of people who do identify for a part of themselves as a financial modeler. You are of course part of the group because you are instrumental in your podcast and in some of the other work that you've done connecting financial modelers. And so the purpose of the council is to generate thought leadership and through a couple of surveys a year that we would turn into reports.
(11:48):
So the first report seemed timely to talk something about AI, obviously that is. But instead of talking about what AI is going to do in modeling, we decided to flip it and spin it and say, no, no, what's the future of humanity? What's the future human role in financial modeling as a result of this new era of AI? So what we did is we surveyed the council. We generated an extensive survey report that went to all 63 council members. We had a 100% response rate. Everyone spent a lot of time on the survey. And from those survey results, we created this report that's coming out tomorrow called the Human Financial Modeler. And it's a very extensive, I think it's a great document. I hope people enjoy seeing the views of these 63 professionals.
Host: Paul Barnhurst (12:35):
Got it. So comes out tomorrow, you surveyed everybody, you put together the white paper. It was on, I like you put a spin on it, not so much AI, but the human side of it. What's the human role going forward? Is that kind of a way to think of it?
Guest: Ian Schnoor (12:53):
What's the human role? Well, we ended up based... So the survey had 30 questions and each question had an optional spot for comments as well. So technically there were 60 places for questions. So we collected a lot of data. Some of the questions were liquored scale, which is sort of like a range of options. Anything from strongly disagree to strongly agree. Some were multiple choice, some were short answer, some were ranges, different ranges. But for every question, the participants had an option to also send us comments. So we collected a lot of information, a lot of data. And we looked and we aggregated all the responses into five categories. And so the report has five chapters. So the first chapter talks about how this group of people is using AI today and the benefits and gains they're seeing from AI. The second chapter deals with where does this collective group believe it's going to be really important for humans in modeling going forward?
(13:58):
Where can we use AI and where is it important that we keep humans involved? The third chapter is all about trust and accountability and governance. What level of new governance is in place, is needed with the combination of humans and AI working together? Then the fourth chapter's on ethics and what sort of ethics is required? And then the last chapter deals with the future. What do we think jobs look like and what does the group think that the future of the profession looks like? So that's how it all plays out.
Host: Paul Barnhurst (14:27):
Why don't you give us some of the highlights? You can either take us a little bit of the executive summary if you want to give highlights of each chapter, but maybe just start and tell us a little bit of some of those key takeaways. Yeah.
Guest: Ian Schnoor (14:40):
I mean, maybe I'll give you a key takeaway from each chapter and I'm excited for people to se the report and share feedback and share thoughts. Again, these are not my views or FMI's views. It's nice because it's global geographic views of a pretty significant number of people, 63 people. So chapter one was all about how is this group? We had a number of questions on the survey of how is the GLC using AI today and how are they using it? How extensively are they using it? And how beneficial are they finding it? Sorry, when I say how are they using it, I am specifically talking about using it for financial modeling purposes. Sure. Of course everyone - We'll be very clear
Host: Paul Barnhurst (15:23):
About that. Not using it to write emails.
Guest: Ian Schnoor (15:28):
Not to plan your next vacation, which is very, very good at that. But how much are you using it for agentic purposes to help you construct audit, review, edit, modify financial models specifically, which is very different we all know now than researching a travel destination. And the key takeaway there is that 86% of the council members are using AI for modeling tasks or have used it in the past year, but adoption rates remain shallow. So using it, but over half of the members have said that it has led to little time savings at this point in aggregate. They're using it, they're trying it. Some have actually given up on it at this point or they're waiting for further enhancements. Only a small number are really gone hard and have built in skills automations and invested heavily in actually having it build and meaningfully contribute to their modeling processes.
(16:39):
Most are still watching and trying it, have not fully immersed themselves in it yet. They're
Host: Paul Barnhurst (16:47):
Not really using it day-to-day to build models, but they're testing it, they're playing with it. They're kind of trying to figure out when does it really make sense?
