Welcome to our session again, week four.

Today, we're looking at an AI tool made for investment banking called Rogo, and we're going to be covering using Rogo across real investment banking workflows. So we're going to look at a session that shows how an analyst might use Rogo on a continuous sell-side process.

Now, by the end of the hour, you should expect to see at least three to four deliverables. We're going to look at a cited target profile, we're going to create a buyer capacity matrix, we're going to look at a buyer book that we can actually export directly into a professional-grade Excel spreadsheet or even PowerPoint, and we're going to build a self-refreshing buyer appetite monitor. That will be at the very, very end.

Now, while doing that, I'm going to, of course, show you Rogo and all its features.

Now, a quick line on me. My name is Juan Cabrera.

I started my career as an analyst at Merrill Lynch.

Then I went back for a PhD in quantitative finance, where I happened to play with these AI tools, of course, at a very early age.

For 15 years, I've been training bankers across the street, both on the buy side and the sell side. And for the last couple of years, I've been working directly with banks in helping them embed AI tools into their everyday workflows.

So I'm really excited to be here, and I hope it's a productive hour.

I'm going to jump right into it.

Just a couple of housekeeping items. If you have any questions, feel free to use the chat.

You can also use the Q&A tool. Okay, both work.

I'm looking at them right now, and I can answer as many of your questions as we have time for. Okay, so let me get started.

And this is an overview of what we're going to be seeing today.

The first thing is I want to give you a conceptual overview of what Rogo is as an AI tool and how is it unique, how is it different from other more generic or broad usage tools like Claude, like Copilot, ChatGPT, and others.

So Rogo, again, is a tool made for investment banking.

So I'll talk about that a little bit, and then I'm going to show you all the different features within Rogo that I'm going to cover over the next 60 minutes. And in doing so, I'm going to walk you through.

So, okay, great.

Now, I'm going to walk you through our entire workflow, and you're going to see the workflow is going to be pretty straightforward.

And of course, we're going to end with the takeaways for the session. All right? So let me get started with just a conceptual overview of Rogo.

And again, don't worry, we're going to jump into Rogo.

We're going to spend most of our time looking at the tool.

Now, the way I would think about Rogo is a layer built on top of the same frontier foundational models that we already know.

So Rogo actually leverages all the main frontier models. It leverages mostly OpenAI, but also it leverages Anthropic models and Gemini, all three.

Okay, so it's just an IB layer that is built on top of our very popular, very amazing models. Now, what are some of the key aspects of Rogo that make it unique, make it specialized tool for banking, specifically? Number one, and I'm going to start with that, that it's grounded in professional-grade, banker-grade data.

Rogo is connected to your main banking data services like Capital IQ, FactSet, PitchBook, and others. It's got access to all of-- Of course, it's got access to public filings in the US, in the UK, and other countries.

So it's grounded on professional-grade, banker-grade data, and that is extremely important, especially since we know that AI tools can sometimes drift and hallucinate.

Now, the other thing is that it provides you with citations, very well-built citations.

Now, that might seem like a small thing, but it's actually a huge thing, as bankers always need to cite and have an audit trail of their work.

That's a key part of their everyday work.

Rogo will speed up that process. It would help you with that process, and I'm going to show you some examples of that. Now, that's one big reason, I think, why Rogo is so unique. The other reason is because it provides templates for IB workflows directly. So an example of that are these things called shortcuts. Now, be careful.

The meaning of shortcuts in the context of Rogo is not the same meaning as Excel shortcuts, like Control + C, Control + V.

We are talking about an actual feature of Rogo called shortcuts. Think of them as quick actions, if you might.

Although they also have quick actions separately. So I'll show you all that.

Now, what's great about shortcuts is that it saves the banker time in terms of prompting time.

You don't have to keep up with a prompt library somewhere.

Actually, you build your prompt library within Rogo.

You can just click once or twice on a shortcut, and you have your actual deliverable after a couple of minutes, as opposed to having to think about what to prompt, how to prompt the AI tool. So that's a huge efficiency feature that helps specifically bankers, and Rogo has its own shortcuts, but it also lets you build your own.

Great. The other and last key unique feature of Rogo is its ability to generateor export and generate output that is professional-grade output. And I'm going to show you some of those.

So you can export to Excel, PowerPoint, and Word.

But the output will be very nice, banking professional looking.

Nothing like text or markdown and so on.

And it's actually very nicely formatted in very nice templates. Of course, your bank might have its own template, and Rogo works directly with banks in embedding those different formats and templates into the tool.

All right. Great. Now, having said that, I have a question here: does it support every region? That's a great question.

Not a question for me. Okay. I would say they have a great customer service email. Send them an email. They'll get back to you.

It might be an AI assistant or a human assistant that will get back to you, but that's definitely a question directly for the company.

Now, here are the features that I'm going to show you over the next hour.

There is nine of them.

That's a lot of them, but every single one will be actually showcased here.

