- 6 months ago
Join Forbes' Assistant Managing Editor Katherine Schwab in a captivating discussion with Gabriel Stengel, CEO of Rogo, an innovative company bringing cutting-edge AI to Wall Street. As a former M&A banker himself, Gabriel offers unique insights into how AI is poised to transform the finance industry. Will these powerful tools replace the roles of human bankers, or are they paving the way for a new era of financial work? Dive into the conversation to explore the shifting landscape of Wall Street.
Read the full story on Forbes: https://www.forbes.com/sites/amyfeldman/2025/08/12/next-billion-dollar-startups-2025/
Subscribe to FORBES: https://www.youtube.com/user/Forbes?sub_confirmation=1
Fuel your success with Forbes. Gain unlimited access to premium journalism, including breaking news, groundbreaking in-depth reported stories, daily digests and more. Plus, members get a front-row seat at members-only events with leading thinkers and doers, access to premium video that can help you get ahead, an ad-light experience, early access to select products including NFT drops and more:
https://account.forbes.com/membership/?utm_source=youtube&utm_medium=display&utm_campaign=growth_non-sub_paid_subscribe_ytdescript
Stay Connected
Forbes newsletters: https://newsletters.editorial.forbes.com
Forbes on Facebook: http://fb.com/forbes
Forbes Video on Twitter: http://www.twitter.com/forbes
Forbes Video on Instagram: http://instagram.com/forbes
More From Forbes: http://forbes.com
Forbes covers the intersection of entrepreneurship, wealth, technology, business and lifestyle with a focus on people and success.
Read the full story on Forbes: https://www.forbes.com/sites/amyfeldman/2025/08/12/next-billion-dollar-startups-2025/
Subscribe to FORBES: https://www.youtube.com/user/Forbes?sub_confirmation=1
Fuel your success with Forbes. Gain unlimited access to premium journalism, including breaking news, groundbreaking in-depth reported stories, daily digests and more. Plus, members get a front-row seat at members-only events with leading thinkers and doers, access to premium video that can help you get ahead, an ad-light experience, early access to select products including NFT drops and more:
https://account.forbes.com/membership/?utm_source=youtube&utm_medium=display&utm_campaign=growth_non-sub_paid_subscribe_ytdescript
Stay Connected
Forbes newsletters: https://newsletters.editorial.forbes.com
Forbes on Facebook: http://fb.com/forbes
Forbes Video on Twitter: http://www.twitter.com/forbes
Forbes Video on Instagram: http://instagram.com/forbes
More From Forbes: http://forbes.com
Forbes covers the intersection of entrepreneurship, wealth, technology, business and lifestyle with a focus on people and success.
Category
🤖
TechTranscript
00:00Hi, I'm Catherine Schwab. I'm an Assistant Managing Editor at Forbes covering technology,
00:08and I am so excited to be here with Gabriel Stengel, the co-founder and CEO of Rogo,
00:13which is building AI tools for Wall Street. Gabriel, thanks for joining me.
00:20Catherine, thanks for having me. I'm excited to be here.
00:21Great. So let's just start with the basics. What does Rogo do? How are you helping bring AI to
00:30Wall Street? Rogo is an AI analyst for finance. We take cutting-edge AI researchers and sit them
00:36next to financial experts. I was an M&A banker at Lazard. Our team is filled with folks who worked
00:41at Goldman, Blackstone, and Apollo, people who understand real-world financial services use
00:46cases. There was $3.5 trillion of M&A and deal-making last year that was done in tools that
00:53were made in the 80s, PowerPoint and Excel. We believe that AI will allow everyone to be a lot
00:57more insightful, to be smarter, to do a lot of this busy work more quickly, and focus on the more
01:02strategic, judgment-oriented work that investors, bankers want to be doing.
01:08Okay. So what makes Rogo's AI models especially tuned for finance? Why not just use ChatGPT or Cloud?
01:16The same way that, you know, when you're hiring someone who's gone to a great undergraduate
01:21institution, the first thing you want to do is teach them about how you work, the tools that you
01:25use, the data sets that you use, and the way you think. What we need to do when working with
01:29investment banks is take these great models like ChatGPT and Cloud and teach them how to think like
01:34an investor, how to use the data that an investor uses, whether it's public company financials or
01:39private company data like PitchBook, and then teach them how to do the same kind of outputs,
01:43the PowerPoints, the Excel models, and so on. And so we actually take those cutting-edge models,
01:48and we train them to more specifically do financial workflows. And so we have a lot more data,
01:53we're more accurate, it's lower latency, and it can actually address the pain points that these
01:58professionals have much more quickly because we can take a targeted approach to creating the technology.
