Claude for Investment Banking: Building a One-Pager in PowerPoint with AI

Graham has a confession: he hates PowerPoint. Always has. The late nights at Lehman Brothers lining up charts pixel by pixel. The IC memos that never end. And when he recently tried using Claude Opus to build slides, his words were blunt: "they looked like cr*p."

So this week, he and Debs gave Anthropic's newest model, Fable 5, one sentence and a template. Build an investment banking style company overview for Nvidia. No test runs. No preparation.

The result was a genuine split decision. Formatting earned an A from both Debs and Graham: clean, well-aligned, properly labeled with units and sources. The basics that junior analysts routinely forget were all handled correctly.

Content was a different story. Graham gave it a B-plus. Debs was harsher: a C. The overview covered Nvidia at a high level but lacked segmental revenue breakdown, forward estimates, and what Graham described as "emotion."

The conversation that followed was arguably more interesting than the output itself. 

Graham argued that PowerPoint might actually outlast Excel as a day-to-day skill in finance: even after you stop building models, you never stop editing IC memos and pitch materials. 

Debs made the case for reusable prompt "skills," pre-prepared instructions that teams can share and iterate on, as the practical path to getting first-pass results from AI that are actually usable.

Debs's closing verdict: "I think this is a superb use of AI in investment banking workflow." 

Not because the output was perfect, but because the tedious formatting work is exactly what AI should be handling.

Key Discussion Points:

Graham's confession: why he hates PowerPoint and what Opus produced when he tried it. 

The Fable 5 test: one prompt, one template, no preparation. 

The split verdict: formatting A, content B-plus/C. 

Why the output "lacks emotion" and what that means for AI-generated materials. 

PowerPoint as a career-long skill: Graham's argument that it outlasts Excel. 

The skills approach: how banks are building reusable prompt libraries.

Why Wall Street Prep?

Wall Street Prep is the trusted training provider for the world's top investment banks, private equity firms, Fortune 1000 companies and business schools. 

Our online training and instructor-led boot camps are direct adaptations of our corporate training, making Wall Street Prep the ideal choice for those looking to break into finance.

DISCLAIMER:

The information provided in this video is for educational and entertainment purposes only and does not constitute financial, investment, tax, or legal advice. 

Investing involves risk, and you may lose some or all of your capital. 

Past performance is not indicative of future results. Please conduct your own due diligence or consult with a certified professional before making any financial decisions.

WTBD Newsletter:

https://webmail.wallstreetprep.com/whats-the-big-deal

Follow Us On Socials:

LinkedIn: https://www.linkedin.com/company/wall-street-prep/
Instagram: https://www.instagram.com/wallstreetprep/
Resources: https://linktr.ee/wallstreetprep

Transcript

SPEAKER_00 0:00

I got a confession to make that I actually hate PowerPoint.

SPEAKER_01 0:04

We've been really putting it to test over recent episodes, seeing how if it rises to the challenge of modeling in Excel, but now we're going to give it a new challenge. We're going to use it for PowerPoint.

SPEAKER_00 0:14

When I was using Opus to pull these slides together, honestly, they look like crap. So I'm really hopeful that Fable is going to tackle this problem here.

SPEAKER_01 0:23

I think this is a superb use of AI in investment banking workflow.

SPEAKER_00 0:29

Today we're going to continue on our theme of testing out AI and finance, and we're going to put Claude to the work in PowerPoint today instead of Excel. And just start simple and see how well it does building a simple company overview.

SPEAKER_01 0:42

Oh, I love it. We've been really putting it to the test over recent episodes, seeing how if it rises to the challenge of modeling in Excel. But now we're going to give it a new challenge. We're going to use it for PowerPoint. What skill are we going to challenge it with?

SPEAKER_00 0:57

We're going to start, we're going to start easy today. And we are, we're going to ask Claude to, in in proper format, we're going to give Claude a template and we're going to say, hey, just produce a company overview on a company of our choice. I think we're going to choose NVIDIA. We've been talking about AI a lot, so it makes sense for Claude to produce a company overview for a company involved in that space. So let's let's put it to the work, or let's put it to the test rather, and see how we go.

