How AI Excel Tools Stackup Against the Hype and How Excel Agent Has Disrupted the Marketplace with Ian and Giles
In Episode 5 of The ModSquad on Financial Modeler’s Corner, Paul Barnhurst, Ian Schnoor, and Giles Male take a hard look at the changing landscape of financial modeling in the wake of Microsoft’s release of Excel Agent. Since launching at the end of September to coincide with Excel’s 40th birthday, Excel Agent has quickly changed the competitive dynamics for AI-powered modeling tools. The team explores the implications: how Excel Agent’s capabilities compare to other tools, why third-party platforms are shutting down, and what all this means for the future of work in modeling-heavy industries like investment banking.
Expect to Learn
Why Excel Agent is pushing competing modeling tools like Rosie AI out of the market.
What makes Excel Agent a “magnifier” of both modeling skill and error.
How fast AI is evolving inside Excel and what that means for modelers today.
Why AI won’t reduce hours in finance, despite speeding up modeling work.
What OpenAI’s Project Mercury reveals about the next phase of automation in investment banking.
Here are a few quotes from the episode:
“You can't hit a prompt, go get a coffee, and expect a working model.” – Giles Male
“If you don’t understand what the AI just built, you’re in trouble.” – Ian Schnoor
This episode makes it clear: AI is not a replacement for skill; it’s a multiplier. Excel Agent may be setting the new standard, but success still comes down to human understanding, judgment, and accountability. As the modeling world evolves rapidly, professionals who stay informed and upskill will thrive. The Mod Squad isn’t slowing down either, more tool reviews and sharp conversations are coming.
Follow Ian Schnoor:
LinkedIn - https://www.linkedin.com/in/ianschnoor/
Follow Giles Male:
LinkedIn - https://www.linkedin.com/in/giles-male-30643b15/
In today’s episode:
[05:29] - AI Tools Recap
[07:26] - AI Hype and Hidden Risks
[10:23] - AI as a Skill Magnifier
[13:48] - Microsoft’s Impact on AI Startups
[16:15] - Rapid Evolution of Excel AI
[21:29] - OpenAI’s Role in Financial Modeling
[29:17] - Understanding Assumptions and Calculations
[31:53] - Final Thought
Full Show Transcript
[00:01:28] Host: Paul Barnhurst: All right, so you and I haven't changed much, but Giles has been up to some stuff, so we're really excited to have you here with us today. And why don't we just quickly go around the horn? And in case someone hasn't been listening to these episodes, just real quick kind of lightning introductions. Giles.
[00:01:44] Co-host 1: Giles Male: Yeah, sure. Giles is, co-founder of Full Stack Modeler and Microsoft MVP. I'm in Sofia, Bulgaria. This week. I have been co-hosting the XL days event with some awesome speakers, and across this series. I've been kind of testing the tool so far on a few esports cases and a kind of bit of what I would call simple, like in the grid data analytics.
[00:02:08] Host: Paul Barnhurst: Perfect.
[00:02:09] Co-host 2:Ian Schnoor : And yeah. Hi guys. Good to see you back again in Toronto. I'm the head of the Executive Director of the Financial Modeling Institute. I spent my whole career working with financial modeling as a banker. Founded a training firm teaching financial modeling at banks, pension funds, business schools, and built a financial modeling module that's now a mandatory part of the CFA program. And, excited to be here with you guys testing AI tools.
[00:02:35] Host: Paul Barnhurst: And we're excited to have you here with us. And so a little bit about me and again, Paul Barnhurst, I'm lucky enough to host, co-host and everything in between for the show Financial Modeler’s Corner. And this is our special ModSquad series that we're doing on financial modeling tools. So really excited. You know, I go by the name of the fat guy for my business, and have been doing that for quite a few years, and been lucky enough to work with both these guys on various projects. Before we jump into things and do a little recap of where we're at so far already episode five. Can you believe it?
[00:03:11] Co-host 1: Giles Male: Nope. It's flown by.