Okay? We're going to look at setup capabilities, things like projects and personalizing your account, and also personalizing your project, as well as shortcuts, which I already briefly mentioned.

We're going to look at some research features, such as citations, screening, and tables.

By the way, I like to refer to these tables as AI tables, and you will see why. They're pretty amazing. And then you have automation and output features, like the spreadsheet feature, export feature, and the scheduling feature. Okay? All of those will be looked at here.

So how about we jump straight into Rogo? But let me set up the scenario for the hour, okay? Here is the scenario. You are an analyst on a mid-market M&A advisory team.

So now you get a sense of what role you're doing, the type of company you're helping.

Although I actually ended up using a public company.

Okay, so it's just easier, and it shows the features of Rogo a lot better.

So let's say you're an analyst on an M&A advisory team. Okay? The MD just mandated a sell-side process for a publicly traded target, okay, which I'll tell you about in a minute, and wants a research package. Okay, so what is in that research package that you've been asked to build? The target profile, a buyer book with a capacity matrix, and a buyer appetite monitor. Okay? And the reason it's in all caps or, sorry, in uppercase for the first letter is because I'm going to actually name it that specifically. Okay? And you need that by tomorrow morning.

Let's not talk about what time it is right now, but let's say it could be 9:00 PM.

So the first step, and that's what I'm going to show you now, is how to set up a deal project within Rogo, okay, and how to personalize it, and how to start running your workflow using shortcuts. Okay? So you can see on the right side the features that I'm going to show you now are personalization projects, shortcuts, and citations. Okay? Well, what about sources of data? They come with Rogo, right? They're embedded in the tool. Okay? And of course, companies can work with Rogo to embed additional databases they might have.

But we have access to, as you can see, Capital IQ, public news, public sources, as well as quarter for earnings calls. Okay? So let's jump right into Rogo.

Feel free to ask any questions anytime.

This is what Rogo looks like. Sorry, I'm just pulling my Rogo here. Perfect. So here it is. Okay? Now, at first, when you look at it, you will see it looks a lot like other-- Let me actually close something because this is a bit distracting.

If you give me a moment.

Okay, much better there. Great. So, it looks like the same user interface as ChatGPT and Claude to some extent. So let me quickly walk you through this.

In the middle, of course, you have your chat box.

That's where you do most of your work, that's most of your inputs.

But you will notice there's a few features around it, and I'm going to walk you through them through the hour. I don't want to get into it right now.

So here is where you can type, build, and that's your task, that's your prompt box. Great. And all the features around it, which I'll explain over the hour. Now, on the left side of the screen, you have your sort of workspace and your history kind of section.

And again, it's very similar to other tools.

You have your Home tab, you have your Search.

By the way, you can open up your search pressing Control K.

So there it is. I can search through all of my different chats, all of my different tables, so many things. Tasks and so on. Okay? You have tables, which I've already mentioned before.

I will talk about tables at the very end of the hour.

You have skills. Now, unfortunately, skills are not something that I'm going to discuss in depth in this session. Skills is a very new feature, and I do want to say this, don't confuse skills with Claude skills. Okay? They are quite different here. In fact, we cannot, and when I say we, I mean human users cannot invoke the skills and use these skills.

They're actually built foran AI agent that, again, I won't really discuss in this hour much.

Okay, so I will skip skills for now.

You have scheduled tasks and you have shortcuts, which is great. Okay? Now I've pinned here a project. Now this is my pre-session run of the same project I'm going to run with you here.

Okay? Now, before we get started, I just want to let you know that we're going to use a public company as our target company in this process, and that's going to be Celsius Holdings.

So I want to make sure I say that Financial Edge is a training company.

There is, as far as I'm aware, no Celsius transaction in the market. We are not advising in any way in any transaction. Today is just a teaching scenario.

There is no investment advice provided. Okay? Now, let me start with setting up the deal project. Okay, so how do I set up a project? Well, if you go over here, you will see a section called Projects.

I'm going to open that, and there you see New Project, so let me click on that. Okay? And here you have your project, and this is, again, very similar to Claude, ChatGPT, and other tools. Right? So we're going to name this project Kelvin, just like that. Now here, I could paste my instructions, and I could do it right now, but I want to do it a different way, just making sure that some of you who've never played around or haven't played around with AI tools in general understand really how projects work a little better. So I don't want to get ahead of myself here. Okay? I just created the project. I do want to say this, very important, especially on this screen.

Rogo provides you with pre-built project instructions. Okay? So let me explain what instructions are.

These are project-level instructions.

So whatever you put on this window will be basically a higher-level prompt.

So as long as you're working within that project, any time you prompt the engine, it will read those instructions first, and then it will process other information. So these are project-level instructions.

So these are very important. So anything you know about the deal, anything that is specific about this sell-side process on Celsius, you might want to put here, and Rogo will read it every time, so you don't have to repeat yourself later on when you're just doing some of your work. Okay? So maybe actually I'll do this. Okay, I'm going to show you what these instructions might look like. Now, first, I want to show you the pre-built instructions. Now we're doing a sell-side process.