Guest: Ian Schnoor (16:54):
Yeah. Or using it for audit purposes or checking or trying to build things, but everyone's encountering some frustrations similarly to having whether it's a new associate on your team or a new senior person, that friction and pain point of no, that's not what I meant or that's not what we talked about. Or wait, you already did it right. Why are you changing it back to something else? Everyone's getting a little bit of a scare like, whoa, wait a second. When I last checked it, it was right. Now I'm checking it again and you've changed something without telling me. Oh yes, you're so right. I forgot to tell you that I changed this. I though it would be good if I modified this formula, but I should have told you that I was... Everyone's encountering a little bit of a scare along those lines. Surely you've had those scares yourself.
(17:47):
Well,
Host: Paul Barnhurst (17:47):
My background is in FP&A. I am also passionate about financial modeling. Like many financial modelers, I was self-taught. Then I discovered the Financial Modeling Institute, the organisation that offers the Advanced Financial Modeler Programme. I am a proud holder of the AFM. Preparing for the AFM exam made me a better modeler. If you want to improve your modeling skills, I recommend the AFM programme. Podcast listeners save 15% on the AFM programme. Just use code podcast. Yeah. So this has nothing to do with modeling, but just today and I was doing something for my website and I made a bunch of changes and I thought we were just planning and it's like, oh, okay, undo that one. You completely stuck that in the wrong section. So yeah, we've all experienced it whether it's in modeling or something else where it does something where you're like, wasn't expecting that.
Guest: Ian Schnoor (18:50):
No. And it's kind of crazy as you sit through it. Part of your brain is thinking, oh my God, this is amazing. It's incredible. And then the other part of you, you might have felt wants to throw something at your screen and say like you're yelling at an intern saying, why did you do that? We didn't talk about that. Or you didn't tell me that or you edited something without informing me. You've had that experience, truly
Host: Paul Barnhurst (19:14):
Haven't you? What were you thinking? Why did you do that? Wouldn't you think to talk to me?
Guest: Ian Schnoor (19:20):
Right. And it didn't. And so if anyone listening has had a frustration using AI for any other type of report, if you've not used AI for modeling or agentic purposes in a spreadsheet, but if you've ever experienced frustration in AI while doing any other task, just overlay that feeling of frustration into modeling. First of all, you might get some great work out of it and you'll be pleasantly surprised and excited. At the same time, you might find that it will do things without telling you that are somewhat shocking, that scare you. And so it requires... So that's what we're seeing here. People are using it, they're cautious, they're hesitant on the council and are not yet finding earth-shattering time shifts. And what some people are saying is, "Hey, it just means I've got more time for other stuff. I've just shifted my time. I'm not exactly checking out at three in the afternoon.
(20:21):
I'm just getting some things done a little quicker and now I have time for other things that perhaps.
Host: Paul Barnhurst (20:27):
Well, it's like you said when we were on the mod squad one time, we were talking about this. He goes," I guarantee you the lights aren't going off at five in Manhattan in the finance district. The investment bankers are just going to be working the same hours. They're just able to get more done.
Guest: Ian Schnoor (20:41):
They're going to get more done and instead of going to the pitch tomorrow and meeting the CFO with one model that's half baked, they'll have five models that are mostly baked and they'll have more insights and they'll have more. But of course, as we all know, I mean, this is what all professionals have done. Whenever we've had productivity tools enter our professional toolkits, it means you can generate more. But every time your pitch deck goes from 10 pages to a hundred, there is more opportunity for a number to be wrong. And the CFO will notice on page 92, you have a number and that's inconsistent with another number on page 12. And then suddenly credibility goes down the tube and you're like, oh my gosh, I missed that. And then you go back and Claude says, oh yes, I didn't catch that. Or I should have told you that the number...
(21:31):
There is just more room for error. Yeah. So that was the first chapter. People are using it. Nobody is exactly kicking up their feet and drinking martinis yet at three o'clock in the afternoon. In the second chapter, we got into, okay, where everyone believes everyone on the council, there's nobody on the council that is saying, stop the train, stop the AI train, jump off the train, put it back in the can, and let's hope it never comes out again. No, everyone knows that we're not putting the toothpaste back into the tube. The question is how is shifting? So in chapter two, we looked at the financial modeling lifecycle. We looked at the classic six stages of a financial modeling project. Not everyone goes through all six stages, but the council was asked which stages are most suitable for AI usage and which ones are least suitable?
