How to approach AI implementation, and Karmel Capital's AI investing strategy outlined

Note: The securities mentioned in this podcast are not considered a recommendation to buy or sell and once should not presume they will be profitable.

In this episode of The Private Equity Podcast, Alex Rawlings welcomes Scott Neuberger, Co-Founder and Managing Partner of Karmel Capital, a private equity firm investing in late-stage software and AI companies. Scott shares deep insights into how Karmel Capital leverages AI within its investment process, how they identify and evaluate late-stage tech businesses, and why they’re placing strategic bets in the infrastructure layer of AI.

Scott explains the firm's capital efficiency-focused strategy, how they rank companies, and what metrics truly distinguish iconic businesses from the rest. He also discusses how AI is transforming internal operations and why firms must go beyond the hype to truly implement impactful AI solutions.

Later in the conversation, Scott offers practical advice to portfolio company leaders on how to begin leveraging AI meaningfully—starting with labor-intensive areas like customer support. He finishes by outlining Karmel’s top-down investment approach to sectors like cybersecurity and why infrastructure plays offer value and growth.

Whether you're investing in tech, operating a portfolio company, or just curious about how AI intersects with private equity, this episode is packed with real-world insight.

⌛ Episode Highlights & Time Stamps:

00:03 – Introduction to Scott Neuberger and Karmel Capital
 01:00 – Scott’s journey: entrepreneur turned investor
 02:19 – The mistake of investing too early in venture capital
 03:47 – Why Karmel focuses on measurable, repeatable metrics
 04:45 – How they assess capital efficiency in tech companies
 06:41 – Key metrics and importance of experienced management teams
 08:38 – Evaluating human capital and talent within portfolio companies
 10:05 – Zooming out: The “mosaic theory” of identifying strong investments
 10:33 – How Karmel Capital uses AI internally for data collection & analysis
 13:22 – AI investing: why infrastructure is Karmel’s focus
 15:49 – Pick-and-shovel strategy: betting on infrastructure vs. applications
 17:44 – Advice for portfolio execs on where to begin with AI
 18:43 – Customer support as a high-impact AI use case
 21:09 – Navigating noise in AI investing: how Karmel decides where to play
 22:34 – Case study: AI in cybersecurity and the top-down analysis approach
 24:59 – The arms race in cybersecurity: AI on both offense and defense
 25:29 – Scott’s reading and listening habits (inc. 20VC podcast)
 26:56 – How to contact Scott

Connect with Scott Neuberger:
📧 Email: scott@karmelcap.com

🌐 Website: www.karmelcapital.com

Raw Selection partners with Private Equity firms and their portfolio companies to secure exceptional executive talent. We focus on de-risking executive recruitment through meticulous search and selection processes, ensuring top-tier performance and long-term success.

🔗 Connect with Alex Rawlings on LinkedIn: https://www.linkedin.com/in/alexrawlings/

🌐 Visit Raw Selection: www.raw-selection.com

Looking to grow your team? Check out our Hiring Guides

for proven strategies, templates, and best practices to make smarter hires. 

Transcript

00:03
Welcome back to the Royal Selection Private Equity Podcast. Joining us today is Scott Neuberger, co-founder and managing partner of Carmel Capital. These guys are investing in late stage software and AI companies. So we're going to dive into how they're using AI for their actual firm, but also their kind of investment thesis, how they're selecting businesses and how they're reviewing them. So really good, interesting podcast.

00:28
Get your pen ready to make some notes.  Scott, if you can share with us a brief insight into you,  Great. Thanks Alex for having me. uh So I'm a lifelong entrepreneur.  I started my first company in college uh and grew it out of college um and ultimately sold it. um And then uh it moved to San Diego,  where I live now with my wife and four kids. along the way of

00:56
of running businesses and starting businesses,  I  became fascinated with investing in businesses. um And around 2013, I  became a full-time investor when I co-founded Carmel Capital with my partner,  James Braylene. uh And from the very beginning,  our mission,  our shared thesis  as investors has been to find a way to invest in value.

