Yesterday I had a conversation with Rex Woodbury at Daybreak, and we talked a lot about AI adoption among consumers. The topic we're going to unpack today is the other side of that coin: how businesses are adopting AI, and the different factors between SMB and mid-market companies and how they're thinking about things versus the enterprise, and where you're seeing some opportunities for adoption in a pretty interesting way. I'd love to start with a quick intro to you and Pathlight and where your perspective is coming from, and then we'll dive into the thesis about AI adoption.

I'm the co-founder and managing partner at Pathlight Ventures. We're based out of New York City, investing out of a $75 million fund focused on early-stage investing. Our sweet spot is usually seed, but we'll also do pre-seed and the occasional Series A. We like regulated industries, as well as applying AI to good old-fashioned born vertical industries. My background prior to starting the fund: most of my career was as an operator. I was the first biz hire and head of growth at companies like Plaid, Ramp, and Alloy. So you can imagine I spend most of my time thinking about and assessing distribution and go-to-market. One of the areas we're really excited about in the application of AI is AI use cases that almost look more prosumer, selling into SMBs and SMEs, but I'm sure we'll go into it in a bit.

I'll start on the enterprise side, because I think that's been a bit more fleshed out. It's obviously been at the board level. The executive team, the CEOs, the board are asking, "What is our AI strategy? We're getting pitched by OpenAI, by Microsoft, on adopting these new LLM tools." And so all these enterprises have opened up AI budgets, or AI experimentation budgets. That's where we're seeing a lot of the early adoption: people saying, "We need to figure out something on the AI side." Whether that's partnering with an incumbent or partnering with a startup, those are interesting entry points into those areas.

Where we've actually been more interested and excited has been this SMB and mid-market use case, where these are businesses that historically have not been super well served by technology. With a lot of the founders we've been talking to in the SMB market, for the first time in a long time, you're able to sell magic to these businesses. With the previous generation of vertical software, at the end of the day, I'm selling you some piece of software. You have to get trained on it, you have to integrate it, you have to implement it. You may have a small team, and you have to get them on board and convince them to use it versus their existing ways. Whereas now, with LLMs and voice and other advances on the AI side, you can actually sell them a product that just works. You can sell them a product that calls interesting prospects for them, that picks up the phone automatically and sub-qualifies customers, that just actually does work for them.

Similar to what I'm guessing you talked about on the consumer side, you're able to go talk to these businesses that have historically been ignored by technology and give them something that feels really custom-built for them and extremely logical. And the most important thing is it has really, really obvious ROI, from a revenue perspective or even a cost-saving perspective. That is something I think we haven't seen in vertical software for a very long time, especially in this SMB market.

When you think about the types of founders building in this space, one of the observations I've seen you make is that a lot of those teams are often younger: young, scrappy builders that are identifying a problem and building for that SMB and mid-market world. Why do you think that is? Is it unique to AI, or is it something else about the current generation of founders?

The teams that we've seen win this SMB and mid-market are often much younger: less than, say, five years of experience, first-time founders. The one thing that they do have is some impressive technical ability, but oftentimes limited domain experience. They're not experts in these spaces, and more often than not, it's usually a couple of pivots away from going into these areas.

What's been more surprising is that the typical young Silicon Valley advice was, "Go work on problems that you've worked on before." What we've seen with a lot of success for some of the younger companies is, no, actually, there are all of these verticals that exist in the rest of America. They're not in the New York Cities, the San Franciscos. You have to go out to South Carolina, you have to go out to Texas, you have to go into the suburbs of whatever region you're in, and actually go talk to the customer. So the one unifying thing that we typically see is a lot of iteration and deep customer discovery across these SMB segments.

And the cool part is that these customers are actually willing to hop on a call. If you're saying, "I'm a cracked engineer, I want to build stuff for you, and I'm going to automate this piece of your work for you," more often than not, those people are down to pick up the phone. And the fun part is they're also used to getting cold called. Their job is often picking up the phone and talking to customers. So they have a young person that's willing to learn about their problem sets and solve a problem set for them, and they're actually open to that. So one, we're seeing a bias of people wanting to help younger people, and I think young people should 100% lean into that ethos and mentality.

And two, the interesting part that we've been seeing on the AI side is that these SMB buyers are willing to pay more than they originally would have for a piece of software. Historically it was, "Oh, you're selling this piece of software. It's going to cost me maybe $5,000 a month, maybe $1,000 a month. Do I really want to pay for that?" There's still this full opportunity cost of, "I have to learn and get trained up, and I don't really want to do it." Whereas now it's, "If you implement my automatic voice calling thing, it'll literally bring you more money, and you don't have to do anything." I'm willing to pay whatever for that. There's some sort of rev-share thing that I'm willing to do, because it's low downside for me and high upside. If it brings in more business, that's a super easy ROI. And all of a sudden, on the founder side, the ACVs actually look much more interesting, and the velocity of your ability to actually close that sale is much more interesting.