Down here, Rogo gives you the option to click on some pre-built instructions. These are generic pre-built instructions.

Okay, I'm going to click on one just so you see.

All you do is just click on it, and there you go.

You have your instructions in place. Of course, this is not good enough.

You want instructions that are specific to your deal.

Okay, so for example, it says here, "You're an investment banking expert built to support an analyst." So you're basically talking to the engine, to the model.

"Running a sell-side process for Project North." Well, guess what? We don't call this Project North. I would have to change that. Right? And as you can see, it's just really generic.

So what I've done is I've created a slightly different set of instructions, that I actually used this first bit, but I added more to it.

Okay, so let me copy-paste that into here.

Let me get rid of this and paste my own instructions. Okay? So it's a little more detailed, but again, this is a training session, nothing too crazy here.

We have the same pre-built section here, and I added who the target is, what the name of the project is, and some information about how the output should be delivered. Okay? Although a lot of this is already taken care of by Rogo itself. Great.

Just a quick thing that you might find interesting. If you click on Improve here, it would take whatever you have typed into the instructions, and it would run it through AI and enhance your prompt.

So it would enhance your instructions. Okay? Now, one thing I like to say every time is be very careful with this, because in improving your own prompt or enhancing your prompt, it may create or generate additional information that you are thinking, "Well, I didn't want that." Right? Full instructions. That's a great question, Stephen. Full instructions.

Yes, that's a great question.

By the way, that's actually a great question on two fronts, because there's a second part that I didn't want to forget to tell you about.

Features in Rogo are... Well, let me take that step back.

Features in AI tools are changing weekly, if not almost daily. Right? Little things.

So when you go into a tool, you will sort of pick up on new things.

Features in Rogo are also changing at the same pace.

Okay, so what I say today, and I'll give you examples later, might not be true tomorrow. Okay? All right.

Anyway, but most of it, 95% of it will remain the same for some time. Okay, let me create the project, and here it is. That's our project.

Okay, if you wanted to add files, you can.

You can add, I don't know, maybe a SIM, confidential information memorandum, indications of interest. You can add NDAs, due diligence lists. You have tons of information that you could add here. Okay? What I wouldn't add here is public filings, public documents, because all that stuff can be sourced directly from Rogo. Right? So let me skip this, and there it is. We've created our project.

It looks great.Here are my instructions if you ever want to edit.

There they are again.

If you want to ever add files, you can add them here. Okay? Now, let's get into the more-- So there's a second part to this.

Besides setting up the deal project, I also want to create a profile for the target. Right? Now, I could do it just by typing into the prompt box, but part of the reason Rogo exists is because it wants to make your IP workflow more efficient. So let's make our own workflow more efficient. Right? What I can do, instead of typing the prompt, I can use a shortcut. So let me explain shortcuts. Okay? Different ways of accessing shortcuts.

One is you can click on Shortcuts here.

You can also click on the little icon for Shortcuts here.

Okay? And it's going to browse shortcuts. Right? So I'm just thinking about something that there's a third place that is not popping up on my screen right now.

But it's fine. Okay? Let me click on Browse Shortcuts. Actually, no. Let's do it differently.

Let's actually go to Shortcuts here because that will take me to the big window.

Okay? Now, here is the place where I manage my shortcuts.

All right. Now, let me explain to you what a shortcut is.

There are Rogo built shortcuts. Every single box here is a shortcut. Okay? So all of these are Rogo built. So you can see here by Rogo. Right? You can star the shortcuts. You can duplicate shortcuts. You can share them as well.

I've created a couple of shortcuts just for overall teaching. Right? So you can see here my shortcuts, and I have one called Adjusted EBITDA, one called Transaction Comps, and I can edit those, duplicate, star those, and so on. Okay? Now, let me click on one shortcut to see how it works and what it is.

Okay? I'm going to use the public company profile shortcut, which is, by the way, pre-built by Rogo. Right? So all you got to do is click on it, and it would take you to the prompt, and there it is.

Now, it doesn't seem like much. Right? You say, "What do you mean?" It's just five words. Right? Write a profile about-- Actually four, and the fifth one is a field for you to enter the company name. Right? You can actually build this thing exactly as you see it. Now, if you're wondering, well, that's just a few words, that's not actually true. The main prompt, it's very short, but there is another layer of prompt behind it. So let me show you that.

If you click up here on Show Details, it would open up that larger prompt. So this is sort of like instructions provided to Rogo that are hiding behind your prompt.

Right? I hope that makes sense. So this entire thing is actually your prompt.

So what is this front end mini prompt? It's just there so you can enter the company name.

Okay? Now, before I actually run this, let me escape out of it.

Let me close it. Get rid of it. I just deleted the shortcut.

I didn't run it. You can also just quickly click on this icon, and you have the same list of shortcuts, but just a list, and you can click on one of them, like this one, and there it is again.