01:26
opportunities inside of technology. uh so the last 12, 13 years of our everyday grind has been how to do that. uh you know, where we've been, just, you know, I always tell people coming to work doesn't feel like work to me.  So I  would do it even if I wasn't paid. so I love to tell people about it. But a lifelong entrepreneur turned investor and I love doing it. That's neat.

01:56
Sounds good. What's the saying something about, I love your work and you never work a day in your life. I've probably absolutely quoted that or misquoted it, but we all get the premise. I'm exactly the same in my line of work as well. What's one mistake that you see out of private equity firms, portfolio companies, or of course venture capital firms making and what would you suggest to correct them?

02:19
Yeah. you know, we, we, we mostly focus in, in, in, I would call it the outskirts of venture capital. We're investing in venture capital backed companies. We're not technically venture capitalists. We're technically a private equity fund because we invest in the secondary market predominantly, which legally technically falls as a sec as a private equity fund. Anyways, we, what I think the mistake that

02:49
that some venture capitalists make  is investing too early in companies.  Extreme early stage investing to us  is  like  you have to have a massive shotgun type approach. You have to invest in hundreds of companies  to make a portfolio, produce results that are compelling  because you're really just shooting in the dark quite literally.

03:17
most of the time, at least in my experience. uh So  what we do and I think  is  we learn and we're absorbing data and  exposing ourselves to as many companies as possible, literally hundreds at a time, but we're waiting to see them hit certain key performance metrics before we start to uh seriously look at investing in them. And I think that's what the, you know, the mistake of

03:47
early stage venture capital is you're  trying to divine something uh from  something that is just  magic versus actual repeatable uh metrics  that are business driven metrics.  So we focus on  finding repeatable metrics versus uh trying to figure out that this guy

04:16
who was left-handed and went to this school is definitely going to be a great founder. Makes sense. So just diving into that then, talk to us about your, your kind of identification, your evaluation of kind of late stage technology companies and your type of process around that. Yeah. So our process is, we've, we've found over the years that one of the key differentiators between

04:45
truly great companies, iconic companies that we all know of today that are publicly traded and are leading  their particular industries within technology and just very good businesses, very good businesses that ultimately usually get acquired by other private equity firms,  certainly do quite well for their investors. But what is the difference between those two  truly good oh companies? And what we would say is,

05:15
The data for us supports  both companies are very good at raising capital. They're very good at attracting  high quality investors to their business during the growth of their business.  The  iconic companies differentiate themselves because they're able to deploy that capital extremely efficiently into converting it into  high margin, high growth revenue. um

05:44
and building a team around them that can continue to build that machine,  that can continue to grow uh year on year at elevated rates  with high  margins.  And so we  seek to measure capital efficiency, which is what we call it, in a number of different metrics,  different ways of thinking about it.  And we essentially stock rank every company out there based on these metrics.

06:14
And we zero in on the ones that have the best capital efficiency. Everybody within our universe of over 500 companies have raised tremendous amounts of money from top investors in the world. But we're focusing on the cohort that is very capital efficient. so when you look into those capital efficiencies, what in particular from a metrics perspective do you look at? And if you look at...

06:41
work from a metrics, what additionally are you looking at from a, because you've got obviously more data than a VC investor, early stage entry seed, whatever.  How much weight do you put on that team? As you mentioned that earlier as well. Yeah. So,  um, team is very important for sure. Um, we, we look for it is  in, our,  you know,  it is the, they're the conductors of this orchestra.  Um,  and you know,

07:11
No orchestra, no matter how good your various instrument players are, will work well if there's not a great inductor.  So  we look for management teams that have uh deep experience in the space.  We  hope to find ones that also have had  prior exits or prior IPOs. um

07:39
at the point in time that we're investing in them.  That to us is sort of the gold standard that they've kind of been through a complete cycle before and are back at it again. And in the data we've analyzed, some of the best outcomes are  from uh second time entrepreneurs,  especially second time entrepreneurs that have had big exits earlier in their careers.

08:09
And beyond  capital or how we look at capital efficiency,  so we'll look at essentially different ways of thinking about the amount of equity capital they're raising  and the amount of revenue that that is producing.  And so we'll look at different sort of time periods  closer and further away from their last rounds,  different subsections on the revenue.

