AI's Next Wave Is Enterprise Software
Vista Equity Partners: Robert F. Smith Interview
Inside Alts with Robert Frank
Introduction
Artificial intelligence is reinventing almost every industry, especially software. Vista Equity Partners is at the forefront of a revolution in enterprise software that is changing the way companies do business.
We sat down with Robert Smith, Vista's chairman, founder, and CEO to talk about the private companies that will be the biggest winners in the next AI wave, as well as Vista's new agentic factory and his own personal AI agent he calls "Q."
The State of Private Equity
Question: Let's talk about the private equity space as a whole. You guys are obviously different for a lot of reasons, but it has been a challenging couple of years for private equity—especially from the perspective of high net worth investors and family offices. They haven't had much distributions or realizations. Returns have underperformed public markets, and it's tough to get a lot of high net worth investors and family offices to believe in the sector, to put in more money. What do you think next year looks like for private equity?
Robert F. Smith: It's interesting. The dynamic that people are now realizing is that those who have the ability to fundamentally create value in companies are able to continue to return capital at similar or the same rates that they have in the past—like us, for instance.
But the public market dynamic has been interesting. This whole idea of AI has sucked a lot of the oxygen out of the air for a lot of investors and pulled them into "Mag 7 plus" type companies. So that's going to be an interesting comparison relative to private equity businesses that, if you're really creating value consistently, will return capital consistently at the rates of return. I think they'll be historically quite comparable.
We've seen in our markets really interesting opportunity. In 2020 and 2021, you saw a run-up in the marketplace; they've now come down in enterprise software. So it's been a wonderful buying opportunity, especially as you're transforming businesses to become agentic.
The Three Legs of the AI Journey
We're going to see that third leg—what I call of the AI journey—really start to kick in.
First leg: Hardware vendors. That's where we're seeing a lot of the capital go.
Second leg: The hyperscalers. Those hyperscalers are now starting to build out the infrastructure and capabilities. Some may argue they're overvalued in some respects.
Third leg: The application providers. That's typically been the way that these cycles have played out. The application providers usually get the lion's share of the economic rent long-term once the technology has been diffused into those markets and technologies.
So that's really where we are in the cycle.
Private Markets: Where the AI Opportunity Lives
Question: You mentioned private versus public markets. For somebody who wants to invest in AI today, it appears that most of the opportunity—or a lot of it—is happening in the private markets. Where do you see public versus private in terms of the opportunity for investors to really invest in what will be the successful AI companies?
Robert F. Smith: The important thing is what will be. Again, application providers—those who are utilizing generative AI—are private today. 99-plus percent of these companies are private application software type companies.
Question: Why should we focus on that rather than the hyperscalers and who has the best LLM at any moment?
Robert F. Smith: There will be a few that not only survive but will thrive in that marketplace. But in the enterprise, you actually don't need large language models—you can actually execute using small language models and smaller footprints, utilizing different technologies that I think will actually create massive economic rent pickup.
Three Categories of Enterprise Software Companies
In our world of enterprise software, I'd say there are three categories—and that's why, as an investor, you really need to know what you're doing.
Category One: Agentic Transformation
Companies can become agentic. What does that mean? It means agents are actually performing the tasks within the workflows of enterprise software at a higher frequency and higher precision than what you could do as a worker using a tool.
That creates a massive economic opportunity. I like to say: AI will enable enterprise software to eat services. This creates a massive opportunity for growth—hyper-accelerated growth in these markets once these businesses are operating.
Category Two: The Rule of 70
In enterprise software, we've always operated under gold standards of the "Rule of 40"—that's a combination of growth rate and EBITDA margin. Well, as you apply AI and Gen AI across the cost infrastructure of enterprise software:
Product Development: We're seeing 30%, 40%, 50% productivity in writing code for new code—already. Existing code is somewhere between 2% and 12% productivity improvement. But over time, new code is going to replace the existing code, so you can actually develop a lot more code, a lot faster, at much lower cost.
Go-to-Market: How do you make your sales people more effective? Enterprise software typically has customer acquisition cost as one of the key issues. Gen AI enables that to come down dramatically because you have a series of agents that act as pre-sales and, in some cases, even sales capabilities.
Services: This is where you're providing services to customers. You can build out a series of agents in an agentic format that operate in a services component that actually brings your services cost down.
Back Office: All the back office operations become much more efficient.
As a result, you'll be able to boost margins pretty dramatically—almost double them in many respects—for existing enterprise software companies that are serving large-scale industry.
Those two categories are great places to invest. With that said, 97% of those companies are private. You're not going to have access to those companies unless you're coming through the private equity markets or investing in companies that are embracing this change in paradigm.
Category Three: No Right to Exist
Here's the interesting—and I think one of the most important—part: those companies that have sovereignty and dominion over the workflows and data sets have an opportunity to become part of category one or category two.
If you're basically repurposing data and information that has been available in the public marketplace and selling or reselling that content, you'll end up in category three. And there's a lot of them.