Or you can just type the forward slash key on the keyboard, and you get to the same place.

These are your shortcuts again. Okay? So different ways of getting to shortcuts.

Right? Let me run it. Okay, let me click on this.

Let me choose the public company profile. It says company.

I'm going to type the ticker here. Remember, this is natural language, right? I could type the ticker. I could just say Celsius Holdings.

I could say Celsius, the beverage company. Right? By the way, Celsius, for those of you who don't know, is a functional energy drink company, a US company, mostly selling in the US as well. So very popular here. Okay? I'm going to go ahead and press enter.

And there it is. Okay? Let me run this. Now, before I run my first prompt, let me just say a few things that some of you might be interested in. That's why I'm saying this.

Do you see here, Felix? Now, Felix is a mode within Rogo. Okay? But let me maybe explain this a little more carefully.

If I click on Felix, I get to choose the different models.

But these are not models as in GPT 5.5, GPT 5.4, or Opus 4.7, Opus 4.6, and so on at 4.6. No. These are modes within Rogo. Remember, Rogo is leveraging OpenAI, Anthropic, Gemini models. Right? So when you choose Instant, for example, in that case, it's going to be using GPT 5.5.

When you choose Deep Research, it's going to leverage the deep research features from OpenAI. When you choose Rogo Pro and Rogo Fast, in those cases, these are actually models that are fine-tuned by Rogo.

Fine-tuned for what kind of work? Financial Q&A, financial workflows. Okay? They are, of course, fine-tuned based on the foundational models, again, OpenAI, Anthropic, but they're fine-tuned for this kind of work. That's Rogo Pro, Rogo Fast.

Fast, of course, is specializing in speed.

Rogo Pro is specializing in depth. What about Felix? Felix is sort of like a router.

Okay? It will use whichever model will give you the best results for your given prompt. Okay? And that's actually a very new featureFelix as a brand and as a technology here. Okay? I'm going to keep it on Felix, and let's go and run our first prompt.

There you go. And I'm not going to wait for this.

It's going to be amazing, but I'm not going to wait for this. Okay? Instead, I've already run these things.

All right, so let me pin this, and let me show you what the end product will look like.

Okay? It will look like this. By the way, I ran this this morning, so it's going to look very similar. And there you go. Now, I want to just highlight a few things. This is a company profile.

Notice that it's in chat, so the output didn't come out as some other deliverable. It's in chat, so it's sort of like a working draft.

You can read it, and you can start from here.

Now, notice the structure.

It has part one, it's got a part two, and notice this, it's got a nice laid out table.

But look at this, the table has citations.

That's one of the key features of Rogo.

So if I go to one of these numbers, let's say this number here.

I'm not clicking on it, by the way. I'm just putting my pointer over it.

It says, "Oh, that number came from Capital IQ." Right? And in this case, most of the numbers will because of the nature of the data.

So it's pulling data directly into this table.

I can click on this, and it can take me to the source.

Okay. If you have a Capital IQ account access, it will actually take you all the way there. Okay? So that's a table that, by the way, you can copy, you can export as well. Okay? And then it's got part three, which looks at the product portfolio for the company. Right? That's what the products are. So here's what I want to highlight.

It is very structured. Why is it so structured? Because the prompt is not just, "Oh, give me a profile." The prompt is much more structured behind the shortcut. Okay? And that's how we get here.

That's our profile. Okay? Now, it's probably still running here. Oh, I moved away from it, actually. So I'm going to have to do it again because once you move away and you get no output, it will stop working. Okay, but you know what? I'll continue with this sample chat just for time savings. Okay, great.

So that's how you run a shortcut. Just one last thing here that is important.

If I click on shortcuts again, can I create my own shortcut? Of course you can.

Okay, so let's look at a shortcut that I created.

Let me open this up. And this is a way to create a new shortcut. You have your prompt. Remember, your prompt is your short message. What you actually see.

Then you have the additional instructions. That's what's behind the scenes.

So in my case, I made it very structured because this is for a precedent transactions output.

And then you can choose the mode. You can choose which sources to pull data from.

In this case, I chose all sources, but you might want to say, you know what, just use Capital IQ. Okay. Because that's going to allow Rogo to be a little more efficient, consume less tokens, because it's going to just browse through one to get you the output, and it's going to save you time.

So when you know what the source should be, I'll recommend that you pick that source. Only, not all sources. Okay? The default is all sources.

You could attach files if that helps you with, let's say, a template. Okay? And you could even say, "Hey, once you're done with this whole thing, also export it." And I will show you exports in a minute.

Okay, so this is an entire screen to build a shortcut. Then you save it, and there it is.

That's it.

Okay, great.

Now, so let's move on with the workflow. So what did I do next? I said, okay, now I have my profile.

Sorry, I got ahead of myself here. That's my profile.