(22:29):
And in a nutshell, in that chapter, the clear conclusion came out that the shift, there will be a shift, of course, from construction to judgement . So there will still be need for work, but we're all hoping that it means that you don't have to be at your desk at midnight building manual formulas anymore like we used to. So the build stage, the actual mechanical construction stage will be expedited, everybody hopes and expects through AI. But the parts around scoping around design, what's this purpose of this model? Who's it for? What do they want out of it and why? And how are we going to design it, structure it and organise it? Those pieces, the strong feeling was that there is going to continue for a long time to be a human lens needed to attach to that. And so that's chapter two.
Host: Paul Barnhurst (23:19):
So I mean, that's how I fel. I think that's like you said, almost everybody in the council, that doesn't surprise me is still needs to be a human lens, but I think we're all excited for when AI can help us be faster. Now, some more than others, some will want to do it 100% themselves because there's some kind of pride in building it by yourself and that struggle and be like, I did all this model and being able to do formulas exactly as what you want. I mean, just like in any profession, there are some people that are very particular about how it's done. The more particular you are on very little things, I think the harder and harder it will be to give it up to AI. Bad financial models can lead to bad decisions or worse. So how do you minimise the risk of a bad model?
(24:07):
You make sure the models you build are great. Financial Modeling Institute developed the Advanced Financial Modeler Accreditation Programme to help modelers like you. The AFM programme offers a step-by-step approach to building world-class financial models. The programme ensures that you know the best practises in model design and structure and will help you brush up on your Excel and accounting skills too. Be the one on your team to build great models. If you want to impress your boss and your clients, get AFM accredited. Podcast listeners, save 15% on the AFM programme. Just use code podcast@fminstitute.com/podcast.
Guest: Ian Schnoor (25:02):
Yep, 100%. And part of it is through what I would call just the craftsmanship element. I mean, if you're an artist, if you're an artist or a craftsperson, if you are a mason or a woodworker, you're right. I mean, I've said this, a financial model is no different than building a wood sculpture. Financial models are just a virtual representation of something that exists in the physical real world. If you happen to be a wood sculptor or a mason and you like working with brick, if the day arrives that a robot can show up and you can stand behind it and yell instructions and it will do perfect work, you still might not like that very much, right? Because there's something about getting your hands dirty that you might like. Now, a finance professional will probably have to let go of some of that, but they'll have to know how to check it and audit it and make sure that what comes out of the final end is the same thing as if they had built it on their own.
(26:09):
And so there will be a piece of that. The next chapter we got into got into trust and governance. And probably the strongest consensus in the entire survey was that zero, zero of the 63 members, zero said, just looking at it and reading it right now, not a single member agreed or strongly agreed that they would be confident relying on an AI generated model for a high stakes decision without independent human review. Zero people said that they would be confident trusting it without human review if the stakes are high. If there's reasonable stakes involved, an acquisition, a financing, major decisions being made. Sure, for a pitch or a preliminary high level discussion, yeah great. But if there's a significant decision to be made, not one person said they'd be confident in the foreseeable future using an AI-generated model without meaningful human review and checking in.
(27:14):
And we thought that was pretty, not surprising, but interesting.
Host: Paul Barnhurst (27:18):
It's telling. It's the reminder that we're not there yet. Whatever there is, we're not there in the sense of what LinkedIn tells us we can do.
Guest: Ian Schnoor (27:28):
Wait, you mean there's influencers on LinkedIn that say we can just get our work done in five minutes?
Host: Paul Barnhurst (27:34):
No, no, there's not. I've never seen one.
Guest: Ian Schnoor (27:38):
Well, this group of people, maybe they're just technologically behind the time.
Host: Paul Barnhurst (27:46):
We just need to take more courses.
