Robert F. Smith on AI, Innovation and Enterprise Software | Vista Equity Partners

Robert F. Smith
4 followers
Nov 30, 2025 #Robertfsmith #Business
This episode of the Altco's Mainstream Podcast is brought to you by Ultimis,
a leading full-service fund administrator for asset managers in private and public markets.
As private markets continue to move into the mainstream, the industry requires infrastructure
solutions that help funds and investors keep pace. Ultimis is a leading full-service fund
administrator for asset managers in both private and public markets, offering a wide range of
capabilities across registered funds, private funds, and public plans, as well as outsourced
middle-office services. Delivering operational excellence, Ultimis helps firms manage the
ever-changing regulatory environment while meeting the needs of their institutional and retail
investors. Ultimis provides comprehensive operational support and fund governance
services to help managers successfully launch retail alternative products. Trusted by
institutions, investment consultants, registered investment advisors, state governments,
and fund managers, Ultimis provides solutions for nearly every investment structure in the
marketplace. Visit www.ultimisfundsolutions.com to learn more about Ultimis's technology-enhanced
services and solutions, or contact Ultimis Executive Vice President of Business Development,
Gary Harris, on email at gharris at ultimisfundsolutions.com.
We thank Ultimis for their support of Alt's Going Mainstream.
Welcome back to the
Altgoes Mainstream Podcast. Today's episode welcomes a pioneer and visionary in enterprise
software investing to Altgoes Mainstream. Robert Smith is the founder, chairman, and CEO of 100
Billion AUM Vista Equity Partners. He sits on the firm's investment committees for Vista's
flagship foundation, Endeavor, and perennial funds, and serves as a member of Vista's Executive
Committee. Vista's portfolio spans 90 enterprise software, data, and technology-enabled companies
that employ over 100,000 people worldwide. He's also heavily involved and committed to the firm's
wealth channel efforts as the chairman and investment committee member for VistaOne,
the firm's evergreen private equity vehicle. Since Vista's founding, Robert has supervised
on over 600 completed transactions that represent more than $330 billion in aggregate transaction
value. Robert founded Vista after a career at Goldman Sachs in tech investment banking,
where he was co-head of enterprise systems and storage, executing and advising on over $50
billion in M&A activity with companies that were foundational players in the early days of the
internet and technology, such as Apple, Microsoft, Texas Instruments, eBay, and Yahoo. Robert has an
innate understanding of technology and trends, shaping the way that companies and people
interact with the world and conduct business. It's no surprise that he was early in seeing
the rise in impact of AI because he was early in seeing the dawn of the internet in the 1990s.
Robert and I had a fascinating, thought-provoking conversation about the evolution of both
enterprise software and Vista as a firm. We covered the early days of enterprise software
and what Robert saw early on that gave him conviction to focus on enterprise software
as a banker and then as an investor building Vista, the investment characteristics of enterprise
software, the power of product superiority in enterprise software, why sovereignty and
dominion of data are so important, and why it matters for AI, whether the rule of 40 can become
the rule of 50, 60, or 70 with AI, what aspect of AI is most impactful for companies, how Vista
approaches value creation, what it took to scale Vista to a $100 billion investment platform,
and why the Wealth Channel is so core to Vista's business and to the firm's DNA. Thanks, Robert,
for coming on the show to share your expertise, wisdom, and passion for enterprise software
and building businesses. We're going mainstream.
Robert, welcome to the Alcove Mainstream Podcast. Thanks, Michael. Good to see you.
Likewise. Pleasure to have you. I think we have such a fascinating conversation on tap
between your career, evolution of the enterprise software landscape, what you've seen from the
early days of your career, now AI becoming a big part of that, how you built Vista. So,
we'd love to start with your background. You started in engineering. I find that fascinating
because it feels like a number of investors in enterprise software have started with an
engineering background. So, I want to unpack what's instructive about that when it comes to
thinking about the enterprise software landscape. Sure. That's a great question.
Congratulations on building a wonderful platform and informative. It sounds like you're delivering
unique content to your subscribers, which is great. Frankly, stick with it. Keep doing it
because more investors are going to need what I think you can extract and deliver to them.
My background, again, chemical engineer. I went to Cornell. I'll call it a generation of chemies.
We're really focused on a couple of things. One, I call it, we're in a whole lot of new
unit operations or the creation of new processes, but we were in a focus on automating processes.
We had just discovered that we could utilize this compute technology for controlling systems.
One of the early projects I worked on was implementing a programmable logic controller
into a, at that time, a 40-year-old plant. These plants were typically, like all these
processes were more manually or analog controlled where you had
operators who would go through an episodic observation of an event and then an episodic
intervention. To the extent they were paying attention and not talking about who's getting
drafted in football, you would have certain control paradigms that would occur in these processes.