Okay, so what I did, I said, you know what? Now that I have a profile, I put this prompt in place. Okay? I just want to get a first sense of what are potential strategic acquirers for my client. Okay, so I've already run this, but I just typed this prompt.

I actually just typed it in myself.

I said, based on the target profile just generated, so now Rogo knows the company well, what kind of strategic acquirers should we be thinking about in the buyer universe? Who are these companies that might buy my client? Right? And then I say, categorize by acquire profile and explain why each category is relevant. Okay? You have to appreciate the output here, how it's really meant to be for this kind of work.

So this is the output it gave me. Project Kelvin.

Strategic buyer universe framework. This is just a first pass.

And it said, look, for tier one, you have PepsiCo, because PepsiCo and Celsius have already a very strong partnership, a distribution partnership.

So that's likely a very appealing acquirer. Then it said, but you have a tier two of other global large companies. Remember, Celsius is trading at about 9 billion enterprise value, I believe, roughly around there, 9.5 or 8.5.

So the acquirers have to be larger in this case.

So they say, well, it could be Coke, it could be Keurig, it could be Nestle. And it tells you why these are potential acquirers.

Then tier three, other energy drink companies like Red Bull and Monster. These are really good potential buyers.

And then tier four and so on. Okay? And it gives you a summary at the end.

Okay, so this is the recommend. So now I have an idea of who my company is, and I have an idea of who might buy the company.

All right, great.Let's recap on what we've done up to this point.

So what have we done with Rogo so far? We've set up the deal project using projects.

We've generated a target profile using a shortcut, pre-built shortcut, and we've checked citations that are coming from Capital IQ.

The numbers are correct. Okay? So your main takeaways are, of course, related to this, right? Use projects for a single deal because projects compounds your system instructions, your attached files, prior chats, everything gets added to the context of your deal, right? And the tool is going to work smarter.

Use shortcuts. Not just use pre-built shortcuts, but build your own shortcuts for your everyday workflow, right? For things that you run on a regular basis to have one click or two clicks, and there you have your input on the way.

And finally, appreciate citations, right? And verify citations, because we cannot work without citations in this kind of work. All right? So let's move on to the next part of our workflow, and it says, well, with the target profile complete, the next deliverable is the buyer book.

So the MD wants two answers, but two answers in one artifact, in one output deliverable. Number one, who are the strategic acquirers for the target? By the way, we already have a sense for it, right? Like Coke and PepsiCo and Keurig and Monster. Right? We just looked at the first screen.

But there's a second part, which is which of these companies can actually afford Celsius? Celsius is not a small business, right? So who are the strategic acquirers for this target, and who can actually afford it? Okay? So what we're going to be building is a buyer universe list, and we're going to attach to it a capacity matrix, right, to answer both questions. Okay? And at the end of this, we're going to end up with a handful of companies in our shortlist that we say, okay, these are the companies that are most likely interested in this target company.

All right? So what features are we going to look at in this next step? Number one, screening.

I want to be very careful here, right? Screening is not what I just did.

It's not just a prompt. It's actually a feature and a tool within Rogo, and it is deterministic. So as you already know from all this AI content available today, AI output is probabilistic. Right? You can put the same prompt and the same context and the same files, and you're going to get slightly different outputs, right? Because AI by nature is based on probabilistic models. Right? But the screening tool within Rogo is deterministic.

You put the same prompt, same criteria, and you're going to get the same output on your screen. Okay? It is basically a feature that is not related directly to AI, but is extremely useful if we want to sort of work within Rogo and not have to leave Rogo, use another tool, and come back. Okay? So I'm going to show you that one.

I'm also going to show you the professional grade exports, specifically when it comes to Excel or spreadsheet exports.

So we're going to look at that.

And I'm going to show you the spreadsheet feature. Okay? It's just one of the features that you can use in that way. Okay? So let's jump into it. Our sources are going to be the same sources, Capital IQ and Public News.

And we're going to build some great deliverables now.

Okay, so let's look at the workflows before that.

The first thing I'm going to do is I'm going to screen the buyer universe using the screening feature.

Then I'm going to build the capacity matrix using the spreadsheet feature.

Then I'm going to export my output into Excel using the export feature. And once I have that, I am going to narrow down my list to only four buyers. Okay? And I think you have an idea of what they will be. Okay? So let me go back to Rogo, and let's get that done here.

All right, so let's go back to our project Kelvin here.

Now, remember the first part I ran, but because I moved away from it, it didn't actually get registered here. Actually, yeah, that's correct.

Let me just check something. That's correct.

Well, that might not be the case. No, it did get built. I just didn't do it within the project. Is that possible? Hmm.

Okay, so I might have done it outside the project.

So it didn't get added to the project. I just added it.

Okay, you can click on any chat in your history, and you can right click on it, or click on the three dots and say, "Add to project." Right? So now you see here, there is the output that we generated earlier. Great. Now I can continue from this.

So let's get to work.

Let me go ahead and do the first part.