Guest: Ian Schnoor (27:48):
They need more courses, this group of 63. I'm certainly in that group, but they're not confident at this point relying on it without human intervention. 100%
Host: Paul Barnhurst (27:56):
Agree.
Guest: Ian Schnoor (27:57):
We asked the group about governance. 48% of the GLC members, so call it 50, half-half, half of their member organisations have no formal policy governing AI usage. And that was telling. So they all feel like they're in the wild west. They know they're using something new. Everyone feels that there is some element of governance required around it. Half the organisations have nothing. The organisations that do speak to governance, it's often very light. Or one funny quote that's in the document, well, it's funny, but it's not funny said, "Our organisation keps working on a AI modeling governance policy, but the AI tools keep changing so fast. We can't even get our first draught out because it keeps changing." And that's really telling as well. This technology is moving so quickly everyone recognises that there is an ethical need and a governance need, but they're having trouble keeping up.
Host: Paul Barnhurst (28:58):
Yeah, no question. I mean, we've talked about that previously on Mod Squad episodes, and I think it's something everybody's dealing with. The AI fatigue, as we've called it, just trying to keep up. It's crazy how quick this moves.
Guest: Ian Schnoor (29:12):
It's very crazy. And nobody, as we've talked about in other episodes, nobody that's serious into AI is just using one AI tool. You're using one for idea generation, you're using something else to generate documents or reports. You're using something else to check and double check the first one or the second one. The AI fatigue is very real and it's hard sometimes to keep track of all that. I mean, I've had that experience in working on documents as well where one of the AI agents will, I think we're almost done. And then somehow it looks a lot shorter than I was expecting. And it says, "Oh, I didn't tell you. I edited down 25% of the words to keep it within a manageable page count." I said, "Well, I didn't ask for a manageable..." And so It will do things right without again sharing. So there is a lot of, there's some hesitation.
(30:11):
Then the last two categories in the report, so if I told you that we had a unanimous view in the report that nobody would trust an AI generated model without meaningful human intervention, the last two sections we were split. The council was completely split down the line, which was first of all, around ethical standards. So most people agree, this is a very nuanced chapter, 70% believe that some form of ethical standards is needed and appropriate, but wide range of debate within the group on what is appropriate. And the 30% that don't want ethical standards, it's not that they don't want them. They're unsure of whether that's the right way to go. Everyone's thinking about the right way to... If you use AI, who do you tell? How do you report it? Who should know that you've used AI in your... And there's a wide range of views on what's appropriate.
(31:10):
Listen, I know most people on the council, this is a very, very ethical, ethics focused group. And so everyone's trying to do the right thing, but there's no consensus right now on how to get there.
Host: Paul Barnhurst (31:24):
I think a good example, if I can hear on the ethics, I remember when AI first came out, when I wrote articles and different things, I would source it that it was AI. And I remember one of the people were like, "Why are you doing that? You don't need to do that. " A lot of other content creators didn't, but now I don't even think about it to be honest. I help it write an article, but it's usually taking a transcript and putting something together. I'm still wording it and changing it, but these five words were AI. So it's interesting kind of like with modeling. "Okay, so it wrote a formula. Do I need to disclose that?
Guest: Ian Schnoor (31:56):
Exactly. And so within the council, there's debates. Well, do I tell my boss? My boss owns the model. Do I tell my boss if I use some AI? Do I explain exactly what I did and where? Okay, do I have to tell? And most people believe that that part is an important reporting line. D we tell the client? D we have to tell the client that we used AI? And there was great disagreement on that. Some said absolutely. And others were like, well, no. I mean, the owner of the model, the senior person at the bank or whoever's building it, they're not going to tell the client necessarily the profile. Well, if it was a person, the person who built it is a first year associate on our team or a third year asociate or they have no modeling experience or they have a lot of...