[00:03:14] Host: Paul Barnhurst: It's weird to think, why don't we start? Giles, I heard you got a pretty cool award. Since we talk a lot about Excel and modeling, and we're all nerds here, why don't you share it with the world?
[00:03:23] Co-host 1: Giles Male: So, Ian, you don't know what this is, but basically all.
[00:03:26] Co-host 2:Ian Schnoor : I have no idea.
[00:03:27] Co-host 1: Giles Male: You will geek out at this. Anybody listening to this will geek out at this, right? So all of the speakers at Excel Days got an award from Mike as the special guest speaker. He was part of the original Excel like version one developer team back in 1984. And every decade, almost all of the original Excel team from Microsoft 40 years ago get together. They just had their decade, you know, reunion. So we've each got a plaque signed by about 80% of the original XL team. Given to us by Mike Koss, who was part of the original XL team. That is the coolest thing I've ever received.
[00:04:10] Co-host 2:Ian Schnoor : That is very, very cool. Good for you. That's it. Is he one of the guys that built the secret Space Invaders back in game? That was in the original XL? Do you even know what I'm talking about?
[00:04:21] Co-host 1: Giles Male: It's been brought up. Yes, it's been discussed over dinner. I don't think he specifically built it, but. But yeah, obviously he knows he was working with the people that did know all the Easter eggs. Lovely. Lovely guy. So interesting hearing stories about what they got up to 40 years ago. He showed loads of, like, home camera footage. Do you know what I mean? Like, behind the scenes, sort of just them camcorders from 40 years ago in the offices.
[00:04:46] Host: Paul Barnhurst: I'm guessing he has to be around 70 or older now, right?
[00:04:50] Co-host 1: Giles Male: Yeah, he's probably in his. I don't know, I would guess in his 60s, having been retired for a few years, was at Google for a little while a few years back. But yeah, a lovely guy. And, what a kind of honor as an Excel geek to meet him and get that present.
[00:05:06] Co-host 2:Ian Schnoor : Very cool. Very cool.
[00:05:09] Host: Paul Barnhurst: Congratulations on that, Giles. Thanks for giving us a sneak peek of that. It is really cool. We're looking forward to seeing it hung in the office.
[00:05:17] Co-host 1: Giles Male: Next to the Leila Gharani flag. Woo!
[00:05:23] Host: Paul Barnhurst: Oh, I don't even know how to respond to that one, so I won't. All right, moving on. We've tested three three tools so far. We've tested Tracelight AI, Microsoft Excel agent, which is kind of throwing a bomb into all of this in a few ways, I think. So maybe let's do a little recap of testing, and then we'll get into how agents are playing with all this so far. Giles, what are your thoughts? Where's your head at?
[00:05:53] Co-host 1: Giles Male: I mean, my head's definitely shifted since we looked at the agent, I think. I would say with tracelight and rosy, it was where I kind of thought things were, which was, on the one hand, just generally very impressive that things can do things that they are doing at any level, interpretation of words and challenges and getting some things right has been pretty impressive. And then equally, it's getting some things wrong and it seems very inconsistent and you can't really tell when it's going to get something right or wrong or how it's going to deliver the solution. So that was all kind of as expected. And then, you know, Agent Mode came out, the hype train went wild. And then we looked at it and it was like, okay, well, it is better. Like it's clearly better than the other tools we're seeing on the market. Still not perfect. It's still going to get things wrong. It's still a probabilistic model that's going to give you different answers to the same questions, which causes huge issues. But yeah, I guess my general sense is I think maybe a couple of things. Number one, I don't know how the other third party tools are going to compete if Microsoft's tool is ahead of the game, that's really tough. But also, yeah, we probably have to move fast as well in our podcast series. How are we going to keep up with the even faster changing environment than we thought it was? So yeah, I have a lot of thoughts and because I've been away for what, four days, I feel completely out of touch with what's going on in this space already. And I'll leave it there.
[00:07:23] Host: Paul Barnhurst: It's pretty amazing. What about your thoughts so far?