Okay, and the first part is the screening feature. Okay? Now, do you see here these five features? I call them quick actions. Okay? Now, if you want a screening table, use the screening tool.

If you want a PowerPoint export or a PowerPoint deliverable, use the presentation tool.

If you want an Excel spreadsheet, use the spreadsheet tool.

Okay? And so on. Right? So let's go ahead and type my prompt.That's my prompt. My prompt says, "Find publicly traded companies globally with a market cap over $5 billion." They need to be large because the target is large.

"Including non-alcoholic beverages, energy, soft, bottled water.

Return company name, ticker, market cap." So very specific, right? You could even make this more specific with a shortcut if you want to build it. Now, I'm going to click on the screening tool, and notice how the screening feature gets added to the prompt.

Now Rogo knows to build a deterministic screening table from this. Okay? That's all I need.

I'm not going to do anything else here. Okay? I'm going to click on go.

And now it's running a screen for me, okay, that is going to actually take my criteria. What is the criteria? Over $5 billion in market cap from these specific sectors, global companies, and I'm asking for specific output like LTM revenue, EBITDA margin, country, and so on. All right? So this thing is going to run, and I'll let it run. Okay? In the meantime, I'm going to go to my output so you can actually see it. Okay? Again, imagine time went by.

Yeah, time went by and we're here, and we already ran the screen. So this is what the output will look like.

Let me go all the way up. There you go. Same prompt.

I did this this morning. Okay? So now it gives me what? Here is the list of companies.

Notice that it's giving me all the information, and notice that I have citations so I can check my work just in case. Okay, but I'm going to show you one more thing here that is unique, and that's the reason I'm using the screening tool.

So it gave me about 18 companies. All right? And some of them are ones that we've already mentioned, like PepsiCo, Coke, Monster, Keurig.

Actually, it also included Celsius in the list, so just be mindful of that. Right? So Celsius is number 16 here.

It's because it's just doing a screen. That's it. Right? I could improve my prompt and say, "Don't include Celsius." Right? That's it, simple. Okay? And then it gives me some observations as to why this list is relevant for my specific deal.

Okay? And then some caveats. In this case, it's telling me that some companies were not included because of the way Capital IQ tags companies into their industries. Okay? But it actually went and found the companies. So it said names like China Resources Beverage. Right? So it's not like it-- It knows there are companies that are not tagged properly in Capital IQ that it found for me.

Okay? So it's giving me that additional layer of information. It's quite important.

All right. Now, what am I getting from this that is different? This is what I'm getting from this.

You can click on any of the citations if you want, or you can come down here and you will see that it created a screening output. So let me click on that.

It's going to open up sorry, the side panel, and it's giving you your actual screening output. And you can see how this is very well formatted. It's got all the information you need or that you asked for across all of the different companies that were listed in chat.

It tells you all of the criteria. There's five criteria. Right? Beverage industry, market cap, beverage category.

So it tells you the screening criteria.

You can actually hide this or show it, up to you.

And you can iterate this. Of course, you can iterate this, right? We can always iterate these things. Okay? Great.

So what happens next? I have my screening. It's deterministic, by the way.

If I re-run it with slightly different criteria, it's going to change. But if I re-run it with the same criteria, it's going to give me the same list. Okay? I'm not talking about in chat, I'm talking about this screening result.

Okay? Great. Now what do I need? Now I need to create a capacity matrix. So basically, for each of these companies, I want to find out if these companies can afford to buy Celsius, given the size of Celsius.

I mentioned it was a market cap of $8 billion, $9 billion. Enterprise value, a little higher, $9.5 billion, roughly. Right? Big company. Okay? So what I'm going to do now is I'm going to...

You can export this information, okay? Now, I've already done this, but I can show you how you would do it.

You click on the download, and there you go. It just exported the information.

Okay? It put it on my downloads.

So I'm going to open up my file so I can just show you what it gave me. Really straightforward.

It's just a table.

Right.

It's taking a moment to open. Maybe already open. Let me check. No.

So while this is, feel free to ask any questions. Happy to answer questions here.

It's a good moment for that.

All right, so what happened? That's my table. That's my data.

Okay? Now, it uses the logo, sorry, the Rogo template because I have not provided Rogo with my own specific template. But this is what I really want, right? I want all of these companies. That's my universe of buyers. Okay? So I've already downloaded this.

You just saw how to download. I'm going to close it now.

I am back here, and I'm going to use that spreadsheet now because I'm going to do the following.

I'm going to ask Rogo down here.

That's my prompt now. Okay? By the way, this could be a what? A shortcut called buyer capacity matrix shortcut.Right? But in my case, I'm just building this for the session.

Build a buyer capacity matrix from the attached list of companies.

So let me show you attachments. I will just click on this clip here, and I will attach the file that I already downloaded.

And now there it is. That's the Excel file I just showed you. Okay? And I'm saying, "Hey, look at this, give me revenue, EBITDA, net debt, leverage.

Give me market cap." But look at the last thing I'm asking to do.