SPEAKER_01 1:24

Excellent. Now, experience tells me we should dive in, build the prompt, and get it running, and then maybe we can chat a little bit about company overviews whilst we're waiting for Claude to do its work.

SPEAKER_00 1:34

Exactly, exactly.

SPEAKER_01 1:36

Having a look in your screen, I can see you're using Fable 5. It's still your preferred model, is it, in Claude?

SPEAKER_00 1:42

It's my it has become my preferred model for everything. I haven't tried it in PowerPoint yet. And while it's building, we'll talk about some of the issues I've had with other models in PowerPoint because I have I have used them quite a bit recently. So I am really, really hoping that Fable does a nicer job than than Opus and some of the other ones, just because thus far it hasn't been great. So I'm I'm curious to see how well this works.

SPEAKER_01 2:06

I'm curious. We've just to confirm we've done no test runs. We have no idea what it's gonna show us. So looking forward to this. Okay, so nice simple prompt that you're creating there.

SPEAKER_00 2:20

Alright, literally said, build investment banking style company overview for NVIDIA, keep this to one page and use the template provided.

SPEAKER_01 2:29

Excellent.

SPEAKER_00 2:30

Okay, let's go.

SPEAKER_01 2:31

Right. Okay, so why are we doing this uh challenge? How important in your view is this as a an element of investment banking, creating a company overview?

SPEAKER_00 2:42

Well, let's let's talk about creating a company overview and also just using using PowerPoint more generally to put together whether it's pitchbooks or investment committee memos. This is this is a skill that really I you know what I'm actually gonna say, I'm actually gonna say people probably continue using PowerPoint longer than they longer than they use Excel, just in terms of editing investment committee memos, you know, that that kind of thing. Because once you become a VP or ED, whatever, whatever the kind of title title is at your firm, you stop using Excel as much day to day, but you always tinker with and and make changes to IC memos. So PowerPoint as a as a skill in finance is is generally pretty a pretty important one. Now, admittedly, when you're when you're a junior banker, you spend a lot of time making everything look absolutely perfect. And that's that's one of the things that I remember about late nights at Lehman Brothers or early days of Aries putting IC memos together, just making sure everything was perfectly perfectly lined up. Like, I don't know, do you remember? I think this was probably from like literally industry season one, where that investment banking analyst I'm not saying I was worked this hard, but that investment banking analyst literally like dropped dead at his desk. But one of the scenes before was him like lining up charts on the on the page. And I'm like, that that one, that scene stuck out in my mind because I'm like, I used to spend so much time doing that stuff when I was uh like a first-year analyst. So the fact that we've potentially got tools to take some of that, uh, some of that kind of pain out of the job, I think is really, really interesting, interesting thing.

SPEAKER_01 4:27

Absolutely. When you said industry season one, I was gonna, I was thinking, where's Gray and going with this? I mean, there's some really interesting scenes in throughout all of the seasons. But yeah, completely, completely, I think we've all had kind of late night moments where we've been struggling to meet a deadline. Um, and yeah, PowerPoint kind of is your friend uh in terms of helping, you know, communicate with clients, with investors. Um, it kind of helps us to bring all of our analysis together. As you say, it's kind of almost the tool that you use more when you become a little bit more senior because you're no longer doing all the Excel calculations, but you are then responsible for sharing ideas with the client, um, be it investment recommendations or you know, an MA transaction, a private equity transaction, and summarising your views in a way that the client or investor can understand. Um, and that's a skill in itself, the communication skill. And I really like what you said there about this almost probably being more enduring as a skill than Excel, because we still need to have those conversations. We still need to be able to communicate with our clients, even if the there's a lot of automation in the underlying calculations. Yeah. Um, but yeah, great if Excel can, but sorry, if Claude can do some of the heavy lifting on the actual PowerPoint, because I don't know about you, but trying to align text, trying to align boxes in PowerPoint, I feel like I'm definitely done with that. Um, and particularly like late at night when you've got a deadline to try and meet.