[00:07:26] Co-host 2:Ian Schnoor : Yeah. I mean many, many thoughts. wide ranging thoughts. On one hand, I, I'm excited about the new technology and what that might mean in terms of efficiency gains, minimizing errors. We have literally a brand new tool at our disposal that we would have never dreamed of that will actually build things for us, check things for us, audit things for us. We have a friend that will work 24 hours a day sitting on our shoulder. Um, as a supporter. So that excites me. What scares me is you talked about the fact that they can still make mistakes. They all make mistakes. They all did impressive work. They all make mistakes. The mistakes aren't what scares me as much. There's a couple of things that scare me, the number one, the overconfidence. We saw that in all of them that you tend to think that if a computer, a supercomputer, is telling you something and it has researched it and looked at it and it's confident with an answer that it's probably right, and yet in many instances, as we found it will confidently, it will more than happily and confidently suggest something that is clearly wrong because it's not sure. Now we know that hallucinations exist throughout AI tools, and we were seeing them, um, manifest in the tools we looked at. So. So that scares me. The other thing that scares me is that, you know, again, I can appreciate the benefit of AI tools to speed up things like coding. And we're seeing a lot less people needing developers, right? I mean, if you can get the code and you can test the code and it's working, then why do you need someone spending ten weeks on it if they can build it in ten minutes? I get that, but a lot of Excel tools and spreadsheet tools.
[00:09:08] Co-host 2:Ian Schnoor : Again, it's about the process of a human getting smart and insightful, um, and being able to deliver and communicate a result to a team, a board, a client. And I think we are a long, long way off from a board of directors or a management team. You know, putting a prompt into AI, getting the answer and being comfortable with that. They're still going to be people talking to people. And what scares me is that the AI tools we're seeing are thinking differently than a human would. So they're developing tools that are more complex. They're developing answers in solutions that are not nearly as simple, um, as, you know, an average user or modeler might, which means I think that might require human users to up their game to try to understand and interpret what the heck these AI tools, because you're going to be on the line for answering questions about the forecast and the tool. And if you can't understand what the AI is built for, you're in trouble. I don't think there's going to be a lot of patience for people saying, well, you know, I don't know what it did. I build something. You know, you can't expect me to understand it. No, we do, um, we do expect you to understand it. And you need to understand what we're talking about for this company. And, um, and that means people are going to have to up their game because it's doing some things that are more complex than most human users can do. So there's a couple of observations.
[00:10:23] Host: Paul Barnhurst: I'll share a few thoughts. One, I think in many ways it's a magnification tool. And what I mean by that, if you're unqualified, it's going to magnify that in the end, because you're not going to be able to find the errors and you're going to leave those complex formulas in there. And if you're a good modeler and you've spent the time to use AI appropriately and it can be used appropriately to save you a lot of time. Don't take me wrong. It's going to magnify your abilities. So I think in many ways it's a magnifier and it's the people that use it irresponsibly or don't know well. And maybe they're, you know, they're using it with good intention. So I don't want to say, like, you know, they're being bad people, but they don't have the skills to really be doing what they're doing. That's where I worry or what's going to be the cost and productivity and efficiency in, you know, deal trouble. Let's go on and on. So there's nothing we've tested so far that I'd say has blown my mind. I was surprised. On the whole, it's been better than I thought it would be, and that's far from perfect. But I think for me, as I've really come to a conclusion, it's for those that take the time. It's going to be a wonderful magnifier in many ways.
[00:11:37] Co-host 2:Ian Schnoor : I think that's a great way to put it. You know a similar analogy. You hear university students, if you're a university student and you're using an AI tool to be lazy, to do an essay for you, and you just need to write an essay on a novel, let's say The Great Gatsby. And you just put in a prompt and say, generate an essay and you don't read it and you hand it in, you don't know much. In fact, recent studies, including one from MIT, showed that brain activity is declining. Literally, brainwaves are declining of people who spent the past year doing nothing but prompting from AI not thinking at all. I think the same will be from, you know, Excel, work and modeling, but those who use it. If you have to build a Great Gatsby essay and you use it to help you generate ideas, and you use it to help you learn the topic, and you use it to refine your own essay, and you use it as a person, a partner to bounce back and forth. I think that'll make you better. And I think that's what you're, you know, you're saying as well, Paul.