Well, once you introduced a compute system into those environments, you're observing an environment
60 times a second and you're intervening whatever is applicable for the process. You have a much
tighter control paradigm. What that led to, at least in the case, the first deal I worked on
way back when, our company implemented one of these is a 26% increase in the productivity of
a plant that was 40 years old. If you start thinking about that in a broader scale, compute
actually brings massive productivity and massive efficiency into almost any environment that it's
placed into. Can you take it from the process industries and then move it into different
environments like an office? You have this same sort of productivity. When I look at today,
at Vista, we're, I don't know, fourth, fifth largest enterprise software company in the world.
We've got to add all our revenues up, just a little under 33 billion in revenues.
What engineering and chemical engineering in particular informed me about doing and managing
a business was how do you think about creating a systemic process to improvement as opposed to an
episodic intervention to a problem? Those are really the frameworks that I like to think about.
I call them engineered solutions. How do you think about an engineered solution that when you fix it
once, it's fixed forever? Of course, there's tuning that's associated with it. Investing
in enterprise software, the thing about it is it is indeed that. I like to say chemical engineers
were designed to create elegant solutions to complex problems. If you take that paradigm
of thought and then infuse it into investing, that's really what Vista is. It's an elegant
solution to a complex problem. Such a fascinating explanation of enterprise software. I think we'll
get to a lot of what you talked about in the context of Vista and how you think about both
investing, but also post-deal value creation around how you think about implementing processes to help
these companies. First, I want to get to you talking about your background. You actually
started at consumer-facing companies, Goodyear is one of them, before going into banking at Goldman
and really were at the early days of the internet. What did you see in the 90s? It's easy to see
today that companies and people want to buy more software as a way to augment processes,
but in the early 90s, what did you see that made you say enterprise software is the place to be?
Going to Goldman Sachs, I worked in, in essence, I call it applied research and development,
broadly speaking, for about six years before I went back to business school. That gave me unique
insights as to how things are made and how things are marketed and how things are sold and all those
sort of things, which was fascinating and fun. I encourage all young people to do that. How do
you build things as opposed to observe things and talk about them? From that, and then joining
Goldman and having a chance to really work in an M&A department, which then became the tech group,
right? You actually had this, you got a chance to see how capital is deployed across the planet,
how you think people think about utility of capital, different capital stacks of capital
structures, and, you know, where, where's the highest and best use of different types of
capital, depending upon where it's formed and how it's delivered. The thing that became very clear
to me was there was a massive arbitrage of productivity that was now starting to infuse
into businesses. Again, my early career, actually writing code, programming these, you know,
programmable logic controllers and seeing how effectively you can actually make a change by
having tighter controls. Think about that paradigm, tighter controls on, on operations.
And then as the world started to disaggregate, I'll call it hardware from software. Back in the
day, you used to sell it as a whole solution and you had kind of hardware and middleware and then
software and all that kind of evolved ultimately where the application layer of software actually
became frankly, the most versatile in terms of value creation. You didn't have the hardware
or the refresh cycles. And so you could actually, you know, increase the frequency of innovation
as a result of the construct of how software actually works. And of course we've got,
you know, higher, higher level languages, which gave it more, more, you know, flexibility and
more capability associated with it. But then as you actually started to look at software companies,
which I had the great occasion to, when I was at Goldman, I got to advise some pretty interesting
companies, little company called Microsoft, one called eBay, one called Yahoo. And in evaluating
those companies, what you saw is everyone approached their market differently. And not
just in terms of the go-to-market approach, but actually how they did things like their product
development, how their customer service worked, how their, you know, support worked, how their,
you know, development, all, all the aspects, go-to-market, back office, infrastructure,
all those sorts of things were all so different. But yet you would see, you know, best practices
in one company that if you applied that best practice on one company to another company,
you'd say, wow, man, there's a 5%, 8%, 10% improvement. So if you think about it as the
paradigm of how do I take best practices out of one place and apply them to the next, not rocket
science, but, you know, just something that could create that massive arbitrage. This company is
operating at 12% EBITDA margin. If you apply these five things can operate at a 30% EBITDA margin
by, you know, being effective and not having to change the product at all. And ultimately
focusing on call it the execution capacity. The next layer of that is, of course, how do you create
the way I think about it is it is some product superiority, which means really evaluating what
it is that the product does for the customers, what it could do better for the customers. And
you have to then listen, listen to what customers have to say. What is it that your product does for
me? What is it that if your product did something else would be even more valuable to me that I'd be
willing to pay for? And so that is the next part of the equation. And so the way I like to think
about enterprise software kind of pre AI, and we'll, we'll get to that in a second. Well, pre-agent tech,
not AI, because AI actually enables that. It is all about product superiority and execution excellence.
Okay. So how do you build systems at scale to deliver both of those things? Okay. So part of it
is, you know, figure out what systems actually work for different components of enterprise software,
product development, customer support, go to market, back office, administrate, all that
sort of stuff, build out sets of systems, and then take those sets of systems. And as you buy companies,
or as you evaluate companies, if you apply that system into that company, you're looking to
underwrite, what does that company now look like? And that tells you what you can pay for the
business. And then you look at what everyone else can do and say, well, this is what they should pay
for the business. And there's some arbitrage there in the process. So that's the whole dynamic.