02:03Okay, so when you're saying you're feeding this data into these, can you be more specific about like
02:13what that process actually looks like? Is it specific to each client?
02:17Yeah, and we partner broadly with FactSet, with Refinitiv, with CapIQ, PitchBook, Prequin,
02:24Crunchbase, just to name a few. So our AI agent has access to all of those tools. So if you go in and say,
02:31hey, Rogo, tell me about a company like Perplexity, we can actually go through and
02:36look at Crunchbase and PitchBook data to tell you the fundraising history and give you data that you
02:40trust. If you say, great, put this into a PowerPoint template, it can go and look at the last PowerPoint
02:45you made on a company and follow that exact same workflow and template too. And so we both connect
02:50to a company's internal data sets, their SharePoint, their precedent workflows, as well as the third-party
02:55data and content that they already know and trust. Okay, so I'd love to hear more about
03:01your time at Lazard. How many hours were you working every week?
03:07You know, I love my work at Lazard, honestly. I mean, it's very interesting work. You put a 22-year-old
03:13in a room with the CEO of Fortune 500 companies. I didn't get to say a whole lot in my first few
03:18years, but I got to see how they thought and how they thought about huge deals. And it was crazy to me
03:23that a lot of these deals were powered by analyses that I was doing late at night in old, archaic
03:29tools. And people spend a lot of money on bankers and they spend a lot of money on data. And really,
03:34what they want is insight or something needle moving, right? You want something that's going
03:38to change the strategy of what you're doing. And it was very hard to ever elicit that for clients
03:43when you're spending so much time in a data room or doing the process work. And so it was very rewarding,
03:49and I got to see a lot of what could be going on. And then I also got to live the reality of,
03:54hey, what actually happens behind the scenes? How did that time in your career lay the groundwork
04:00for Robo? You know, I saw firsthand that a lot of the work I could be doing could be completely
04:07augmented or replaced by AI. And so when GPT-3 came out, I was using it a ton. I was developing little
04:13tools. I was both a banker, but then I also was on their sort of software and data team
04:18trying to figure out internal use cases. I'd actually in undergraduate life published a paper
04:24on AI assistance for finance. And so at Lazard, they had kind of staffed me on some of these
04:28exploratory projects. And when GPT-3 came out, it became really clear that no longer was a managing
04:34director going to have to ping me at 11 PM and ask for revenue multiples on a company. They were going to
04:38be able to do that directly themselves. And so that's when we left to get started building Rogo.
04:43It was far beyond the point where these tools were actually capable of doing that yet, but the writing
04:48was on the wall that front office work was going to get more and more automatable on Wall Street.
04:53Okay. Interesting. You did use the word replace. So I'm curious, do you see tools like this fully
05:04replacing junior bankers? I think tools are going to replace work and workflows. I don't know if
05:11they're going to replace people. You know, the best bankers on Wall Street are not the best bankers
05:16because they're way better at Excel than everyone else. They're better because they have relationships.
05:20They can think commercially. They know who to call to get a deal done. A lot of the work that takes
05:25place now. Yeah, sure. It'll be replaced, but it'll transform and allow people to focus on more
05:29strategic, more interesting work. And the org structures might look very different, but you
05:35know, I think net new, if you can make more money on better margins, you'll probably continue growing
05:40and continue hiring. And at the end of the day, these are not institutions that are going to kind
05:44of just take a step back and be like, oh, we're, you know, we, we can grow so fast, but instead we're
05:48just going to, you know, lean the team and return more to shareholders. No, they want to be aggressive
05:53and take market share. They want to compete.
05:54What were those early experiments like that you were doing, uh, with AI kind of
06:02re-chat GPT and where was it falling short and how have you seen this technology kind of develop
06:10over the last few years? The early experiments were like, you would, you know, ask a question,
06:16you get an answer that looks right. And then you would take a closer look and you'd say, this is
06:20absolute gibberish nonsense. Uh, uh, the trajectory of AI and finance, I don't think is that different
06:25from the trajectory of AI broadly, which was two, three years ago, you know, chat GPT was helpful for
06:31writing poems in the style of Roald Dahl or Shakespeare and like doing kind of like fun stuff.
06:37And then over the last year and a half, you said, Hey, maybe this is helpful for sequel or for
06:41copywriting. And today I can use chat GPT for just about anything, right? It's a thought partner.