[00:12:28] Host: Paul Barnhurst: Exactly. I think we're all on the same page there.
[00:12:32] Co-host 1: Giles Male: I had one more thought. Yeah. So again, it's partly just because of some of the things I'm reading and thinking about. My guess is that it's going to significantly change again very quickly. There's a few things that are on my mind. Number one is where you can really train your own tool. I know to an extent you could upload, you know, models and and say build this way. Well, we get to the point where you can really train tools, you know, based on entire companies sets of templates and be like, look, this is how I want to work and how I want you to model consistently. And it remembers well. I think that would be a big shift. And then I think even things like when, when the kind of voice prompting gets to the next level, which I think is, again, we're on the cusp of it now. When you are just dialoging with your LLM and you've got that memory and it's trained, I could see this shifting so quickly. and I guess where the agent update left me was again thinking, oh, this is not years and years and years away. So I wouldn't be surprised if we're having another chat in three months and we're going. My God, even what we talked about then was just feels so out of date now.
[00:13:48] Host: Paul Barnhurst: I agree. I mean, you look at Excel agent, I think we should talk a little bit about the impact of that came out last week of September, same time as the 40th birthday of Microsoft. Happy birthday everybody. Check out our agent. Right. That was kind of the messaging. Let's start vibe working as they call it. And that impact's been huge. In just those few weeks we've seen two tools shut down. One that we released the episode because we found out late the night before that they no longer existed. That was Rosie AI. It changed, you know, in the words of their founder, it changed the funding environment. Investors went from, this is a great idea and we're willing to fund it. This is a great idea, but we don't want to compete with Microsoft. Yeah, right. And just just on a dime. And so it'll be really interesting to watch how many of these tools go away. My thinking is 85 to 90%. There are some that have done special things, focusing very much on. Analysts have brought in FactSet and a lot of other data sets, and have done enough that they'll probably be there for a while, and I think there could be room for a couple. But would any of you expect most of these to survive, kind of, given the current environment?
[00:15:00] Co-host 1: Giles Male: Surely not. And actually, I think, Ian, maybe you made that really good comparison to the car industry in one of the most recent episodes, and it stuck in my mind for this episode, thinking you were exactly right with that prediction about I think it was you, Ian.
[00:15:16] Co-host 2:Ian Schnoor : Yeah, we're experiencing an exciting time. It's a gold rush mentality. Everyone's trying to stake their claim. But that's right. This seems typical of most early stage industries and businesses where there's a lot of capital flowing in, a lot of excitement, a lot of upside. Yeah, I think I had mentioned that in an article I read that there were over 3000 car manufacturers in the United States in the early 1900s. And of course, now we're down to a handful. So it's not unexpected. And being able to differentiate yourself will get harder and harder and more and more expensive. Yeah, I agree with you, Paul. I think some will survive if they're doing it if they build the right partnerships and, and develop the right niche product that can layer on top of what's there. Sure. Why not? And make people's lives easier and faster and allow you to work more and get more done if you're just, you know, putting yourself on top of an LM and doing, you know, a wrapper on top of an LM, chances are you're not offering much unique differentiation.
[00:16:15] Host: Paul Barnhurst: Yeah, I think we're all on the same page there. The other thing that with aging is interesting, just to watch how quickly, you know, Microsoft is moving in Excel historically. I remember as a kid, right. Every three years you'd get the new office version. Then they went to 365 and every month you'd hear something. Right now? It feels like every week. It's just amazing. The development. Like, you know, they now have autocomplete your functions in Excel. Cloud hasn't rolled out to everybody, but in beta they've already taken agents from hey you gotta install this add in to yes it's inside copilot now. You know you still have to have a copilot, but you know what it is? Next week will be on the desktop for all we know.