SPEAKER_00 5:55

I got a confession to make, devs. I actually hate PowerPoint.

SPEAKER_01 6:00

Oh yes, I think we all do.

SPEAKER_00 6:01

There is something, there's something kind of, I think, fun and interesting about the analytical skill set you build in Excel, but I don't derive any pleasure at all from using PowerPoint. And admittedly, there have been a few times recently where I've been pulling pulling materials together. I actually I had this uh I had this class that came together very, very late in the day and was pretty new. So I was pulling PowerPoint slides together. And I was trying to use Claude as a really helpful resource to do that. And when I was using Opus to pull these slides together, honestly, they look like crap. So I'm I'm really I'm really hopeful that Fable is going to tackle this problem here because I just I just don't want to be doing PowerPoint myself anymore. I don't know about you.

SPEAKER_01 6:47

Absolutely. Yeah, definitely. I mean, we used PowerPoint in a slightly different context in research. We were using it to um sort of come up with a slide deck that summarized our investment thesis and our view on stocks so that we could then use that to turn the pages with clients. Um so it's similar in the sense that you're trying to sort of aggregate all your analysis into a nice digestible format. Um, and we definitely had the same issues in terms of trying to make sure that everything was well formatted, all the numbers were consistent. And when you change a number in your analysis, you're changing it in your slide deck as well. All of that felt very clunky. I know that you know you can link numbers in PowerPoint to Excel, which helps improve it, but ultimately accuracy and layout is really key in PowerPoint. So we'll see how Claude gets on. Um, like you, I'm a little bit anxious to see uh how it copes. Uh so far, my experiences haven't been amazing in PowerPoints compared to Excel. Uh but yeah, let's see how Fable gets on. Uh, what's it showing on the screen now?

SPEAKER_00 7:44

Thus far is at least looking more professional than the stuff I was I was looking at literally just a few weeks ago. So whether whether anything on the page is actually correct or not, we'll we'll see in a sec. But in terms of the structure and layout, it's it's doing an okay job. Yeah, so we got a few main sections on the page. We got company overview, key facts, recent developments, summary financials, the world's leading compute world's leading accelerated computing company, providing full stack infrastructure powering AI at data center scale. I mean, as a as a one, as a one-line overview about what Nvidia does, you know, that that is pretty impactful. Right? We've got just you know, the the kind of high-level company facts, ticker, where it's founded, headquarters, the CEO, uh, some recent, you know, some recent news and summary financials. We've got, let's see, 2024, 2025, and 2026 actual. I don't know if we have, I mean, one thing that I would appreciate here, okay, fiscal year end last Sunday of January. All right, because I see 2026 actual, and of course, my my first thought is what what date is this, is this as of? Is this actually an actual or is it looked at at some kind of one-year forecast? But so far, so far this looks right. Have we got some numbers on on bar charts? It would be helpful to have a revenue kegger here. Uh button. Yeah, I mean, I guess from our perspective. That's that that's something that you can just with a simple prompt just say add a revenue kegger to the top of this bar chart. And you know, it's not it's not done, so maybe maybe that's still coming.

SPEAKER_01 9:26

Maybe it's still coming. I mean, I guess a bit of from a high-level bit of feedback from me. Um we usually see the company overview as four quadrants and try and you use that to kind of highlight different aspects that altogether um provide a really comprehensive view of what the company does and how it's performing at the moment. They've kind of the left hand side, we've got kind of recent developments and an overview. That's fine, but we often sort of squish that into one quadrant to allow a little bit more narrative about maybe, you know, the partners that it works with, some of the key customers. We usually see that. On the right hand side, yeah, with the financials for a public company, we would definitely want to include forecast numbers as well as actuals. Yeah. And that chart at the bottom, it's doing nothing for me at the moment because it's just literally showing your revenue line from the table in a bar chart. What I really like to see is some kind of segment analysis, get an idea of the relative contributions that those segments provide, the total revenue numbers. And even if they had that as a stacked bar, it provides a little bit more information than is in the table. I always think you only need to use the real estate once on a slide. You're basically duplicating information now. You're not adding anything new with that bar chart. So I have to say, there's at the moment, it's, you know, has it finished its uh its thought process? I think it has. Um as a kind of first cart, it's okay, but it's lacking the colour that you really want to get a sense of you know NVIDIA's business, particularly the segments. And we know that NVIDIA is hugely uh important in terms of or it's selling a lot into data centers. That's a key segment for it. I think about 90% of their revenues are selling the chips uh for data centers. And we can't see any of that in the at least in numerical analysis on that slide.