I think that brings up a really interesting question, which is in part because of the process
that you have at Vista in terms of what you bring in terms of operating playbook,
the understanding of enterprise software businesses that you've invested in.
When you underwrite a business, how much do you look at, here's the business today on its own,
versus here's what we can do if we work with the business and take those operating processes
and then can help scale the business? Do you wait one more than the other?
Like all things, it isn't as linear, but the framework is, what are the critical factors
for success as an investment that we have under our control? And that's how you have to think
about it. Okay, what is it that we know how to change, or we know how to encourage, support,
underpin that may also already be in place that in doing so now decreases the risk in that business
and enhances basically the convexity of the upside? And that's how to think about it. So
it's a gradient and it's a function. In some cases, you may say, hey, look, this particular
management team or this particular individual in this particular area is doing a great job,
better at this than anyone else. So how do we underpin that person, that team to be very
successful on that part of the business? And part of that is maybe incentive comp. It may be the
way that you actually approach the marketplace and maybe how you distribute knowledge so that
you have less risk in the business. You then will look at other things and say, well, okay,
they actually don't have a go-to-market strategy that includes this channel. Well, why? And what
can you do there? Or they may not have their pricing right, or they may not have their
contract administration right, or they may not have their ability to scale infrastructure,
which might be people getting enough developers to do the work at the pace that you want,
which we'll get to it on the Gen AI thing, I'm sure. All of those are elements that you underwrite
to. You can say, okay, what do we know how to do? What do they know how to do? What do we think is
required for this to be a successful outcome? Given a market today, you can project in some
cases where you think the market or what the market for that particular asset might be in the
future, that if you get from here to there, this will give you those sort of returns and how do
you de-risk that process? So that's really the methodology that we undertake.
When you talk about de-risking process and you talk about something like product superiority,
when it comes to enterprise software, where my mind goes, or at least one area where my mind
goes is these contracts tend to be pretty sticky when it becomes something that becomes hard to
replace because it's a, in some cases, mission-critical software. What do you look for
when it comes to product superiority? Are there things that you're looking for to discern says,
this is product superiority that a company has? Yeah, there are a few things around the product
superiority, but as a general category, you have to look at it relative to its competitive set.
That's an important part. So it isn't, oh man, this is a great product, but relative to what
everyone else has relative to what the customers want. So you really have to do some deep dives
into why do people buy this software in the first place? What is it that makes this unique?
What is it that makes it differentiated? And what is it that makes it protectable? Is there
anything that's really protectable? The world of Gen AI changes some of that for a lot of things.
If you don't have, I call it sovereignty and dominion over workflows and data sets,
you don't have a right to exist as an enterprise software company. If you do,
then you have a right to dominate. If you enable and identify that business more effectively.
And so what was a very important part of how you underwrite is now an essential part of how
you underwrite because you also have to understand the impact of agentic solutions
in the world of enterprise software. If you are just repurposing content, then you don't really
have a right to exist because these agentic solutions ultimately will, I mean, they're the
best search engines on the planet and they can aggregate, refine, define, and deliver and dispense
content much more efficiently than many other things that are out there.
But if you actually have the ability to say, listen, I have a unique workflow,
I have a unique data set that only 1% of the enterprise data actually lives in these
public environments for LLMs, broad scale LLMs to train on. And if you have the ability to protect
that and utilize it effectively, then you have a unique advantage that no one will have
for a long period of time. And so there's a different underwriting criteria you have to
think about from a product superiority now versus what it was in the past. And in the past,
there's a lot of feature functions and do you have the capability, do you have data sets?
Remember, by having unique data sets and unique workflows, you can actually go to your customers
and understand, I call it the derivative products that only you can deliver to them because you're
the only ones with the workflow, the only one with that data, the only ones who can use that
to create a new product that creates a unique set of advantages that you can actually price relative
to their utility. That is a concept that is very different enterprise software than almost any
other industry. And I like to say it's the only business I know that every day you can develop
a product that only you can sell to your customer. That's a dynamic that will be enhanced
in an agentic world. On that point, is it the scaled players that will benefit more from
the application of agentic AI than smaller companies or startups that may have the ability
to use AI to help companies? I got to break your question up a little bit. It's not necessarily
scaled players, but it's incumbents who actually have sovereignty and dominion over workflows and
data sets. There may be scaled players and there are some scaled players out there today that don't
have sovereignty and dominion over the workflows and data sets. They actually don't have a right
to exist. There are smaller companies that may not have very rich data sets, but have some utility
in using their capabilities and actually some other data sets that actually can propel them
to be quite successful. So you got to discern some of this stuff. It's not as easy as big will
survive, but there are some advantages to incumbents if you have actually been managing
your business well up until now. And at managing the business, and I can't underscore this enough,
is really that sovereignty and dominion over workflows and data sets. You don't have to own
them. You just have to have sovereignty and dominion over them. Sovereignty so that you can
protect them and dominion so that you can actually utilize data sets that are unique to your business
and often are created from the workflows that are unique to you as well. That will create a,
depends on the industry and what the value is, a massive economic advantage in an agentic world.