06:45I can tell it about my personal life. I can ask it for help drafting email. And I think that's what
06:50we're seeing in finance too, where the sort of breadcrumbs were there, that it could be helpful,
06:54that it was starting to think, you know, it was the first time in human history, you could say, wow,
06:58that is a thing capable of thoughtfulness. That's not a human, but it wasn't really that useful.
07:03And now as the plumbing has been getting done, as the models have become smarter, as they've been able
07:08to connect to the right data sources, integrating the right systems. I think there's, there's very few
07:12workflows that won't be able to get automated by these sorts of tools and models.
07:18How did you solve the hallucination problem? If to take your example of like, you know,
07:24instead of a senior partner having to ping a low level analyst, like what are the revenue multiples
07:32on this company at 11pm at night? Like, wait, is it good enough for that? Is it good enough for kind
07:40of a full report? And what are their trust barriers there in terms of double checking
07:48to make sure? There was one, there was one instance when I was an investment banking analyst,
07:53when I think I sent out materials that had the complete wrong number for the share price of a
07:58public company, and it went to a company CEO, that was a human error, right? That was a human
08:02hallucination. And I lost trust with my associate. And the next time my associate, she was great,
08:07and she checked a lot of my work. And eventually I earned her trust back and could be autonomous.
08:12These tools are probably already more capable than I was as a junior banker. And maybe I wasn't
08:16a great junior banker. They still hallucinate. I mean, there's no solving it. It's like saying,
08:20can we solve every car crash through self driving cars? No, but we can get to a point where we trust
08:26them, they obviously add value, we still want to check them. And we need to integrate them into the way we
08:31work, the way that we integrate humans and human error into the way that we work.
08:35Okay, so walk me through kind of a new workflow that would not have been possible when you were
08:43an analyst, like with a specific example, just to, I'm trying to kind of get at like, what are those
08:49kind of specific tasks beyond just kind of getting answers to like, how much does this public company
08:55stock trade for? Imagine a year ago when Dylan Field was thinking about taking Figma public and
09:02he's, you know, someone got wind of that and you're an MD at Goldman Sachs or whatever bank ended up,
09:08you know, leading that leading that IPO, you say, Hey, we need to get in front of Dylan in 24 hours
09:14and pitch him on why we should take Figma public. That deck used to take a week or two weeks to create
09:20that bake off deck. It was, it was coined or termed, you know, you have to put together the
09:24precedence analysis. What is Figma going to be worth? What's the story that you want to tell
09:28to Wall Street and the investors? What's a draft of the S1? That takes a while to do. Now someone
09:33can go into a system like Rogue and say, Hey, who are the comps? Who are the peers? How do they trade?
09:38What's the story that might resonate today on Wall Street? Can you draft all of those materials
09:43in 30 seconds so that I can go and iterate and be done within 24 hours so that I can get in front of
09:49the Figma management team before anyone else does? That's what's happening.
09:53Was there a kind of light bulb moment for you when you decided to start a company based on this idea?
10:02No, I mean, I wish I was, you know, had been prescient and just knew this was all going to happen.
10:07I was really excited about the technology. I've always been a kind of technologist at heart and
10:13GPT-3 felt like magic. And the work that I was doing in investment banking did not feel like magic. And
10:18I thought there was a really clear opportunity to kind of bridge this coming technology with the
10:23work that I was doing. I had no idea how fast it was going to go. I had no idea, you know,
10:27all the use cases that would emerge. And I think it was a, you know, sort of a story of
10:32right person, right place, right time. I was passionate about the technology. I had the investment
10:37banking experience. I lived in that world. And so I was able to, you know, connect with the right people
10:42to get started. And then it's a game of, you know, making sure you make the right product bets and
10:47make the right bets on what you're going to build. But, you know, it was not a, I did not,
10:51you know, sort of sit from an armchair and envision the world we'd be in now three years ago.
10:57So correct me if I'm wrong, but you started the company before kind of the big chat GPT moment.
11:04So what was it like in those early days kind of before everyone woke up?
11:10Yeah, it was, it was horrible. You know, it was, I felt like I was selling snake oil,
11:16right? Like, it's like you say, hey, AI is now capable of thinking, it's going to be able to do
11:21so much of the work you're doing. Why don't you try this tool? And people go, what, what on earth
11:25are you talking about? Chat GPT was a huge educational moment for the world. GPT-3 was out. And so that was
11:32the catalyst for me. I was, you know, a lot of people saw it coming before GPT-3. GPT-3 was the
11:38moment I went, wow, this is going to be a very capable tool and technology. But, but the world
11:44hadn't seen that yet. And I think partly what Chat GPT did, the consumer tool is now more and more
11:49people were using it in their daily sort of personal life and starting to see how profound
11:53the impact would be. I think that was enormously helpful for us, for enterprise, for industries to
11:59think about adopting the technology. What were those early pitch meetings like
12:05with investors kind of pre-Chat GPT? You know, I still shudder to think about the
12:11deck we actually use for our pre-seed round. We didn't call anything a chat bot. We didn't call
12:15anything AI because we thought it was kind of cringy and AI sounded a little bit like snake oil.