I'm asking it to build a capacity to pay formula. I'm asking it to be two formulas, if you notice.

Net leverage and capacity to pay. Two formulas. Okay? I'm basically saying take the EBITDA of the company, multiply it times 4.5.

That's me thinking that's the most the company could borrow, right, and keep its current rating.

And add any cash they might have and take away any debt they might owe. So I want to see how much room they have, to borrow, basically. Okay? At a 4.5 multiple. Right? And then I say return all values, and I specify cite each cell to its source filing.

Okay? So I am asking it to go back or go into the source of truth here. Right? So when you run this...

Let me just check something.

Yeah. When you run this, what's going to happen? Well, it's going to take a moment, because what is Felix do-- Sorry, not Felix.

What is Rogo doing right now? Actually, Felix as well.

Felix and Rogo are both one and the same thing. Sorry, I just got distracted by a pop-up.

They are actually building an Excel file with formulas, and they are building formulas for me, right? And the data is being pulled from a specific database.

In this case, actually, it's being pulled from 10-Ks and 10-Qs and actual source documents, not just a data service company. Okay? So this might take a moment.

You never know, right? But it might take a moment.

Steven, the question is a great question, by the way. Two great questions.

The question here is, do you have to give it the exact formula, or if you just say capacity, would it know what that means? So I'm going to give you an answer that might not be that satisfying. Okay? You can try not giving it the exact formula, but the benefit-cost trade-off is not worth it. Right? Because the tool, the model might drift. It might not interpret correctly.

Right? It might give you a formula that is not what you wanted.

So I always say this across all of my sessions, right? If you're dealing with AI tools, you have to be specific. It's for your own benefit, not just because you want to be precise.

Of course, you want to be precise because you're going to save time, right? I'd rather type the formula. So Felix follows exactly that specific framework than just see what it gives me and then iterate. Right? When you know what you want, put it in there.

If you're not sure, then you can iterate, of course. Okay? It's not the best answer, but honestly, that's the only thing I can say there.

All right. Zoe, I'm going to get to that very soon, okay? So Zoe asked: What are the main differences of using the screening function within each chat versus the table function in another section? I'm going to do it now. What's coming is that table function.

Well, huge differences. Okay? The screening function, which you can see here, is deterministic.

Right? By the way, a question that is even-- Oh, it looks like it's restarting again. But a question that is even broader is screening versus Excel spreadsheets versus tables. Right? Well, screening is deterministic, but it's not an export.

You can export it, but it's not that, and a spreadsheet is actually going to be a working Excel document where you can interact with the formulas directly.

And then tables are AI cells. That's the best way I can describe tables, AI cells. Each cell in a table is an AI prompt output kind of window. So you'll see that next. Okay? All right.

I'm not going to wait any longer for this because it's not pulling up, but I want to show you what the actual answer is.

So I'm going to click on it, and this is what should have happened, and it will happen eventually. It will work out, as you can see here, buyer capacity matrix at 4.5 x capacity to pay, it computed that. You can see how Coca-Cola, PepsiCo, Monster have huge capacity to pay.

Monster because it has no debt, Coca-Cola and PepsiCo because they're huge, right, and they have tons of assets and cash.

So these three companies at the top, and actually the fourth company as well, have tons of room if they wanted to acquire Celsius. Right? So that's the whole point.

But if you look at Keurig Dr Pepper, look at that.

In that case, we have negative capacity to pay, and that's because Keurig Dr Pepper has a huge amount of debt still lingering from back when Keurig and Dr Pepper merged.

I think it was 2020, 2018. A few years back.

Okay? Now, look at what else it did.

It created this, and that is an Excel sheet. By the way, you see these errors here? I noticed this earlier this morning.

These are actually not errors once you open up Excel.There are formulas that are in this, they're not rendering properly here. Okay? But what I want to show you is this, it created this entire thing.

I mean, you can actually just deliver this to your supervisor, right? And say, "Look, I did the work." Okay? A couple of interesting things here.

First, it is color-coded, just like it should be for a professional-grade banking kind of output.

The next thing is nicely formatted, and it looks even nicer when you actually open up the file. I won't do that just to save time.

But look at this, it has citations.

For every single number, it tells you exactly where it came from.

Right? And this all came from 10-Ks and 10-Qs because these are LTM numbers.

Cash on hand came from cash and short-term investments in the 2025 balance sheet for the company's 10-K/10-Q, depending on which one is the latest.

But wait, look at the capacity to pay. If you click on any of them, look at what happens.

It shows you there's a formula. It takes this number, the EBITDA, it multiplies it times the 4.5, it adds the cash, takes out the debt. That's it.

It's basically a fully built Excel workbook with your answer. The other thing that you might find interesting, and I do, is that the output, Rogo recognizes that this is a 4.5 leverage assumption, and it's giving you the option to change it.