SPEAKER_00 11:10

Yeah, I think that's all that's all that's all fair and and valid. I mean, my my take here is on the basis of a like we didn't we didn't we didn't ask for any of that up front, right? We just said produce a company overview. I'm actually compared to compared to some of the output I was seeing literally with this same template a few weeks ago. I was seeing just just zoom zooming out for a second and thinking about this from a from a structural perspective and some of the work that I think makes PowerPoint really annoying, just in terms of making everything look good, not just lining things up, but just having things being visually balanced with like font sizes and sizes of tables and white space on a page, that kind of thing. That's where that's where I was seeing real, real issues before. And just if I just, you know, kind of zoom out and and look at the page here in terms of how it's used the space, actually I think this is pretty decent. Because a part of part of this job, or part of the PowerPoint job anyway, is making stuff look, look nice and look professional. The fact that the fact that I think we've delivered a lot more with with this year to a few weeks ago, I think is a really good step in the right direction. And I have no doubt if we if we did a couple things, we could get this in a much better shape. And one of those things for me would be, you know, this is probably this is probably an argument where if you're producing something like this on a repeating basis, you produce some kind of claude skill. And it says, all right, here's the here's the company overview template for for my preference for my firm, whatever, whatever the case may be. And to your point, you're going to include forward, forward estimates, you're going to include some kind of some kind of segmental revenue breakdown for companies where that's appropriate, and whatever other stuff that you're looking to, looking to pull on the page. That way, when you produce a simple prompt, say, saying produce this one pager on NVIDIA, it's already got that that skill to hand and kind of goes through, goes through and builds correctly. Um, but also if we just, I'm sure if we prompted here and said, hey, provide this revenue on a segmental basis, it would just do it. Right. So as a starting point, I'm I'm I'm getting I'm getting happier with this. And uh I wish I wish a few weeks ago I had more fable credits in my token bank because I just didn't. Uh so it was a lot more, it was a lot more manual fine-tuning to get stuff looking the way I wanted to. Um but from from my perspective, at least on on the the PowerPoint pain points, just getting stuff looking nice and well formatted, that's one of the areas where I think these tools, at least today, can be, I hope can be really helpful. And it feels like we're moving in that direction.

SPEAKER_01 14:02

And Graham, one thing that you mentioned now, I think really important uh is the use of skills, um, which in my view are kind of like pre-prepared prompts almost that you can then recycle. Um, because we're now teaching a lot with AI in the classroom. We're introducing how AI is used in key workflows across investment banking. And we know that efficiency is important, and it's not just about efficiency in terms of getting AI to do some of the key work, but it's also about how you prompt in an efficient way. And I know in all of our episodes so far, we've kept our prompts really simple. But when you're at your desk, you can't go through all of these iterations that we have, you know, been experiencing during some of our episodes where you go back and you say, actually, can you change this? Can you prove this? That you really want the first attempt, the first cut to be really robust. And having a skill, a pre-prepared prompt that you then save into your skill bank, you can then keep reusing that. So every time you come back to your company overview, uh, you have that pre-prepared and you save yourself that time. And this is what we're seeing is being used in investment banks, isn't it? The use of these skills um already is and you share them with your team. It's not just you that hosts those skills, isn't it?

