Software as a Coworker w/ Vista's Robert Smith

**Hugh MacArthur** (0:04)
Over the last three decades, enterprise software has gone through a series of transformations, from on-premise software to the cloud and to SaaS. Ironically, one of the biggest ideas in AI may bring the industry back toward a very old concept, keeping the most valuable data and intelligence inside the enterprise.

**Robert F. Smith** (0:20)
You have sovereignty in demanding over the workflows and the datasets. If you do not, you have to question your right to exist as a software company.

**Hugh MacArthur** (0:28)
That's Robert Smith, founder, chairman, and CEO of Vista Equity Partners. Today on Dry Powder, I'll ask Robert why he believes the firms with the best customer relationships, workflows, and proprietary data may have the biggest advantage in an AI-driven world.

**Robert F. Smith** (0:42)
I tell my team, this is the early days of the messy middle.

**Hugh MacArthur** (0:45)
We'll unpack the new rules of the game, how inference costs could erode margins, why data sovereignty matters more than ever, and what the messy middle of this transformation really looks like for investors. We'll also talk about exit pressures, amid headlines of a coming South Apocalypse, and why Robert believes the news of enterprise software's demise has been greatly exaggerated once again.

**Robert F. Smith** (1:04)
We've converted more companies from on-prem, the cloud than any institution on the planet. But what it did was it ushered in a major re-rating of enterprise software. Well, our sense is this is going to be the same sort of dynamic.

**Hugh MacArthur** (1:16)
I'm Hugh MacArthur, Chairman of Bain's Global Private Equity Practice, and this is Dry Powder. Robert, it's terrific to see you again. Welcome back to Dry Powder.

**Robert F. Smith** (1:24)
Thanks, Hugh. Good to see you.

**Hugh MacArthur** (1:26)
Now, the last time we had you on the show was at the end of 2023, and obviously, we've talked a bunch since then, but during that time, you were describing some of the early wins that you were seeing in AI. Are you seeing more measurable bottom line results across the portfolio today?

**Robert F. Smith** (1:41)
Yeah, I would say so. Again, we now have been in full swing with our agentic factory. We now have not only an agentic factory, but an agentic lab, and I'll talk a little bit about that. Fifty-four of our software companies that are through the factory, another 20 that are going through at this point.
By 54, the companies actually have agentic offerings, well over a half a million agents in the market that are actually producing revenue. We see at least 10 of our companies, again, early days that we think should achieve over 300 million and call it agentic revenues over the course of this next year, and we think there's going to be a massive expansion on that. And then, of course, there's a whole new concept around the cost of inference that is now coming in everyone's psyche that we've been dealing with in terms of how to deliver a low cost inference infrastructure for our own portfolio companies, but I'm sure we'll talk about all of those things.

**Hugh MacArthur** (2:35)
It's impressive because when I hear about AI and software companies in particular, the first thing I usually hear are cost efficiencies and productivity, and you hinted at those, but the first place you went was revenue. You're actually seeing real revenue gains from these tools.

**Robert F. Smith** (2:49)
Absolutely. The dynamic that we're all facing is, okay, what is the impact of AI on enterprise software? Of course, there are two components. One is expanding value and the second is increasing efficiency. The efficiency point, of course, is the cost dynamic. We have some real proof points there which are interesting. The last 90 days have changed everything in many respects, but we now are getting, in some cases, 80-90% productivity improvements in code generation across every single code base, with the exception of one, believe it or not, across all of our portfolio companies. These are things like completion rates and faster delivery and an ability to sustain productivity improvements in the product side of the house, which of course is one of the key P&L line items we've had to deal with as software developers.
Customer operations, in some cases, we've seen 95% call it deflection, and in some cases, 95% backlog reduction. All of these sort of dynamics are affecting the P&L of our businesses positively. And of course, on the front end, which is called the sales and marketing efficiencies, enabling your sales agents to be more effective in enhancing bookings. But I think those are ultimately going to be table stakes in the world of enterprise software. If you haven't gotten to it, you better get to it.