08:38
uh And so that gives you a very good picture of kind of, um and we expanded beyond that. We expand into like the actual human capital of the businesses. So the management team is super important, but the overall human capital is also super important. One of our learnings and findings is that um companies are really a collection of people. um And if those people  are

09:06
That's probably speaking to your world pretty well. We all can argue.  It's a two-way street. People have to want to work for you and you have to want to employ them as the leader of the business. It creates a very perfect, in our view, way  of comparing companies together uh next to each other. How long are people staying at the company?

09:36
Are you adding more and more people to the company?  Where are you adding those people? Geographically,  role-based, et cetera. So anyways,  we kind of dive into all that and  really geek out on the numbers and  zoom out and look at it across hundreds of companies at once. And that is when it's sort of this mosaic theory.  Zooming in on any mosaic, it just looks like a rock.

10:05
But when you zoom out, it's like a picture and  it's fairly clear. So I'm going to dive into your investment strategy and what you're seeing from an AI perspective. But I first want to look at this as we're talking about your identification, evaluation of companies. How are you as an investing firm, which is a bit of a bee in my bonnet with regards to private equity firms, venture capital firms, all talking about AI investments and how they're going to do within their portfolio and how they're going to put...

10:33
technology in, then when you ask them what they're doing about it, they say, well, Excel spreadsheet is pretty good. So what are you guys doing from an AI perspective to get you ahead from what you're seeing?  Well, let's just look at  what you were doing as a firm for  AI  in your actual. Yeah. Great question. And uh we're, we're deeply using it in all of our processes.  You know, as I'm

10:59
sort of pulled the covers back a little bit on a few of the different metrics we use. There's actually like 80 different metrics that we are constantly calculating and pulling data in on all these companies. And so you can do it in an Excel spreadsheet, but you know, there's all kinds of human error possibilities. There's all kinds of, there's just only so much time you can do. We used to literally hand collect this information and update our spreadsheets. And we would spend, you know,

11:28
more than a whole person was spending just most of their time updating this data.  you know, there's some, there's some,  the bespoke nature of the data, the fact that there's a lot of judgment involved in sort of  collecting the data.  It helps to have a human uh component to it. What we found with AI is that, is that not only is it able to do things extremely quickly, but there is, there's, there is a judgment capacity to it. There is a...

11:57
There is an analysis that it's doing. um we've been employing that and being able to collect data faster than ever before.  We think more accurately. what it  puts our team at is  now, not  spending much more of their time analyzing that data.  AI also analyzes a lot for us. uh We ask it to come up with new ways of sort of thinking about these  different data inputs and how to...

12:26
how to plot them or how to analyze them. comes up with some cool things. think about it, we look at it. Some of them we don't use, some of them we do use. So it not only is sort of like a brainstormer for us, but it also does a lot of the nitty gritty work so that we can spend time really at the analysis level. we found that to greatly benefit our investment decisions. Makes sense. Makes a lot of sense. Sorry, TuneTwits.

12:54
Just a quick mention of our longstanding partnership with Grata. As you all probably know, the private equity scene is constantly evolving and deal flow is moving now to proprietary and data-driven processes. Grata provides you with the data and information of over 7 million private companies. So if you're looking to improve your proprietary deal flow and improve the data access, then reach out to Grata today. Now back to the podcast.

13:22
So you guys invest in AI, which makes things interesting because AI is a hot topic at the moment. What are you guys seeing as in those kind of later stages, which is nice to hear or, you know, with regards to investing of these types of AI businesses, which gives us an insight into a bit of a crystal ball to love, but what the future looks like within artificial intelligence play. Well, if I had a crystal ball, wouldn't be here right now, but, but I'll do my best to show. mean, uh,

13:52
You know, look, AI, like a lot of technology  is a multi-layer sandwich. That's the way we think about it, the way a lot of people think about it. You know, at the bare bottom is the semiconductors, the NVIDIAs and  other  semiconductors that are running it or they're creating AI to exist. The very top of the sandwich is these applications.  know, somewhere in the middle  is  the models themselves.