[00:17:00] Co-host 1: Giles Male: Probably. That's probably next.
[00:17:02] Co-host 2:Ian Schnoor : And it's interesting. You're right. And of course, in the world of technology and in the world of AI, speed seems to be right now valued more than anything. Yet I think many of our long standing, trusted institutions around the world have demonstrated that that speed is not always optimal. And sometimes it's important to have, you know, a ring fence around it and a mechanism that deliberately slows things down. I mean, I think about, you know, in England, the House of Lords or the Senate, in many countries, those are deliberately in place to slow things down, reflect, pause, um, and understand the implications of what you're doing. And one could argue both sides. Right? that sometimes maybe it slows you down too much and doesn't let you get enough done. But in the environment we're in right now. Well, we saw the release of that, another explosive letter yesterday, right? By a prominent group of people all over the world in all walks of life, saying, you know, slow down. This is dangerous and moving too quickly. And so, yeah, there seems to be a priority on speed. And we'll find out what the implications are of moving as quickly as we are.
[00:18:14] Host: Paul Barnhurst: I think you bring up a great point, right. It's easy to break things when you move really, really fast. And we're moving fast. Also, it's easy to get caught up. Shiny object syndrome, right. That's the shiny new object. I need to go look at it. I'm not going to do anything for you in six months down the road. So you know why waste your time type of things? I think you bring a really good, kind of counterbalance to how quick this is all moving. Right. That reminder that it's not necessarily a good thing.
[00:18:41] Co-host 2:Ian Schnoor : You know, I'm pretty connected with the world of capital markets and people that work in capital markets. And yes, you keep hearing rumblings of jobs and job losses, and I'm sure it's true. I suspect it's mostly in what you'd call the middle office people that are doing mostly repetitive manual tasks that can be automated and should be automated. And that's not a bad thing. Um, my strong sense from what I've people I've talked to is those that work on the front lines, those who are working in investment management, those are who are working in investing money, real money, sizable money, doing deals in private equity, venture capital, credit, you know, really on the front lines of those things. I haven't heard of meaningful job losses as a result of AI yet in those roles. I think it's still very relationship driven and people driven, and I don't expect to see 80% of those jobs cut in the short term.
[00:19:35] Co-host 1: Giles Male: I was going to add as well, what you just said resonated so much with me, Ian. And one of the big kinds of headlines we're seeing in a fair few places is that, you know, this tool can build a financial model in ten minutes. And I quite like the fact that I think we are collectively kind of saying, okay, but why? Why would you want to do that? Like, is that the right target? which I don't think it is in a lot of cases.
[00:19:59] Co-host 2:Ian Schnoor : Sorry. Like so in some instances it might be and in some scenarios. Right. I mean, I can imagine a scenario where you are a junior banker and your boss says, hey, I'm getting on an airplane at 9:00 in the morning. Here's the thing. You know, before Excel was around, you couldn't do certain things. So in most of those roles, certain requests weren't even possible. Um, if you wanted to analyze, if you wanted to build, let's say, a chart, even a simple chart or a graph of a company's stock price, and your boss wanted you to, you know, crank out 50 of them in an hour. Well, you couldn't do it. So then Excel came around and allowed you to do it. So then it suddenly meant they started asking for it. Oh, hey, now we have got the capacity to ask you to give me a presentation with 25 graphs in it, and you can do it in an hour. So it's possible. So in the old days, your boss would have never asked you to throw ten rough, crude models in an evening because that wasn't possible. And now that it is possible, your boss says, hey, I'm getting on a plane at nine in the morning. Can you throw together ten quick and dirty crude models of companies that I can look at on the airplane? Okay, so you'll do it. You're still going to need to spend a couple hours in AI cranking them out, generating them, looking at them, etc. but you still need someone to do that, right? So there's a time and a place for everything. But you're right. I don't think it's going to replace the actual real deal model that gets built to analyze a truly rigorous, important transaction where you're allocating capital. I don't see that changing anytime soon.