Question: So AI will eat those companies?
Robert F. Smith: No question about it. But categories one and two—you actually have the ability if you create agentic components within workflows that you have unique dominion over.
Remember: less than 1% of enterprise data is actually in the environment that has been used to train these large foundational models. Less than 1%. I think IBM put a study out about that. As a result, if you continue to maintain that dynamic, you can be in a unique position to deliver agentic solutions to an industry that no one else can, and capture what I call the economic rent associated with it.
That's why I say: AI will enable enterprise software to eat services—the services being people and dynamics that are providing the tasks that these agents could do much more efficiently at a higher rate and higher frequency.
The Agentic Factory
Question: Tell us about the agentic factory. How did you come up with the idea? Where is it going?
Robert F. Smith: For us at Vista, we've always thought about how you transform enterprise software and how to do it at scale—really bringing a unique value-added capability to the investing landscape.
In the early days, 25 years ago, all enterprise software companies were on-premise. You had some system—typically hardware like IBM 370s—with systems administrators writing code, and users accessing those tools.
Then we had a really unique technology come in called "hosted" or cloud. So we built a factory to transform on-premise software companies to cloud companies. As a result, we've now converted more business from on-prem to cloud than really any institution on the planet.
We found basically two-and-a-half to three times economic rent pickup when you transform a company from on-prem to cloud—reduction of the hardware refresh cycle, reduction of the systems administrator components. It became a much more efficient way to deliver enterprise software to customers.
But now we have a new technology, a new general-purpose technology: Gen AI. So now we have built the factory to go from cloud to Gen AI.
We have over 30 companies now through the factory that are actually producing revenue associated with this conversion. We'll finish the rest—the next 30 or 40 of our companies—over the next quarter or so.
It's unique in that what we've done is build out the capabilities to transform these businesses and actually create agentic engines and agentic products that work within the workflows and give the high precision that enterprises demand.
Consumer vs. Enterprise Precision
I like to say it this way: In the world of consumer AI, 93% precision is okay. You and I can say, "Let's go out to dinner, let's find a Thai restaurant." We go to the LLMs out there, and 93% of the time it'll find the restaurant and book it at the right time. Seven percent of the time, we're going to show up at a restaurant that actually has a Thai dish on the menu, and we might be okay with that.
That does not work in banking. It does not work in insurance. It does not work in automotive. You need to have a higher level of precision. You actually have to build agentic workflows with a higher precision output based on how these probabilistic models actually work.
That's what our agentic factory does. We now have the ability to do it at scale. Over two and a half years ago, we built out the infrastructure, and now we have the right partners—hyperscalers who have capacity and technological capability that we can then infuse into each of our companies to make this a reality.
Real-World Agent Examples
Question: For non-enterprise software experts—for people that don't really understand exactly how an agent works within a company—can you give an example of an agent that you guys have created that's deployed, what it's done, and how it's helped a certain kind of company?
Robert F. Smith: We have over 30 companies with agentic solutions in the marketplace. Everything from:
Healthcare: Utilizing an agent to help identify a healthcare worker to solve a problem that you may have utilizing some solution or system. We have a company called SimplePractice—we're now integrating it into two of the major hyperscalers so that as people are incurring issues, it will identify the right practitioner, actually book the appointment for you in a place that's close by, and has access to their calendars.
Tariff Management: We have agents that actually manage tariff dynamics for large-scale enterprise companies. They identify: "There's been a change in tariff in a particular location. Where can we get substitute goods, a different product at a different price, and actually deliver in the same period of time that we need to keep our manufacturing organizations operating?"
Sports Analytics: For our sports business, you can get real-time information and do "what-if" analysis. "What if we played that player?" A lot of people like to armchair quarterback or Monday morning quarterback these games. The system will actually run through various scenarios that have high precision in the output as to what, based on the statistics, it could have been.
It's a full range—it affects every single industry. Part of our goal is to ensure that our companies get there fast first and have the ability to deliver agentic solutions that provide real, tangible value to the customer set.
The High Net Worth Investor Channel
Question: How has that changed your investor base? For a long time, you were focused on big institutional investors, sovereign wealth funds. Recently, you've been very successful attracting high net worth investors and family offices. How do you see that group and its importance for Vista going forward?
Robert F. Smith: It's been an organic journey. Typically, we look for enterprise software companies to buy. The owner/founder of that business ends up with a liquidity event. They look at our returns and say, "Hey, can I invest with you?" And we're like, "Yeah, sure, you can invest with us."
Over 25 years, all of a sudden, you end up with a pretty large high net worth channel—made up of high net worth individuals, some of which we've provided opportunity to, and others through word-of-mouth dynamics.
From there, we've been contacted for years from the large warehouses and RIAs asking, "How can we invest?" So we started opening the window. We're one of the few that said, "Okay, we're going to limit the amount"—and we've gotten overrun with demand in some respects.