14:22
The open AI is the,  you know,  you know, anthropic, mistral, when, go here, et cetera, et cetera,  And, but also right around that middle or just above the layer of that middle oh model layer is,  is the infrastructure providers. um These are often called these, they're hyperscalers. They're also what's being called Neo cloud these days.

14:51
companies like CoreWeave and Crusoe and Fluidstack and um NScale  and uh Lambda, all companies that we've invested in by the way. uh The um infrastructure layer is super interesting to us because similar to the model layer,  there uh is an essential component to everything that happens above it. uh

15:21
A lot of the investment dollars right now are going into the layers above it, the application layers that are taking  the model, taking infrastructure and doing something to it to produce a product or service  sold to a business or a customer or a consumer. uh in that, in that  very top layers, we find it to be pretty hazy right now. Who's really going to win that in the long run?

15:49
which areas are the models just going to do themselves versus are they going to leave for  the applications specific companies to  run themselves?  And  so what we've really focused on is the infrastructure layer. We feel like the infrastructure layer  is showing the same kind of explosive growth that the models are doing largely because they are really serving the models. It's kind of  what's called like the pick and shovel strategy, you know, in a gold rush.

16:19
sell the shovel, sell the bucket,  don't try to mine the gold yourself.  And so you see much lower valuations uh with very similar growth profiles. So it really fits our value uh mindset towards investing in technology.  from our perspective, uh with no perfect crystal ball, but from our perspective,

16:47
No matter who ends up winning the models, the application layer, et cetera. There's going to be a lot of infrastructure built out.  so, so, so we've, we've been investing heavily there. So the talk and you guys are obviously using AI, both in your firm and obviously investing in those businesses. Now everybody listening here kind of knows or should know that AI is here to stay. And the, need to be leveraging this.

17:16
What is more difficult and be interesting to get your perspective is, you know, I've, we've, found somebody that's supporting us and, uh, and helping us as kind of a consultant to deploy the right things and building some of these things through an, and I think N8, NNN or something like that, and using all these 10 different software programs, all just to get this one thing to happen, whether it be our resume formatting, our lead generation for scrapers and all these different elements, right? But for, for you guys,

17:44
from your advice and from your perspective, if we just look at this at portfolio level, and you've obviously got businesses in software and AI, so there's  tech enabled  already businesses clearly, what would you be saying to portfolio chief executives, CFOs, COOs listening when they're going, well, I don't even know where to start, but I have downloaded the Chatgy BT app, I'm now AI enabled, but equally, I'm just talking with a robot at this point, how do I actually make material difference? What advice would you be giving them  that are listening?

18:15
Yeah. So what I think about is, and what I would advise, you know, anybody that was, you know, thinking about their own business and thinking about how to implement AI and, and, and get benefit from it very quickly. Um, I would look at the,  uh, the  most,  um,  uh, the most labor intensive,  um, aspects of your product or service.  Um,  and

18:43
For example, for many companies, customer support is a very labor intensive, process driven cost of a business. AI is amazing at repetitive, process driven,

19:13
work. and so not surprisingly, customer support is an area that, that AI is doing fantastic in. have a few investments in that space. A company called Intercom is our highlight investment in that area. A company we think very highly of. And, and, and, and, and most businesses have customer support at some level or some version of customer support.

19:43
And, uh, and, and, it's the, you know, our view on it is that, that, that first and second sort of tier of service can so efficiently be provided by, by, uh, AI, um, you can get almost instantaneous answers,  um, both on in text and email and voice.  Um,  uh,  and,  um, and, and, and it is just as good, if not often better than a lot of.

20:11
customer support because it's basically trained on your best support agents within your organization. And then those agents can then be focused on the overall customer experience.  And what you'll find is one, I think  your cost structure comes down and your service level goes up m and your culture is better because instead of the  constant sort of

20:39
rep repetition of customer support.  Um, you know, your, your team gets to work on more interesting, less, you know, wrote exact examples.  Um,  and, um,  and, so it's, it's a win-win across the board. So you mentioned, I'm thinking about the kind of first question on, you know, the mistake that you see VC making and that being a, basically a bit of a minefield as to how you invest. Now you guys have see a little bit more traction.