(32:41):
You don't report on the exact nature of the person who built the model. As long as the owner signs off on it, that's all that was done historically. So there's debate. Do we tell the client? Do we not tell the client? And there's no agreement there right now, which is telling of how early we are in this part of the process. And then the last question, the last chapter where there's not much agreement is what does the future of the profession look like? Roughly half right now are seeing some downward pressure on the career as a modeler. About 10 or 15% are sort of neutral and about 40% see upward pressure. They see more models being built, opportunity for more models. Models will become more democratised, a little less expensive, et cetera. So again, what is the future? And I'm not surprised because it's so unclear, nobody can possibly know the answer and that's born out in the thinking on the report.
Host: Paul Barnhurst (33:43):
That doesn't surprise me. Like you said, it doesn't surprise you either that there'd be a variance of opinions because none of us know where it's going. It's like when we build a model, we don't know what the exact right answer is. That's why we're building a model to try to estimate something and to get an approximate of what we think is going to happen. And it's no different with AI and trying to figure it out. We're all looking at it and say, "Okay, what do I think is going to happen?"
Guest: Ian Schnoor (34:09):
Right. Exactly. And back on that, well, a couple things. Is it different in the old days when you had a junior person building versus a more senior person? Is it a different responsibility to report back if AI was involved in doing some of the research and generating the assumptions? So these are all important ethical questions. This was not in the executive summary, but it is in one of the later chapters is there was a very, very strong, almost near... It was almost unanimous, but a very strong consensus. Over 90% of council members felt strongly that having human financial modeling skills was still going to be very important and knowing how to build a model yourself from scratch was still going to remain important even though I think everyone hopes and expects that you won't have to be in the weeds as often anymore. But the ability to do that, many people are worried that it will be difficult for future financial modelers to check models, to audit models, to provide instructions to AI, to check AI if you yourself don't know how to do it.
(35:21):
And so having the skillset was something that most of the council members believed was still very important.
Host: Paul Barnhurst (35:27):
Again, not surprised to hear that. So any other maybe key takeaways, things that surprised you? Any other thoughts you want to share about the report?
Guest: Ian Schnoor (35:37):
No, not much. I mean, I think it was a fascinating, to me, it's a fascinating report. It was really impressive and amazing to see all 63 people devote so much time to populating the survey, providing optional comments and sharing. It's given us such an amazingly rich report and a view of this profession and where it might be headed. Some parts feel more cemented and comfortable and others, as I said, the views are across the board. But no, I'll let people take a look at it. It'll be available. And I hope that people find it really instructive and helpful to think about where we're at as a profession.
Host: Paul Barnhurst (36:23):
Great. Well, when that's published, send us the link. We'll make sure it gets in there for this episode so people can go download that themselves. I'm going to have you go through rapid fire again one more time. You've done this in the past. Several questions have changed. You get to pick a side. At the end, you can elaborate a little bit because I recognise there's nuance. That's the reason we're asking these. There's not a right answer. So we're going to hold the elaboration till the end and we're going to run through these.
Guest: Ian Schnoor (36:49):
Okay. You know how much I love this.
Host: Paul Barnhurst (36:53):
I know how hard this is for you. Get through it.
Guest: Ian Schnoor (36:55):
Very hard. In my role at FMI where we, just to say at FMI,
Host: Paul Barnhurst (37:16):
All right. So let's see how this goes. Circular references, yes or no?
Guest: Ian Schnoor (37:21):
I'll say no.
Host: Paul Barnhurst (37:22):
VBA?
Guest: Ian Schnoor (37:24):
No.
Host: Paul Barnhurst (37:25):
Lambdas in financial models?
Guest: Ian Schnoor (37:28):
No.
Host: Paul Barnhurst (37:29):
External workbook links?
Guest: Ian Schnoor (37:31):
Nope.
Host: Paul Barnhurst (37:32):
That was the easiest one for you so far.
Guest: Ian Schnoor (37:34):
Yeah.
Host: Paul Barnhurst (37:34):
Easiest
Guest: Ian Schnoor (37:35):
One. That was the easy one. That was the only one that I actually truly in my heart believed absolutely no. Yeah.
Host: Paul Barnhurst (37:39):
I think you'll believe in this next one as well. Should modelers primarily use keyboard shortcuts?
Guest: Ian Schnoor (37:44):
Yes.