We're finding that this has a nice symbiotic dynamic. They have run businesses. They understand our value creation dynamic because we are software operators at the end of the day. They say, "Oh, we get that. We understand how you bring unique value to your companies." So that resonates with them.
It's been a wonderful journey. We'll continue to drive our institutional business quite aggressively, but it's been a nice benefit to have a new set of investors come along to help us expand our capabilities.
Is AI in a Bubble?
Question: When we look broadly at AI and this question of "Is AI in a bubble?"—obviously, in the beginning of any new technology, we have over-investment, misallocation of capital. Clearly, there will be winners and losers. The bubble question is not a binary one. How do you look at the types of companies that you think will be winners versus those that might struggle?
Robert F. Smith: For us, we focus on enterprise software that can become agentic. First of all: sovereignty and dominion over workflows and data sets. Do you have something that no one else has, that when you "agentify," you can capture more and more economic rent?
Question: In other words, broadly, it's the application of AI as opposed to the AI thing itself, the LLM?
Robert F. Smith: Right. It's the application of AI as opposed to the infrastructure. Infrastructure is a cost of goods for us. As more capital flows into it, it becomes a lower cost of goods.
So we are beneficiaries of, in some respects, this overcapacity build that's occurring. As the cost of tokens—which is really how we're measuring it—comes down as a result of all the capital flowing into that environment, it gives us the ability to actually increase the gross margins of the utilization of our agents.
It's a wonderful dichotomy if you're on the back end of the applications—a user of AI across these enterprise businesses.
Agent Q: Robert Smith's Personal AI Agent
Question: I know you personally use AI. You have an agent. What are some of the most surprising things that you've found, used, or done with AI in the past few months?
Robert F. Smith: My agent, Agent Q, has about 20 different categories that it monitors or utilizes every day—from the way we run our business here at Vista to the information and insights that I want from different parts of the world.
Every morning, I go in and look at: "What's actually happening in Asia? What is happening in LatAm as it relates to either global economic activities or specifically as it relates to AI?"
I have one of the agent components that looks at what have been the latest releases, what are the things happening. I've trained it now as to what I'm interested in—to understand what's happening in the releases of AI, Gen AI, and the companies that matter as it relates to our businesses.
I don't know if there's a whole lot that's surprising, outside of the depth capability and the speed at which it can return useful insights. And then, of course, you can query against it.
I also have the ability to look inside our own portfolio companies—we have about 90 or so companies—to monitor what is actually occurring: what pickup rates, what KPIs are important to me and to us that I need to focus on.
It's been a marvelous journey, and every week I talk with my team about what I need to tune a little bit better, make a little more useful for my application uses as an executive here.
AI's Impact on Employment
Question: When you talked about your portfolio companies, I know there was a lot of back and forth—maybe some confusion—about remarks that you made about employment. How it could affect productivity and employment more broadly, and also within Vista. What's your perspective on the job impact both at your company and broadly in society?
Robert F. Smith: It is clearly a highly effective technology that will create disruption in different forms at different levels.
In some cases, it's going to be a massive enablement tool for people to become more effective.
In some cases, it's going to eliminate the need for certain activities—those who are doing engaging back and forth with data systems. These agents, when taught and trained effectively, actually eliminate a lot of that work and can do it faster, at lower cost, more efficiently.
I'm going through a whole review now. We are actually actively deploying these agents to create efficiencies across our entire infrastructure.
Like I said: Not all knowledge workers will be affected in the same way. For some, there will no longer be that job category. For some, it will be a hyper-accelerant of their capabilities.
I tell some folks: "AI's not going to replace your job in some businesses, but the person using AI will replace your job."
Those who become a little more familiar and understand its capabilities first will have the ability to actually be a little more extensible in their capabilities in the institutions in which they work.
Optimism and Historical Perspective
Question: AI is clearly making a lot of people very uncomfortable—a lot of companies, a lot of workers. They're not prepared, they don't understand it. How optimistic are you about Vista's future, about AI's impact on companies?
Robert F. Smith: The way I think about it: AI is a true general-purpose technology. It is going to affect every single business. And it is either going to be exponential for some or existential for others.
I'm optimistic in the way that we are embracing it, embedding it across all of our companies. This is now an exponential opportunity for us.
Question: You've been in investment banking, software for so long. Over your career, how does this moment feel to you as an opportunity versus all the other technological revolutions that you've seen?
Robert F. Smith: All the others have, in some respects—with the exception of the distribution of compute—been significant transformations in the global environment. We've distributed compute throughout the world with all the things attendant to embracing compute.
This is at that level—and potentially more—because the capabilities actually give the average person the ability to utilize technology at a higher rate than we had in just providing compute technologies. Typically, it was a higher-order employee that had the ability to use those tools.
If you utilize these tools well, actually any employee has the ability to utilize the capabilities of AI and become more effective. I think it actually has an inflection opportunity for individuals to be more productive, more capable, and frankly create a whole lot more productivity throughout our global economy.
It's going to be an exciting time to be an investor in this space.
End of Interview