12:21We called it a natural language interface instead of a chat bot, which is the worst, least salesy,
12:27marketable term you could possibly think of. And so it wasn't really, you know, it didn't resonate.
12:33And I, we didn't know how to sell it. And partly what we knew was, hey, we know what the financial
12:37workflows are and we understand the technology. And what Chat GPT and AI did for the world was
12:43teach everyone how to sell and teach everyone how to apply it. And so then as that was taking off,
12:47we were able to, you know, take our unique experience and our unique insights and apply
12:52it to this vertical. Where, when did you feel like you kind of found that product market fit and,
12:58and started to build that momentum post? You know, we, when we first started deploying
13:06the current version of our tool, because we went through a lot of iterations,
13:10the current version of our tool, and it just started getting a lot of use, right? Like you just
13:15saw that something that investment bankers, juniors and senior folks were using every day to supplement
13:21and replace some of the work they were doing. And I remember one of the magic moments I had was when
13:25an enterprise client came to me and said, hey, Gabe, you know, the web app version of Rogo is
13:30great, but we need a mobile app. Our senior MDs are asking for a mobile app. They need a native iPad
13:36kind of UI and integration because they need it on the go. And I thought to myself, well,
13:41I barely knew what MDs did when I was a junior banker. I was not expecting it to cater to, you know,
13:46both senior bankers and junior bankers. And the fact that I'm feeling this pull by users
13:51to get to use this thing that makes them smarter. I mean, that was a real kind of light bulb moment
13:55for me. And that was probably 18 months ago or so. Okay. Yeah. I wanted to ask you about that.
14:03If you, because a lot of what I've read about the company and my understanding was that this was kind
14:07of targeted toward junior bankers and analysts. So, but it sounds like many more senior people are using
14:15this. So what's the product roadmap look like to kind of penetrate every level investment bank?
14:26Our goal is to be the best analyst on your team, right? You know, the best analyst on your team is
14:31someone you can reach via calling them. You can text them, you can send them an email and outlook
14:35and ask for a deliverable. That's the most natural way to work with a coworker. We think that's going to
14:40be the most natural way to work with an AI analyst. And for me, you know, I want to be the tool and the
14:46teammate that is as kind of seamless as working with a real junior banker. And if I'm a managing
14:52director, there's going to be a whole category of questions where I'm just going to think I'm
14:56going to ask Rogo instead. I'm going to ask Rogo in order to get these results more quickly.
15:01And then there's going to be a whole category of work where it's, it's more helpful for me to engage
15:05with a human and to sort of have an intellectual sparring partner where I don't need to turn to Rogo.
15:09But our roadmap is, you know, how do we do the analysis that takes so much time
15:15and kind of informs the more thoughtful work that senior and junior bankers do?
15:21And tell me a little bit about your business model and your customers right now.
15:28So we sell a co-pilot like product, right? It is. It helps with workflows. It helps with PowerPoint. It
15:34helps with Excel. It helps with research. And we sell it per seat to some of the world's largest
15:39investment banks. On top of that, we will soon be selling full work products. So the ability to
15:44click a button and get a full, you know, company profile page of a target acquisition you're looking
15:50at or the ability to click a button and get a full working DCF model for a company too.
15:55But right now we are, we're still in the world of selling the co-pilot.
15:58Got it. And so when you say a full work product, do you mean like a new app,
16:04like a separate app or a website or what will that look like?
16:08I mean like a deliverable. Like if I think about, you know, what are the work deliverables that you
16:12might create in your day to day? One might be a PowerPoint page on a company. Another one might
16:18be an Excel model that, you know, models out the future cash flows of the company. Another one might
16:23be an earning summary that you email out to your team. I don't just want to be the tool that you
16:28use to do research as you assemble those materials. I want to be able to know exactly what you're doing
16:33in your day to day. So you can go in and button and generate that work product that before might
16:38have taken you a week or two weeks of work. Got it. Got it. Got it. Okay. So have you found that
16:44large banks are building their own internal competitors to what you offer? I mean,
16:52they obviously can come by your tool or they could build their own because this is built off of
16:59basically commodity large language models. So do you see that as a source of competition?