Host: Paul Barnhurst (37:45):
Should models always be print ready?
Guest: Ian Schnoor (37:48):
Yes.
Host: Paul Barnhurst (37:50):
Is there ever a situation where merge sales are acceptable?
Guest: Ian Schnoor (37:54):
Yes.
Host: Paul Barnhurst (37:55):
All right. Should financial modelers learn Python in Excel?
Guest: Ian Schnoor (38:00):
No.
Host: Paul Barnhurst (38:01):
What about Power Query?
Guest: Ian Schnoor (38:03):
Nope, not for modeling.
Host: Paul Barnhurst (38:05):
How about Power BI?
Guest: Ian Schnoor (38:06):
You're making me say yes or no. I'll say nope.
Host: Paul Barnhurst (38:09):
That's what I figured you were going to say. I get a mix on that one. Should every financial modeler be able to build a fully integrated three-statement model?
Guest: Ian Schnoor (38:18):
Yes.
Host: Paul Barnhurst (38:20):
Will Excel ever die?
Guest: Ian Schnoor (38:22):
Nope.
Host: Paul Barnhurst (38:22):
Have you used AI to help you build a model in Excel? I know the answer to that one, but I'll ask it.
Guest: Ian Schnoor (38:27):
Yes.
Host: Paul Barnhurst (38:29):
What financial statement is most important for modelers if you had to pick one? Income statement, balance sheet, or cashflow?
Guest: Ian Schnoor (38:35):
That is truly putting me on the spot. For anyone who understands the financial statements, they are literally equally valuable. I mean, gosh, for a financial modeler, people want me to say a cashflow statement, I'm going to say the income statement.
Host: Paul Barnhurst (38:52):
Alrighty. I always love the different answers I get on that one. It's really interesting. All right. Do you have a favourite LLM that you like to use? Quad, Copilot, ChatGPT?
Guest: Ian Schnoor (39:04):
Right now I'm using Quad, so I guess that's my go - to for now.
Host: Paul Barnhurst (39:09):
All right. So I'm curious to see what you say to this one. If you could pick only one for all your models, would you choose sensitivity analysis or scenario analysis?
Guest: Ian Schnoor (39:20):
Scenario analysis.
Host: Paul Barnhurst (39:22):
All righty. And then what is your lookup function of choice? What's your favourite lookup function?
Guest: Ian Schnoor (39:27):
I'll go choose.
Host: Paul Barnhurst (39:30):
I figured you'd pick choose. I was going to be surprised if you didn't. All right. Is there something you want to elaborate on there? You got through that pretty good. You did well.
Guest: Ian Schnoor (39:37):
Yeah, thank you. Yeah. Yeah. I'll tell you what, if anyone's wondering the thread and the theme that is weaving, so of course as you know, what I wanted to answer every single question was with it depends. But the theme that weaves through all my answers is... So as you know, I've been a financial modeler and identify that way back my whole career. As I said, I didn't identify at the beginning of my career that way, but now I mean that has been my identity. If you look
Host: Paul Barnhurst (40:13):
Back your entire... It's been central to your entire career.
Guest: Ian Schnoor (40:16):
From day one in my banking days, I was building financial models, literally. The thread that flows through every single one of my responses has little to do with technology and techniques and everything to do with clarity, communication, presentation, delivery, and being able to make clean decisions. So I have often found, listen, I've used circular references. I don't have a problem with circularity if you know what you're doing and if you can properly explain it and use it to a client. The default should be avoid it unless you need it. I mean, that should be the case for many things. Avoid complexity unless you need it. So I don't want to add a circular reference unless there's no way around it or we're all on the same page. But even some of them that might be questionable like print ready, people say, "Really? Print ready?" Well, yes.