21:09
than those guys, but I can only imagine there's still a lot of noise at your level and there is across any investing level. What's your kind of, you mentioned your due diligence phase, but if you just take one step back before that, you mentioned the focus around the infrastructure of the deployment of AI, but how do you kind of come to the conclusion of where to play? Because AI is brand new. It's hard enough when it's hey, manufacturing and that's a big enough market in the world and you're like, well, there's

21:39
hundreds of thousands of these companies, millions possibly, these businesses where we're going to invest in. Now we're going to tie it down to North America.  What's your kind of process of how you understand where you're to play, what's going to be the area and how did you come to that conclusion?  great question. So, I mean, we do, we start with like a top-down analysis of any of the overall ecosystem and then each individual portion of  it.  So, um,

22:09
Um, and, and, and, know, so like I mentioned kind of that, that layer diagram of,  of, you AI, know, from semiconductors to applications with models and infrastructure in the middle and a bunch of different steps in there. So we'll build very similar versions of that for each, each particular area. And we try to understand what are the key components of a particular, uh, sector.

22:34
Like for example, cybersecurity is also an area of focus for us and is a perfect example of this. There's lots of, know, cyber threats come in many different forms  and there is no one, uh you know, cybersecurity software package today that can prevent all of it.  And so if you're going to invest in cybersecurity, you have to understand what are the different threats, where do those come from and what are the most effective

23:04
uh,  mediation method, remediation methods or protection methods for it. And then once you understand that none of that is really at the company level yet. That's all of the, at the technology level, the understanding,  you know, why something is happening and, and, and, and, and how to counteract it.  Um, then you, start to zoom in and you say, okay, this is the particular area, you know, whether it's, you know,

23:30
Email security, for example, is an area that we've invested in because, know, perhaps unsurprisingly, email is one of the easiest open door policy or it's one of the easiest doors for a bad actor to break in. We get thousands of emails all the time. You know, it's a game of, you know, you just have to win once out of

24:00
out of many thousands or hundreds of thousands to make a big impact. So it's, it's a really easy place. So  who, how are, how are you, how is, how are people using our leading companies using AI, for example, to protect uh email.  And because again, AI can look and scan things much faster than any human can than they even human program software again.  So how, so it's a great use case of that.  Also.

24:30
You know, cyber security is one of these areas where,  um, where, know, it's, it's so interesting to invest in because it's one of the few areas where there's a, there's an adversary, you know, you have competitors in most businesses, but, but, um, but you don't always have an adversary in the case of cybersecurity of an adversary. And so they're also using AI. This is a huge industry unto itself. And so they have resources, they're using AI to improve their side of the equation. And so.

24:59
You have to use AI on the protection side of the equation. anyways, to come back to your specific question, I'd say we really start top down. We try to learn the industry that we're focused in. We try to learn why particular areas in the industry are growing faster than others and assess which ones are we think are going to be the most durable. And then, and only then do we then start zooming in on the companies that are benefiting from that trend or theory that we've developed.

25:29
So, um, I'm interested. What do you read, watch, listen to? Do you recommend that others should check out Scott? Yeah. So,  um, I read a lot. I listened to a lot. love podcasts personally. Um, so, you know, um, so, you know,  um, I, I, I try to absorb,  um, as much as I can,  different perspectives.  Um,  um, so,  um,

25:57
I'm always, you know, reading about companies that were either investing, invested in, or thinking about investing in,  trying to get as wide an exposure as possible to that. So,  so like, for example, a of the, early stage venture podcasts, 20 VC is one of the ones I like and to listen to a lot.  But a lot of the early stage venture podcasts, we're not exactly investing at that point in time often, but

26:26
It is a great way to get smart about what is coming in the future. It feels like looking into a time machine a little bit. I pay attention to all of that. Reading all the normal sources that you would think about. Perfect. If anybody wishes to reach out to you post this podcast, how best to get in touch, Yeah. So direct over email, Scott at

26:56
CarmelCap.com  and would love to chat with you. Well, thank you very much for coming on to the Private Equity Podcast and sharing everything you have done. Thank you for having me. Thank you again to all of our listeners. Till the next time, keep smashing it, and thank you very much for listening.