And so there's an underwriting view that you have to have on how that works that actually is going
to change the way that enterprise software is being fed by AI. And the way I like to say it,
oh, AI is going to eat software. No. AI is going to feed software that's going to eat services.
Because what agentic systems actually are, they're agents. If you think about it, they're workers.
They work 24-7, 365, so long as you call it, build, tune, and enable them the right way.
If they're fed with the right kind of data sets and the proper workflows, you can actually create
a massive workforce of these agents who will do work that you otherwise couldn't do or otherwise
couldn't do at the speed or the pace or the volume with human workers. So there's a dynamic there,
but not the largest of the software companies out there don't necessarily possess the knowledge of
the workflow of the data sets. And there's some small ones who might have some incumbent advantages
in that regard. So you've got to look at them individually. How do companies make sure that
they can get that either competitive or comparative advantage in terms of having the
data sets to be able to then use agentic AI to have the workers to make sure they own that data
and it becomes valuable to the customers that they're working with, either through software
and or helping them with their services. Their best answer is to call us. It's interesting.
I'm actually shocked, to be frank with you, in certain companies that we'll meet with and CEOs
will meet with, that they actually haven't done a good job preserving that sovereignty and
dominion. I'm shocked. On the one hand, I'm shocked. I guess on the other hand, I'm not as shocked
because many of them were on a ARR race. What I mean by that is they're chasing growth that they,
in a way that they've actually leaked intellectual property, they've leaked intellectual capacity,
they've leaked ownership in some respect of the very thing that frankly would enable them
to be more successful in an agentic world. It wasn't necessarily something they were focused on
and a lot of people weren't. Because of how we focused on product superiority and product
development, we had a natural motion that enabled us to have more sovereignty and dominion over
workflows and data sets than we otherwise would. I'd like to stand here today and say,
oh, that was all conscious and deliberate, but sometimes you just get lucky, right?
But I think that luck and sincerity came from our desire to utilize data and workflows that
enabled our customers to use that to create new products. And so as a result of how we
approached that contractually when we bought businesses, it enabled us to actually have a
higher utility of the data and the workflows than we otherwise would. I've met with some CEOs of
some very large software companies, they didn't do that. And it's just a matter of time before
a 24-year-old who can command six, eight, 10 agents can eat the lunch of those businesses.
It brings up a really interesting point around making sure the executives are the founders of
a firm. It could be an investment firm too. I think I would actually group this question both
as investment firm, Vista itself, or the companies that you're working with
and investing in, but what type of culture needs to be created by the executive team founder
to be able to really think about and make sure that you as a firm have that sovereignty and
dominion over your data and workflows so that you can do all the things that you want to do.
There's a component of, no, there are many components of culture that make that effective.
One of them actually has to be an openness to understanding what it takes to be successful
in the world of enterprise software. It's not just product. It's not just your go-to-market
engine and your sales, and it's not just better developers. It's things like your
back-office administration. Those are the things that people don't pay a lot of attention to
that ultimately define what is the economic, you call it, opportunity of the business.
I like to describe it this way. Enterprise software companies'
economic position is defined by their contracts, and you're only as good as the quality of the
contracts that you have signed, but if you will talk to 99 software executive CEOs and you tell
them to name five most important things, most of them won't name contracts, the quality of
their contract. Why do you think that is? They just don't think about it. They think it's all
about, let me sell products fast, let me sell products, get more customers, provide support.
Those are all important things, but that is all defined by the contract, and so you actually need
to have exceptional contracting capabilities that you have to constantly refine. I say it this way,
Michael. It's kind of funny. Just take an average salesperson. An average salesperson is focused on
their commission, and so you will say, here's your territory. Go sell this, and they will spend,
oh man, I've got a great client, great customer, and that customer has a really strong contracting
organization. It might be a Fortune 500, and then you give them their standard contract,
and they just eat it up, and your salesperson now spends 95% of his time trying to convince
home office as to why they should take this contract, as opposed to spending 95% trying
to convince them that the contract that we're giving them is better for them in the long run.
Remember, salesperson sells, and they toss over there, and they're, oh, I'm giving my commission
and move on down the road, right? You have to have business processes that manage that
intentionally. Otherwise, you will have every salesperson, and every regional sales manager,
and every national sales manager, you'll just have variance that actually ultimately is the
underlying construct of the economic opportunity is compromised by the variance in those contracts.
Those are the dynamics that have to be thought through intentionally and built culturally in
an organization as to why that's important. How malleable do you think that is? What I mean by
that is maybe there's a great company, good product, but doesn't have the discipline around
how they're thinking about their contracts or how they incentivize their salespeople to create the
outcome that you just mentioned. How much of that is malleable, and how much can be changed by
processes that if you were to come in as an investor at Vista that you could change something
like that, or is that such a cultural phenomenon that it's hard to change? That gets back to the
first part of your question, which is what culture makes you successful? Openness, thoughtfulness,
openness, and willingness to change. Often what happens is a product person comes up with something,
they build a product, they sell it, it's great, it solves a problem that nobody else solved,
and it's flying off the shelf. Now they're focused on selling, installing, customer support.