(41:06):
Even if you're never going to print it, thinking about print ready forces your brain to think about the model from a clarity, from a communication standpoint. As a president, it allows you to think about the delivery, the flow and the way you're going to treat the model as a communication tool even if no one ever sees the pages. The fact that you've had to think about it that way means you're thinking about it as a delivery tool, a communication tool. So again, I don't want external references. I'd rather not have VBA. It's one other element of creating risk and confusion amongst clients. Sure, of course I've had VBA for printing and formatting and there's nothing wrong with that. But I mean, if you're giving me yes or no, I'd keep it simple. Same thing with lookup functions. I love X. I'm happy to use X lookup.
(41:56):
I've just seen people get... I've senior executives that don't know Excel that well get awfully confused by an X lookup within an X lookup. Whereas I've seen numerous light bulbs go off and joy on people's faces when they look at a choose function and realise I actually get it. I understand it. I know how it's working. So for me, my all answers have nothing to do with technology and more to do with how do I bring this person or group of people from the beginning to the end of a process so that they feel good about the decision that they need to make. I will err on keeping it simple.
Host: Paul Barnhurst (42:33):
Yeah. So one of my favourite quotes, and then we'll wrap up here, favourite thoughts. I remember someone on one of these episodes said, "Keep the model as simple as possible to start. Always keep it simple because your business partners, the people you're working with, they will make it more complex than you want by the time you're done. And so don't add that complexity to start with or it's just going to kind of fell under its own weight. You make it as simple as possible. Let them force you to make it more complex to meet their needs basically. And I really like that way of thinking about it. I
Guest: Ian Schnoor (43:05):
Promise you it is much, much easier to start a model smaller. Having a five-year annual model with five columns and 300 rows and the cost schedule is 80 rows, it's much easier to start that way and then be in iterative meetings with clients and have them say, yeah, you know what? That's really good. We like that. Or for them to say, no that's not enough detail. Can you elaborate this? Can you elaborate on that? And over time, maybe it becomes a few hundred rows longer or maybe it evolves in the periodicity evolve. But that's still much easier to do than start with a monthly model for five years, which is suddenly 60 columns and 10,000 rows of detail and then chop it back. That's much, much harder to do. So yes, I would agree. Now the right answer is to scope it properly and make sure everyone agrees upfront to know exactly what we need and what's the level of detail, but that doesn't always work that way.
(43:59):
We all know that it's an iterative process, but it's still easier to show something to somebody that's a little simpler and have them want to add detail than to take it away.
Host: Paul Barnhurst (44:07):
And I agree, you get the scoping right. So we're up on time. Any last thoughts? Anything you want to share before we let you go here?
Guest: Ian Schnoor (44:14):
Listen, I continue to believe that financial modeling is an incredible skill within many professions and it's an incredible profession. If you choose to make it your profession, now you have the option. You have the option of, but if you want to work in finance or accounting, I continue to believe it remains an incredible skill because it's never a good modeler knows it's never been about or just about building functions in Excel. It's about thinking and it's about a forced... I have become... People don't know this about me. If anyone has ever seen me in a talk, I often will tell them that I used to be the shyest person in the room. I used to be horrified at the idea of speaking publicly, about presenting, about communicating. And I can promise you that modeling is a discipline that allows people to elevate your accounting knowledge.
(45:03):
I've had people tell me that they really understood accounting much better. They've understood finance much better after they became strong mother. They became better communicators, they became better researchers. And so I highly encourage people to still get into this sport, this discipline, because I know I've seen it with me and with many, many others. It has the capacity and ability to really turbocharge and elevate many elements of your professional development and your career. So I will leave it at that.
Host: Paul Barnhurst (45:30):
Well, thank you so much, Anne. Appreciate you telling us about the survey. Allowing yourself to be put on the seat for a minute and have to answer without it depends and a little bit of career corner. So thanks again for joining us. I really appreciate it.
Guest: Ian Schnoor (45:42):
Always great to be here, Paul. Thanks for having me in.
Host: Paul Barnhurst (45:44):
Thanks. Financial Modeler's Corner was brought to you by the Financial Modeling Institute. This year I completed the Advanced Financial Modeler Certification and it made me a better financial modeler. What are you waiting for? Visit FMI atwww.fminstitute.com/podcast and use code podcast to save 15% when you enrol in one of the accreditations today.