[00:21:27] Host: Paul Barnhurst: I think we're on the same page there. So you had shared with us before it came on. Yeah. Loved. If you could bring that up, maybe talk a little bit about it. Bloomberg just wrote an article about some stuff OpenAI is doing that I think applies here to what we've been talking about and the whole modeling world and what's going on.
[00:21:46] Co-host 2:Ian Schnoor : Absolutely. I mean, it was an interesting article that came out, you know, again, this if you are a third party provider, you know it. This again is further cause to be concerned, I think. And I'm just going to share my screen here which is so true. I mean if, if we want to quickly take a look, there was an article that I'll show you. Take a look here. I found it fascinating to see this. Openai has been hiring bankers, um, ex Goldman Sachs staff to cut down on junior bankers grunt work. It's an article that, um, they've got a project name for it. And, you know, I won't go through it sentence by sentence, but they've basically hired more than 100 ex investment bankers to train its AI on how to build financial models, as it looks to replace the hours of grunt work performed by junior bankers across the industry. So that's fascinating. Again, if history is any guide as to the point that I was making, it's great. They've got a name for it. It's codenamed Project Mercury.
[00:22:51] Co-host 2:Ian Schnoor : Contractors are expected to submit one model per week with instructions on how to write it. Um, to do their teaching, you know, AI to build models. Fantastic. You know, in my heart I believe that's just going to mean that the junior staff will be expected to do more models faster, um, than they would have been previously. It doesn't mean they're going to. It doesn't mean that there's going to be meaningfully less bankers. It doesn't even mean that they're going to go home earlier. That's just not the culture. You are not going to see investment bankers going home at 6:00 for dinner. It's never going to happen. It just means that instead of getting one model built by midnight for a pitch, they'll be expected to get ten models done and they will still be expected to check. And now they'll be expected to crank out a model in 15 minutes. But still, check it, double check it, make sure it's working, summarize it, get it, put in a deck. They're going to be working more hours if you ask me.
[00:23:51] Host: Paul Barnhurst: As in just being a wet blanket here or am I missing something?
[00:23:54] Co-host 1: Giles Male: No, I think he's absolutely right. But it's funny because I don't think I've heard anybody else say it. So it's. That makes me laugh. I think you're absolutely right.
[00:24:03] Co-host 2:Ian Schnoor : It's the culture. Here's the thing. I actually.
[00:24:05] Host: Paul Barnhurst: Agree.
[00:24:06] Co-host 2:Ian Schnoor : As well. I don't think it's going to change the culture. I've been around investment banking capital markets for many, many, many years now. Um, you know, everyone thought that spreadsheets were going to change the culture and make, you know, work life balance better. Then everyone thought that the internet was going to make work life balance better. Then everyone thought that tools like FactSet and cap IQ were going to make work life balance better, because then, you know, in my day, when you had to do what was called a comp, a comparable analysis on a public company, you had to download, um, the the filings or go to the library and get them out and then sit at your desk for hours and pump the numbers in. And you could get through a few companies in an evening. And then tools like FactSet and cap IQ came out, and suddenly it meant that you could create a financial analysis, a comp set, you know, in an hour by just prompting it and checking it. So it just meant that you could do more. Nobody's going home earlier. You could do more. Um, it just means that you can get done more things, um, than you could before. I hate to break it. No one's going home at Bay Street and Wall Street and Fleet Street. Is that, um. No one's going home. The windows will not be dark by 8:00 at night. I hate to say it.
[00:25:21] Host: Paul Barnhurst: Giles, I think we have the title for this episode. No one's going home early. I hate to break.
[00:25:29] Co-host 2:Ian Schnoor : It's on that one. I'll put money on that one.
[00:25:32] Host: Paul Barnhurst: Yeah. I never worked in the investment banking culture. The never going home early was a big reason why.