It's saying, "Okay, look, here is your output." But you know what? If you want to change or if you want to flex your leverage assumption, here is the cell. It's in one cell. So it didn't hard code it into the cells. This looks like a real financial model. Okay? And this is the benefit of Rogo, right? It helps you with your IP work directly. All right.

Now, we have a few minutes, and I want to show you tables.

So what's the last bit? Let me go here and let's talk about the last bit, the last scenario.

By the way, from this capacity matrix, I'm going to pick a few companies that I think could afford Celsius, okay? And that's going to be Coke, Pepsi, Monster, and we'll leave it at that.

That's our priority subset. So with the priority subset locked, the deal team needs two things. It needs a table that tells me about the buyer's appetite for M&A. Okay? And this is going to be qualitative now, right? Because I need to scrap all the documents, and I need to find out what the potential buyers, what is their stance in terms of M&A.

So I need to find out what each buyer has said publicly about their own M&A appetite, what kind of targets they would like to acquire, whether anything material has recently shifted.

Maybe a new 8-K came up, maybe the CEO left, a new CEO came in. Now, that could be one thing that you need, and I'm going to build that. But the other thing is you want this thing to refresh every week. Why? Because this process is going for 4 to 12 months, and I want to keep my buyer appetite monitor, that's what I call it at the top, I want to keep it updated. What if new documents came up and the company said something new about what they think about their M&A strategy? So notice two things here.

I need this to be refreshed. A lot of the data is qualitative, and this will take me so much time to actually get on my own by going through every 10-K, every 10-Q, going through some sort of database. Okay? We're going to do it now with Rogo, and it should, at least as a first pass, it should be pretty easy. All right? So let's look at the workflow.

I'm going to pull the buyer data. By the way, I'm going to use three companies only. I just mentioned them, Coke, Pepsi, and Monster.

I'm going to extract on a per buyer basis their appetite signal, okay? And I'm going to use tables for that.

And then at the end, I'm going to actually make it a recurring workflow.

So let's go to Rogo in the last few minutes.

Let's go back here. I'm back in the project. This time, I'm going to click on Tables, and I'm going to create a new table. Okay? Brand new.

There you go. That's my table. This is brand new.

Again, I'm going to need to do this for how many companies? You could do it for 20 companies. I'm going to do it for three.

So I'm going to just paste the list. I'm actually going to type it.

PepsiCo, Coke, Monster. I could have 10 or 20 companies. Okay? Now, watch what happens when I add the rows.

Now, Rogo put each ticker here, and now my table is ready to go. By the way, I could just type here as well and add another company and keep on adding rows, right? But now let's actually get our data.

Now, for this, I pre-built some stuff, and I'm going to show you the finished product in a moment, but I want to see how you would do it yourself.

I want to have a column called M&A stance. What is the company thinking about their own M&A strategy? So I'm going to click on Add Column, and instead of using the prompt window here, I'm going to add it manually, and I'm going to call it M&A stance.

And now I am going to take a prompt that I already built, okay, and put it here. This is a prompt, just like I would prompt anything in any tool. Okay? And I'm going to go save. Oh, by the way, you can say, do you want the answer to be text? Yes. In this case, yes. Do you want it to be a number, currency, a date? You can choose the data sources. Where do you want the data pulled from?I'm going to click on save and look at what happens.

What happens is each cell is an independent AI output cell. It's going to read the prompt from the column, M&A stance, and it's going to process it for each ticker.

And it's going to give me the type of output I want.

It could be qualitative, it could be quantitative, depending on how I set up the column.

Now, let me show you the finished product.

This is my finished product, and look at what it did.

Now, in this case, I have more companies, but look, M&A stance.

Let me expand this.

For Coke, Coke is selectively acquisitive. John Murphy said with respect to acquisitions, and it gives you the exact source. This takes me back to an earnings call that I would've had to actually read or summarize.

It's all in one cell. Every cell is independent.

I can refresh each cell.

If I want to refresh each cell here, I can rerun the cells. I can do so many things.

Each cell is basically a prompt. Each cell is basically an output.

Okay? So lastly, I know we're almost at the hour, but I cannot stop the session without showing you how you will take this and make it a recurring update. Okay? So I obviously have more information, like I have disclosed deals. I have what is the largest deal size in the last two years.

Rogo is refreshing now.

And what you can do once you build your table and you're satisfied with it, is you can click on schedule, and it will take the table name, and you can say every weekday, or every week on Wednesday, or custom, every week on Monday at 9:00 a.m.

I want you to refresh this table and have it ready for me.

But even better, you can provide your email, which is your email on the account for Rogo, and you can enable this, and it will send you an email every Monday at 9:00 a.m. with the updated table.

And that's how you would schedule a task.

Great.

I know it was a lot. I wanted to give you enough from Rogo so that if you ever play around with it or if you're working with it, you can use these features.

Hopefully, this was a useful session for you.

Please make sure you fill out the survey that Sophie just put in the chat. It's a quick survey. It was great working with you this hour. I hope you found it helpful, and we'll see you in a week.