SPEAKER_00 15:14

Yeah, exactly. Exactly. And I think as we explore this topic deeper and as we explore how to use these tools properly, and for everyone's benefit and up in upcoming episodes, we'll start putting out some content on actually not just what do you get from this prompt, but how do you use these tools to their their fullest potential and get the best result. Uh so that's both from uh from an Excel modeling perspective, from a PowerPoint perspective. Stay tuned because in the coming weeks we're gonna start, we're gonna start actually, actually showing people how to use these tools in a way that's gonna get the job done the best.

SPEAKER_01 15:52

Okay, so I think we've kind of put it to the test here with a very simple prompt. Obviously, uh there'll be some improvements that we could bet uh could make. But I know what we've done previously is we've graded Claude. How do you think Claude's done this week? Give it a mark.

SPEAKER_00 16:06

Okay, I'm gonna split my grades into two components here. I'm gonna give a formatting grade and then a content grade. Formatting, I think, I think I give this, you know, I actually give this an A. It looks, it looks clean, it looks tidy, everything is lined up. It doesn't look, it doesn't look crazy, it looks much better than the stuff I've seen before. Uh in terms of content, I'll give I'll give kind of a B plus. It's all it's all there, it's all it's all right, but I agree, I agree with you. It lacks it lacks some of the important detail and context.

SPEAKER_01 16:38

Yeah. Um I'm complete with you on the formatting and kind of included in that, not really formatting, but things like labeling stuff, having uh your sourcing at the bottom, your units in your tables. They've done all of that. The basics which so many analysts, when they're starting out, they forget to include. It's essential. If you've got revenue figure, it doesn't just float 300. Uh it doesn't just float as a number, you actually have a unit and a currency uh there. Um so it's done all of those basics. And as you say, it's not too cluttered. The other thing we often see with campaign overviews is someone someone thought, oh, I've got a bit more information, I'll just add another chart, another line, and then you almost don't know where to look because there's so much information. So from a formatting perspective, I agree. An A, very solid A for me. Uh, but the content, I always have a bit more uh, I don't know, punishing in my grades. I give it a C. I think this is a bit of a boring.

SPEAKER_00 17:30

Oh, okay, okay. Maybe it's a maybe it's a B, not a B plus. I mean, yeah, yeah, it lacks it lacks it lack is it crazy to say it lacks some emotion? I don't know. It's just Oh my gosh, that's beautiful, Grant.

unknown 17:44

It's beautiful.

SPEAKER_00 17:45

It's just uh it's just a bit generic, but you know what it means. But it's it's fine. It's fine as well.

SPEAKER_01 17:51

It's fine. And the reality is, so two things. First of all, um we sometimes do see company overviews which are a bit basic, uh, particularly if you're looking at a private company, you know, unless you already have, you know, if you're working on the banking side, until you've got information on that private company, the the number the information is going to be quite limited. Okay. Um for a public company, it's usually a bit richer in data. Um so a little bit basic, uh, given that NVIDIA is public. Um, but bearing in mind that bankers can be producing a number of these. If you're you know proposing an MA transaction, there might be five potential targets that you need to produce company overviews for. This is a bit of a you know, a bit of a pain, is a nice way to say it, uh, for you to produce when you're working, you know, for a client. You know, it's far and away the least exciting bit, I think, of a lot of the work that you do as an investment banker. So if we have got a situation where we have a skill that can automate a lot of that process, great. I think this is a superb use of AI in investment banking workflow.

SPEAKER_00 18:51

Agreed. Agreed.

SPEAKER_01 18:53

Fantastic. Okay, well, I think that was it for this week. A nice short and sweet episode on using AI for creating company overviews, which, as we said, is a core skill in investment banking. Um, I hope our listeners have enjoyed learning a little bit about Claude's performance uh here, and we'll keep revisiting this topic. And in as Graham said, we'll start to look at some of the ways we can optimize our use of AI for some of those workflows. That's it from me this week. And over to Graham.

SPEAKER_00 19:21

Indeed, stay tuned for more. It's been fun, been fun this week, and we'll see everyone same time here next week.