The last thing to think about is what contracts have been signed. Part of when we come in,
it's like, okay, now we'll look at all that and say, ah, okay, here's where we can actually buttress
what I call the real economic expression of that company by changing the way your
contracting process works. You may not see it, but that is one of the most important things
ultimately in a sale process, an IPO, whatever it is, because it now gives you layers of foundational
layers of economic support that now enable you to grow off of a strong base as opposed to dealing
with low retention rates because you had bad contracts and people are doing different things
in the contracting process that ultimately two years from now, oh, that was a great client, but
they're leaving now, or they want to change it because you didn't have a good strong contract
for them. Part of the best practices approach that we take is that let's go in and let's,
you got to go in and segment all of the different contracts, which are the ones that are in great
shape, great, what's the standard that we want? What do we need to change? How do we change it?
What's the process in order to change it? Because it just can't come in with a, hey,
guess what? Everybody gets a new contract. Okay. No, you've got to actually go through in a systemic
way to build capabilities and ultimately to build these layers of support for the business.
One thing that's interesting in what you just said is that-
Only one thing.
That specific thing, at least I picked out enterprise software.
Seems like it could be construed to have a slightly different risk return profile than
other parts of private equity, even yes, it's still an investment in a company,
so it's still equity. But some of these contracts feel a little bit infrastructure,
like they could be long-term contracts, maybe not the same length as an infrastructure investment
contracting with a government entity, but they have a different risk return profile.
And I want to unpack that because I think that's so instructive to enterprise software
as an investing category. And I want to go back to when you started Vista early on,
obviously there's been multiple waves of software that have since changed the market. We'll get to
AI in this context a bit, but to start, what did you see in enterprise software at the time that
started Vista that made you say, this is a market that I want to be investing in and this
is where I want to deploy capital? Yeah. So if you look at a very high level,
if you do this well, enterprise software has mission critical and business criticality.
What's that mean? That means that you look at renewal rates and retention rates and you look
at, oh, this person's been a customer for eight years. They are completely reliant on this
software. Can you build contractual obligations with that class or group of customers that
actually gives you certainty? And remember, these are 95% gross margin products at the end of the
day, right? No CapEx, negative working capital, right? No inventory, build it once, sell as many
times as you can, right? So to a great extent, that sounds like a very stable investment
opportunity, which, oh, by the way, sovereignty, dominion, workflow, or data sets, you can now
create new products and now create a new set of economic rent capture opportunities, right?
Because I'm now selling a new product that only I can sell you because I'm the only one with your
data and understands the utility or your data utilized in your environment in this product
that I'm going to now sell you. So you think about, you can baseload these businesses with
some convexity in the upside. So from a risk return perspective, these are great businesses
to own. And then it's a matter of how do you use newer waves of technology ultimately to capture
more economic benefit? Okay, how do you take this product that was on-prem back in the day,
legacy, to hosted? What's the advantage of that? Well, one of the advantages is when you move to
a hosted environment, your clients no longer have to go through hardware refresh cycles,
right? It's a browser-based interface in essence. And so now all of a sudden you had, you know,
millions of dollars that you were paying in hardware refresh to go away. Well, I should
capture some of that value, right? I'm selling you a product that decreases your cost of support.
Okay. And your CapEx goes down and goes to OpEx. I should get some, my fair share. Okay. Oh,
by the way, in that environment, I can now make changes more quickly to the product and
update the product more quickly and actually give you a feature function benefit that you
otherwise wouldn't have in your environment. So I should capture more of that benefit as well.
So you have this ability to what I call stack economic rent curves for existing customers over
long periods of time, which actually create a more unique advantage to, you know, capturing
economic value as an investor. So that's the point of it. I think it's a great segue to something
you've said more recently, which is rule of 40 in software is kind of the gold standard for how
people look at investing. You add something like agentic AI, maybe that takes it to rule of 50,
60, or even 70. Can you unpack what you mean by that and what that means for those who are
investing in enterprise software? Yeah. Let me break that up into two separate statements. Okay.
Rule of 40 was a gold standard. I will argue we coined the phrase. I'm not going to get many
arguments against that as my guess. And for those of you in the audience, it's really a combination
of EBITDA margin and growth rate. 25% EBITDA margin, 15% growth rate, great company, 40%,
you know, rule of 40. Gen AI enabled, before we get to agentic, is utilizing AI to make that rule
of 40 more efficient. In enterprise software, 30% of our costs, in essence, are product development.
Okay. So you think about the first instances of a, you know, gen AI enabled product. It is a
developer who now can operate three or four agents who can be as effective as five or six developers.