[00:25:37] Co-host 2:Ian Schnoor : Yeah, I mean, you can see here and, you know, like they've brought in people from, you know, all the big banks and they're paying them per hour. I assume they're still working. So I don't exactly know if they're paying them full time. My guess is that they're doing it on the side.
[00:25:52] Host: Paul Barnhurst: I think it's a contractor. So I imagine almost all of them probably have a day job.
[00:25:57] Co-host 2:Ian Schnoor : Yeah, and they're probably working in banking and sneaking in some hours late in the middle of the night to do it anyway. Sure. I just think that it was fascinating. I'll stop the sharing now. Or you can stop. But, I mean, I believe it will continue to be a new technology and tool that allows us to work differently, hopefully more efficiently in some ways, but I don't expect it to have the life altering impact that some are predicting.
[00:26:23] Co-host 1: Giles Male: Paul, you mentioned the term, the vertical part of this a while back. And again, that really stuck with me. And, you know, when you look at which companies are going to dominate this space, I think we would probably all guess Microsoft is going to open AI and, you know, the big players that are around, it's that vertical part that I still don't fully understand how that's going to work. But if somebody really does like the financial modeling part of this uniquely, well, I guess that's where you carve out a space alongside Microsoft's agent and all the other llms behind that, you know, Claude and everything else.
[00:27:03] Host: Paul Barnhurst: Or an example. You do real estate really well, or you do analysis in the sense of there's one tool called drift AI. And their focus is, look, we're not as good at building models. We can do that with our agent, but we've created personas. We're bringing in FactSet. Okay, you're a credit persona, and we're helping you figure out a loan. And hey, if you're a little better at those different personas enough so that, hey, I get it right three, 4% of the time, or you don't have to be a lot better for the cost to make sense. And so I think some of these different tools will figure out whatever those unique angles are. They're able to load things in faster. But the vast majority, if they're if they're just a wrapper with, yes, they have some instruction, I get it, they've done some frontier models, but at the end of the day, there's no competitive moat beyond the Lem. I don't see any of those surviving.
[00:27:57] Co-host 1: Giles Male: Yeah, yeah, unless we're missing something. But I in principle, I.
[00:28:01] Host: Paul Barnhurst: Wouldn't be the first time. Let's be honest, at least for me.
[00:28:04] Co-host 2:Ian Schnoor : I agree with and I don't. I don't want to sound like I'm, you know, being negative on all this. I'm extremely excited. I mean, I like it. If I could, um, you know, I've always been one to say, you know, I've taught Excel in modeling for 20 years, And my big joke is that I'll often teach for three days. You know, in my old days as a trainer, when I would teach a three day advanced Excel course, I would joke at the end and say, you know, you guys all probably think that I love Excel. I don't, I mean, I do, but the reason I want you to learn all these great Excel skills is so that you don't have to spend your whole life in front of Excel so you can spend get Excel to do your work for you so you can spend your time doing higher value added work so you can be thinking and analyzing and communicating and delivering and elevating your own game. And I think the same thing about AI. I love the idea that AI can help, you know, build me a fundamental model, a merger model, an LBO model, and get it done for me by 7:00 at night or whatever it is, so that I can spend the rest of the evening really thinking about how we're going to present it and deliver it and get smart for a meeting instead of waiting till midnight when I'm done and having no time for the higher value added stuff. I think it's brilliant if it can do it, but I still have to understand it. So that's my feeling. I'm excited about the tool.
[00:29:17] Host: Paul Barnhurst: Exactly. If the tools get to the point where they can build the model for us, we still have to understand the assumptions that were used. There still has to be a validation, and maybe it's another model that kicks out a report and we review it and we're comfortable with it.
[00:29:29] Co-host 2:Ian Schnoor : Well, my point is you have to understand these assumptions, but you will have to understand the calculations as well. You will have to understand how to test it and know, I mean, if you're building a complex piece of debt that's got some really interesting features and covenants and modifications, you need to understand what the formulas are doing so that you can ensure that it's working and stress tested under all scenarios. You can't be. I don't think you'll be able to just hope that it's working. So yeah, I think you're right. You're gonna have to do that.