17:06So I would say two things. One is the most sophisticated institutions are doing all of the
17:11above. They're trying tools, they're adopting tools, they're building their own. We're still in,
17:16you know, minute one of inning one of generative AI. It's so early and the best institutions
17:23are being very thoughtful and experimental. I would say we have a unique advantage in that,
17:28you know, we see the scale of, of not just a thousand bankers or a thousand investors
17:32using the tool. We get the feedback of tens of thousands of hundreds of institutions,
17:37and that compounds into both a better model, more workflows and a better product experience.
17:43Okay. Yeah. That makes sense. Who, who do you see as your biggest competitors then?
17:49You know, when we're, when we're selling, there's all sorts of tools that address
17:54pretty minute point solutions, right? So there's all sorts of tools that can just cover, hey,
17:58I can just do diligence, or I can just look at, look at, you know, benchmarking a bunch of
18:03credit agreements. What we tend to see is that clients want a solution that is
18:08one platform that connects to all their data and understands all their workflows.
18:12And really there's very few people, if anyone in market, who's able to offer that for finance.
18:17And so, as you said, often there are internal tools working on all sorts of things.
18:21I think most folks see us as, as additive additionally, and there's less of a kind of
18:25direct one-to-one comparison for what, for what we're doing, frankly.
18:33I'm curious about the role of the junior banker, given this technology. We talked a little bit about
18:42does this replace those people, but I also, I am not a junior banker, so I don't know, but I imagine
18:50that the practice of building a revenue model or pulling together all of this information would
18:57be like a really good way to learn how to be an investment banker. So I'm curious, kind of from a
19:03high level, what you think about the problem this will create in terms of training people and what those
19:12kind of early years within this industry will look like if suddenly that can all be fully outsourced to AI?
19:22Look, I think there's a lot of, there's a lot of talk about junior banker anxiety, about workflow
19:28replacement, and, and, you know, they're not gonna, they're not gonna have to spend hundreds of hours
19:32in Excel and PowerPoint anymore. I don't think very many junior bankers complain that they, they might not
19:38have the opportunity to move around logos on a PowerPoint slide. I think there's a huge amount
19:43that they can do that's going to be thoughtful that these models won't replace. And at the end
19:47of the day, finance, like a lot of knowledge work industries, is an apprenticeship model. You need to
19:52learn from a mentor, you need to learn from someone, you know, who's done the reps in front of you. And
19:57the problem with banking right now is there's so much grunt work that there's so many junior bankers
20:02that for every one MD that could be a mentor, maybe a hundred junior bankers, right? Like that,
20:07that model doesn't work right now for learning as it. And I think what you'll see is, as folks can
20:12focus on the more human work and actually learn, you'll get a lot more actual learning of what's
20:17important as opposed to learning of, you know, all of the macros and all of the shortcuts inside of
20:22Excel, which is not really what it takes to be a great investment banker.
20:26So what's your plan for the next year? And what challenges are you focused on right now?
20:33I think within the next 12 to 24 months, we will see these agents and Rogo specifically able to
20:41create almost all of the work products that junior investors, junior bankers have to do and sort of
20:46spend inordinate amount of time doing. And so for us, that means we're continuing to push the boundary
20:52of the types of work AI models are able to do in finance, connecting to more data sources that our
20:57clients want, creating more types of outputs that our clients want, and continuing to embed deeper
21:01and deeper into the institutions that we partner with. Moonshot, 10 years from now, how do you think
21:10banking and finance is going to look different? I think it's a much bigger market and a much bigger
21:16industry. There's, you know, hundreds of thousands, if not millions of businesses in the US and abroad
21:22that can't afford to pay Goldman Sachs a fee to raise money or afford to pay, you know,
21:27Lazard a fee to sell themselves. There's a long tail of businesses that don't get access to these,
21:32you know, sort of rich man's financial services of M&A advisory work and so on. And I think as we are
21:39able to help, you know, augment and replace that work, it's going to open up a huge portion of the
21:43market that can actually be served by these institutions too. And it should actually grow
21:47the pot. And so I think 10 years from now, you'll see private markets that transact more quickly and
21:53are more liquid because people can fundraise more effectively and they can, you know, sell themselves
21:58or buy companies more effectively too. And I think as you see more and more, not just AI bankers,
22:02but AI native banks, you'll see a model of the world that looks a little bit more like high
22:07frequency trading in public markets rather than, you know, 22 year olds and Excel and PowerPoint at 3am
22:13as it is in private markets. Totally fascinating. Gabriel,
22:17thank you so much for your time. It was a pleasure. Thank you very much.
Comments