So all of a sudden you go from, well, wait a minute, I was spending 30% of my cost on development,
I can now spend 20% of my cost on development. More effective on new product and code writing
than older. But, you know, again, there's a tipping point dynamic that's going to occur in
our marketplace where ultimately the gen AI enabled or the gen AI generated code will ultimately,
you know, supplant what was the legacy code. And so then you're going to see a massive shift
in the cost dynamics. So now all of a sudden you're increasing your margin without selling
one additional product because your cost of support, cost of development goes down,
0.1 development. Customer support. As you think about utilizing these tools for customer support
calls, you know, 40, 50, 60, 70%, 80%, 90% deflection of these calls with agents, higher
NPS scores, higher net promoter scores, higher customer service that you can do with fewer
people. So now all of a sudden you're reducing your support costs. Go to market. Okay. If I talk
to my CEOs, well, most of them, they're going to tell you that their salespeople spend 100% of their
time selling. If you actually do the work, they're only spending 15% of their time selling. 85% of
the time it's the administration around selling, sell support, you know, capturing information,
putting it in some repository, using that to, you know, to make them more effective sale and
all those sorts of things. Gen AI enabled agents can do a lot of that work. And so all of a sudden
your salespeople now spend 40, 50, 60, 70% more time selling as opposed to the back office
administration part of it. All those dynamics right there, I just lifted your margin from 25%
EBITDA margin to 40% EBITDA margin. Okay. So now you go from rule of 40 to rule of 65.
And now you say, okay, now how do I enhance my go-to-market part of this that I can actually
be more effective in sales lead conversion. So now all of a sudden you're lifting your growth rate.
Okay. Now your rule is 70, right? So that's a dynamic, which is what I call Gen AI enabled
enterprise software. The second part is agentic where you say, wow, I'm now actually able to
develop agentic tools, agents that actually are delivering a completely different set of products,
utility of those products, or actually addressable market going to a different addressable market
that actually has an exponential opportunity. What is that exponential opportunity? At a high level,
it is AI enabled software that now can go and display services because you now have workers,
which are agentic workers who do the things that an individual might do. And especially
you create them in an orchestration layer. Those are the important parts of how you move from
rule of 40 to rule of 70 to agentic. And some agentic may not necessarily be the,
you know, the, or some rule of 70 may not be agentic in terms of their products,
but you can still operate them at almost two times EBITDA margin that you had them operating a year
ago. That's what we're pivoting all our companies towards. Are there industries where services
are going to be impacted by agentic AI where software companies are going to be the ones that
eat into those services or generate revenues from capturing part of that market? Yes. Which,
which industries or categories do you think it'll be? And if I tell you, Michael, then everybody on
this podcast will then go buy those companies. So I'm not going to tell you. We'll have to find out
that. Watch and see. On the flip side of this, where do you think AI still falls short? You've
probably seen, you know, we've announced this deal with Microsoft to, in essence,
build out our capacity to create agentic solutions. We've got our first company through
the factory. We've got another 12 going through now and flights of them. This is a lot harder than
what people say. And if you look, and you've probably read the MIT study, oh, you know,
95% of this stuff is in personal productivity. You're not seeing much in the enterprise because
it's hard. And you actually have to, to create highly precise agent and agent action in the
enterprise. So here's the way I like to characterize it without going straight down that rabbit hole.
Agentic solutions for consumer, we can deal with, I'll call it the fuzziness of outcome.
You know, the hallucinations in essence. Okay. You and I decide, hey, this has been great.
Let's go have dinner tonight. Let's go to a Thai food restaurant in New York. Tell my agent, book,
you know, dinner for two, Thai food restaurant, seven o'clock. Great. 95% of the time you and I
would show up in as a Thai food restaurants and the restaurant in the book at seven o'clock,
all that sort of stuff. But then that 5% will show up. And it actually is a restaurant that
actually had a Thai dish, but not a Thai food restaurant, but a Thai dish for that restaurant
may have not been in business for three years because of where they're actually drawing
the information, excuse me, the information from. Right. That's not adequate for the enterprise.
So you think about every month someone's getting paid and a wire transfer goes into your account.
And every month you, you know, give your daughter or son, you know, thousand bucks.
And so an agent in an enterprise might say, well, why am I doing that? Why don't you just
send a thousand bucks to the son or daughter? Right. And in some cases you may say that's okay,
but for the most part, no. Right. You still do not want to change that workflow. Maybe don't
proceed to be more efficient. So that's a different use case in the enterprise
that still would deal with, call it a reasoned answer with a fuzzy, you know, fuzzy outcome
that may or may not be sufficient for you. And in certain environments, certain cases, certain
enterprise businesses, that's completely unacceptable. And so you have to manage that
in a context of building out enterprise agentic systems, which is what we're doing across our
infrastructure with our partners. So that's an important part of what we're doing.
On that point as well, I think it's a good segue to living your ethos as well. Internally,
you've thought a lot about AI from an investment perspective and applying it to your companies.
You've also used it internally. Where do you feel like you get the most
impact or benefit from using AI at Vista? The most. It's across the board for me.
I've got an agent. My agent's name is Q that I use every single day for conversations like this.
I'm about to go sit with an investor. I'm about to go sit with a CEO or someone who's interviewing
me. I'll type in the five questions, eight questions, 10 questions, whatever it might be,
and it will come out in my voice. Here is how Robert would answer that question in bullet point
and in text and prose. We should try that afterwards and see if what your answers were
relative to what Q's answers were. How similar do you think Q is to you now? And I'm sure it
trains and gets better relative to what it was and how you would answer. It's multimodal now.