[00:29:59] Host: Paul Barnhurst: All right. So what's what's next? Let's talk a little bit about that. If anyone's stayed four and a half episodes we thank you. We're not done yet. We have more excitement, more testing to do. You definitely Excel agent is one of those things we want to test further. But there are other tools. And so Tab AI is one that's really exciting for us to test. I know, Giles, you've been, itching to test a shortcut that's been high on your list that you've been excited about.
[00:30:29] Co-host 1: Giles Male: Very high up. But again, you know, it's one of those where I mean, I had a chat with Nico. It's a high priced product. It's been high up on that benchmarking bit of analysis the whole time. And Microsoft has recently come out with an agent saying, hey, we're now top of the pile again. And I saw Nico come out a week ago and said, actually we think we are. So yeah, I think we need to test shortcuts.
[00:30:53] Host: Paul Barnhurst: Yep. And I think a potential index is in there. They're funded by OpenAI. They raised $14 million still. So there may be a little bit of a wait. But we're definitely going to keep testing. We're going to bring you more information. You know we're not done with this series despite that little Excel agent bomb that came out at the end of September, we still think there's definitely benefit from looking at some of these different tools. So I hope you'll stay with us throughout this journey. And we have some other exciting things we're thinking about, and we'll share those as we get there. You know, maybe looking at some MCP servers or some different kind of competitions and fun and all this because we want it to get a little bit of entertainment. But more than anything, we want you to learn and hopefully at the end of this, everybody will feel like they're better prepared for AI and how to think about it as a modeler, and also just better prepared to be a better modeler at the end of the day. So why don't we kind of go around the horn and maybe we'll do some closing thoughts and we'll, we'll wrap this. So in any final thoughts.
[00:31:53] Co-host 2:Ian Schnoor : Oh gosh I shared a lot of final thoughts. We are truly in the midst of revolution. I mean, that's probably understating it. It's an exciting time. And I for one, as you talked about the magnification, which I really like, Paul, I for 1 a.m. excited about how these tools will make me better. I mean, you know, I like you guys. I'm someone who is always looking to improve, to get better, to learn. And I think as people, if for those of us looking to learn and get better and have a permanent helper, wow, it's exciting times. But you're right. I think some of these tools will leave people in the dust if they are struggling to keep up and now struggling to understand what's going on. So it will see where we get to in this journey. But excited to keep looking and testing with you guys and seeing where this all goes.
[00:32:39] Host: Paul Barnhurst: Giles.
[00:32:40] Co-host 1: Giles Male: Yeah, I think I probably covered most of my kind of general feelings in the episode. I guess what I would say is I just continually think the three of us could keep doing this for a very long time. I really enjoy the kind of, you know, the I'm learning as we're talking and doing this anyway, and I think we could genuinely rush to test every tool on the market, and one week later there'd be a bunch of new tools, or they'd have all updated, or agent would have updated, or every major LM would have gone through another evolution cycle. So I could imagine us doing this for a while and I'd be very happy to. And secondly, a possible alternative headline is Einstein says I don't like using Excel. Just throwing that out there. I'm sure you said that.
[00:33:23] Co-host 2:Ian Schnoor : Well, I sort of do.
[00:33:25] Host: Paul Barnhurst: You're not going home early. Forget seeing the kids.
[00:33:28] Co-host 2:Ian Schnoor : Look at that. I love it. You're pulling all sorts of headlines out of this. You didn't know that. That was all in me. I've been hiding it.
[00:33:35] Host: Paul Barnhurst: Well, on that clickbait note. We'll go ahead and wrap up here, but thank you for joining us. We're very excited. We'll be bringing you another tool next week. So again tune in. We'd love to hear from you. Dm any of us. Find us on LinkedIn. You know, send us an email. We're really passionate about this project and we'd love to hear your thoughts. So thank you everybody for joining us.Bye bye.
[00:34:02] Music: The mod squad. We are the Mod squad.