We've changed the modality a few times now because in the early days it was
efficient. It would take somebody in my team a few hours, whatever it might be, to prepare for
this interview. Now I can literally type it in on my way over here. If I bother to read it,
it'll likely get close to it. But there's other parts of utility. How do we deliver information
back to our investors? The data constructs. This investor is in this fund, this fund,
not in this fund, this fund. I'm about to go sit down with them. I need to literally type in,
and it drops down and says, okay, here's what they're invested in. Here's what they've returned.
Here's their cash flows. Here's what they've gotten back in the last 30 days, 60 days,
90 days, 120 days, whatever. Massive utility. What it took people weeks to do,
it's done in seconds for me on my way walking into the meeting. Great utility there.
Even GPTs that we are now using for analysts, associates in terms of how they're preparing
their memos for investment committee. And how I'm contrasting them versus, well, we've looked at 12
businesses in this industry in the last 15 years. What went right? What didn't? All those sort of
things in the industry. What were the presumptions? So every part of our business is being
affected by this. So when you said, okay, which one's more? I can't say it all depends on what
I'm using that day at that moment in that time, but I'm seeing the efficacy. Again, this gets
back to that personal productivity is really, really, really high for that, for me being a
knowledge worker in the role that I am in. But I can see the utility of our industry and our
business and building. Again, I feel confident about what we are building internally and how
it is affecting, I'll call it the data, because of the way we've built it. I'll call it the
RAG architecture. And the way I like to think about it, we bring models to our data. We don't take data
to our models. Okay. And that's an important paradigm to think about in this environment.
Unpack what you mean by that.
Less than 1% of the enterprise data actually sits in these LLMs today.
Okay. I think IBM had talked about that a couple of weeks ago. Okay. If you now start taking your
data and putting it into others' models out there, it gets absorbed in ways and used in ways that
you don't get the utility and benefit. But if you actually take these models and apply them into
your construct and utilize them in your environment, you're the only one who gets the benefit of your
data. Sovereignty and dominion over workflows and datasets. I know, right? Yeah. Okay. That's
all important and important construct around utilizing AI in the enterprise.
If we think about that concept, sovereignty and dominion over your own data, let's take that in
the personal concept. So neither of us grew up with AI. Younger people or investors or
executives, founders now are growing up with AI as a tool that they can use.
What do you think are the skills that they need to have in the age of AI? Different, similar to
the skills that we had to have when growing up and thinking about how we wanted to build our careers?
We had a lot of emphasis on STEM education. Being able to solve, come up with algorithms
to solve complex problems. That isn't going to be as useful because these systems will be more
effective at it later. So there's one argument that says, let me be effective at managing agents.
And that is a skill you and I didn't have to have. That is going to be for certain people in their
jobs, critically important. For other people, it will be more important to understand how to
communicate as a human being to another human being. To spend time rather than the sales agent
worrying about, okay, how do I pull together this information? How do I connect with that person as
a human being so I can actually make this sale or whatever it might be? So it depends on your job
and your role and what your plan and ambition is as to how you should utilize a gentic world
or a gen AI world that you're now living in. A couple of my kids are creatives.
And I remember two Christmases ago, you know, you should start to use these for, you know,
you're creative. No, dad, that's not. Okay. Okay. Okay. Okay. Now I can see how they use it to
advance part of the work that they are doing, which is still creative. But now it's things
that would take them weeks to do. They can do in minutes, which gives them the chance to think
about the next order of opportunity. So I want to make sure we unpack some definitions here. So
tell us a bit about gen AI. Okay. We've been utilizing what I'll call, I'll distinguish a
little bit, you know, artificial intelligence for over a decade in our companies, machine learning,
artificial intelligence, we were actually able to create, you know, unique algorithms
that create highly efficient tools to deliver, you know, very specific outcomes using those
algorithms. When you start thinking about gen AI, it's a generative dynamic where you're generating
not just the next letter, word or phrase, but, you know, ultimately a, you know, an outcome based on
what the context is. Okay. Which now creates and gives you a whole different utility of opportunity.
And of course, you know, you've got some great people out there and, you know, and, you know,
Microsoft and Google and open AI and all those who are developing very unique models that can do this
at size and scale. And with the fidelity that actually creates new products and new opportunities
and fascinates the imagination, right? Agentic actually can start to not just take this generative
information, but actually reason with it and do something and actually act. And in some cases,
act based on an instruction or by an agent, an instruction from a person or an agent orchestrator
or human orchestrator, orchestrating multiple agents to do series of tasks, specific tasks
to do work. Okay. As opposed to individual tasks. And so the context of that actually creates a
whole nother set of dimensions of opportunity in the world of enterprise software. That is what
we've been focused on with our partners, Microsoft and Anthropic. How do we take this agentic idea,
make it real in the enterprise with the precision, with the capability that no one else can do? So
if you think about it, I've got 750 million users of our software. Each user is a knowledge worker
that you can actually take that knowledge worker's job and break it into a series of tasks
that agents can affect on. Some will be introduced or managed by the orchestrator, a person or
another agent, which can do it. And now deliver that into our customer's environment to create
value for the customer. So I tell my team, we're going out of the user business into the
agent resources business. So 750 million users of our software basically is equivalent to about
8 to 10 billion agents that we are going to actually be delivering into the marketplaces
of our companies and our customers that actually do work. And that's kind of creates a really
interesting paradigm of opportunity. One question on that last point. When do you think we'll see the
impact of all of this on companies? And I asked that question kind of zooming out and saying,
you saw the first wave of the internet. People got excited. It then obviously got over exuberant.
And then the impacts of the internet started to permeate and reverberate. 20 years later,
Amazon in 2015 was much different company than Amazon in 2000. How do you think about that?
What you saw in the first wave of the internet as it relates to AI?
A similar paradigm. The first wave went, of course, to the hardware vendors,
then infrastructure, and then software vendors really started to capture the opportunity set.
I actually struck a deal with this young man by the name of Andy Jassy back then, because he had
been building a really interesting technology called AWS to support this other business called
Amazon, had some compute. And we were able to come up with some unique opportunities,
leveraging their capabilities, our infrastructure, to now create a factory to go from on-prem to
cloud. As far as I know, they're converting more business from on-prem to cloud than anyone else
on the planet. That was a process that we saw still play out. I think there's still only like 50%
or so of the enterprise software companies that are now cloud. We're still on-prem. So it's 20
years later, 16 years. That's the kind of dynamic we're dealing with. I think something similar is
happening now. Of course, the hardware vendors today, NVIDIA, probably the most prominent of
which capturing a vast majority of the economic rent and attention. The hyperscalers are saying,
hey, we're going to build that infrastructure and capability because we've got a lot of free cash,
but we were infrastructure providers, and now we're going to provide a different type of
infrastructure, great capturing economic rent. Then you have NVIDIA deciding, well, I'm going
to get on that game too. I'm not just going to provide chips, I'm going to provide infrastructure
as well. And you've got a bunch of sovereigns out there saying, I'm going to do it as well.
So there's a whole infrastructure dynamic where you're now going to start to see the swelling of
overtime capacity, GPU capacity, but I still believe it's going to be the enterprise software
vendors that are going to capture, who? Sovereign Dominion, right? That are capturing the vast
majority of the economic rent because they actually have the workflows and data sets that no one else
has. We're seeing it now because we're doing it now, but it is not trivial. This is not something
you just plop an agent in, hey, go update these databases, go figure it out yourself.
You've got to now construct it in a way that actually creates the, I'll call it the high
precision outcomes in the workflow environments. So it may take a while to manifest broadly,
but as far as I know, we are on the forefront of making these enterprise class solutions
effective today. I think that's a really interesting point in the context of
the evolution of Vista as you've built the business also to now include a new set of
investors. You've worked with family offices and wealthy individual investors for a number of
years, but the wealth channel in particular, you've built a business to help them get access
to enterprise software. And I think it's a pretty interesting time to think about that because of
all the opportunities as you just laid out when it comes to enterprise software, the application
of AI and genetic AI to that. What made you decide to build a wealth business that would
help the broader wealth community access enterprise software? Yeah. You know, like all
things you sit back, you know, our institutional investors have been fantastic and great supporters
of our business and we will continue to support their businesses, their ambitions, you know,
to expand in this space and many cases diversify what their opportunity set is. We have found,
however, that there are a lot of private investors who don't have access to this. 97%
of these companies are private. You know, we have the ability to actually, I'll use the word
democratize the opportunity set. And so we should do that. I think it's the right thing, frankly,
for our country, you know, to be able to deliver a set of investment opportunities that can actually
change the retirement slash investment vector of a broad group of investors and not just those
who are investing through large institutions. So that's really what the thought and ambition was.
And I'm pretty excited about where we are so far. The other aspect of the wealth solutions
business you've built, you've really thought about AI and how AI can help wealth managers
run their own businesses better. There's a lot of software that helps wealth managers and has
helped innovate their own business, but AI is now helping them do that even more.
How do you think about applying Vista's knowledge and network when it comes to AI in the context of
helping the wealth partners you're working with run their own businesses? Yeah. And we get asked
that question a lot and we are working with many of them to help them understand how we're using it
and what the utility is. And I think we'll see over time what I'll call the emergence of very
specific applications and use cases of agentic solutions that enable them to be more effective,
not only in the services, but ultimately the sales of their product to their customer base.
Well, one thing that's interesting about wealth managers is they have sovereignty and dominion
over their data because their clients, that is unique data to them. They can leverage that.
That was something that you illuminated on this entire conversation and you've done it as well
at Vista with the business you've built. So fascinating conversation. Thanks so much.
Appreciate the time and good luck and have fun. Thank you. All right. All the best. Thanks.
Thanks for listening to this episode of Alt Goes Mainstream. I hope you enjoyed it. You can read
more about alts at my sub stack, altgoesmainstream.substack.com. Thanks a lot and have a great day.