So there's this combination of the market having now opened up into new use cases and new modernization opportunities for SMBs. I just think that these are users that have historically been very under-covered. I think venture historically has considered this category to be quite uninvestable. Vertical SaaS is almost a four-letter word. You constantly run into TAM issues. Whereas now, I think AI has dramatically increased the quote-unquote TAM of what you can sell into these markets, and it's a market that has been historically very underpenetrated. And I think young people are really good.

Do you think the opportunity here is primarily unlocked by an increased appetite from people in these verticals to dabble with AI, where, to your point, there's a low-downside, high-upside opportunity? Or is there something about these verticals that has gotten more accessible? Even five or six years ago, by and large, the closer you got to the SMB customer segment, the less attractive you were as a business, because these are highly fragmented, they're low ACVs, and they're difficult to sell into. It feels like today that narrative has really shifted, and it's something that's much more attractive to go after. I'm curious if you feel like that's a very unique AI angle, or if these verticals have just become easier to sell into.

I do believe that on the business side, there is increased competition. One of the interesting tailwinds that we've seen is that you have both new traditional PE roll-ups, as well as now venture-backed roll-ups. And the typical motion for a lot of these companies is usually, "We're going to start off with a small SMB, maybe a single-owner, sole-prop shop, and then work with them on the product." Some of those will have multiple locations and will start to expand out to different locations. Then a very common go-to-market path now is figuring out the top-down motion via a private equity firm. That gives you the benefit of size and scale. You can start to do the one-to-many approach by going top-down via PE and PE roll-ups.

And on the same side, the SMBs, the sole owners, are competing with the private equity shops, and so they want to figure out edges for the first time. I do believe it's actually become increasingly difficult to start a small or medium business. And I think more and more of these business owners are becoming more sophisticated, in that they have to be using best-in-class technology to be able to compete, to acquire users, to retain them, and to grow their businesses. So there is also, I think, this time and moment where SMB owners and operators are more open to technology than before.

There's one category I've heard you point to specifically as one that feels primed for some interesting opportunities: GovTech. What's the unlock there? Do you think there's an opportunity to build sizable businesses in that category?

We believe that we are starting to enter a once-a-generation window to build technology for state and local government. I'll leave the national government for another bucket. You have the obvious things around DOGE and the administration, and there are some interesting tailwinds around there. But even across state and local government, the other part that I think has been quite interesting is, yeah, we have a lot of cultural, political ties. Citizens just want better technology. They want to interact with the government. And so we're starting to see a generational shift finally happen on that demand pull.

It's a similar thing with our recommendation for starting in this segment. We've backed one company called Obvio that sells into the public sector. When you're selling into local government, even state government, at the end of the day, within a specific city or council, there are really just a few decision-makers that actually matter, that control the technology budget, that want to bring in new technology. And it's not that hard to find. It is literally publicly available on a website. If you take the effort to actually go talk to these people, map up their workflows, and build product for them, there is actually a high degree of excitement at the local level to try and figure out how to use best-in-class technologies and find some of that to actually listen to their problem sets. That's the interesting opportunity set that we're starting to see.

There's a timing in the market. Consumers are asking and demanding better products. We've just seen a lot more openness from state and local governments to partnering with newer companies to build products and co-develop new products.

One of the questions I think most people have in evaluating any AI company is how durable the revenue those companies are able to generate is, and how much of it is hype, experimentation and a willingness to dabble and play around with things, versus what feels like long-term, durable, high-quality revenue. I'm curious what your scorecard is. When you're looking at all these companies, yes, there's opportunity, but how do you know that opportunity sticks around?

That could be pricing moats, partnership moats, branding. I consider all those to be potential moats. The reason I actually like this SMB market, though, is that there's this common trope that SMB buyers are not that sophisticated. I actually think they are very sophisticated. If anything, it's the enterprise buckets of AI revenue that I get worried about. I think a lot of that is just experimentation revenue, top-down: "I have $100 million, and I can spend it on AI." But at some point, when push comes to shove and they have to talk to shareholders about what earnings they're delivering, that's the part that's going to get cut, because it's not actually driving real value. Whereas if you're selling to a business owner and your product isn't delivering value for them, I 100% guarantee they're not going to pay for it. Those business owners actually care about the P&L. It affects literally their salary directly.

And so the thing that we look for is: what did you say you would deliver from an ROI perspective, and are you actually delivering on that? If you're able to deliver on that, it gives you the right to win to build the next product. And then the key thing over time is building multiple products, building further penetration, and owning more of the workflows. But the first critical thing is: are you truthfully delivering some sort of ROI to that business owner, and does the business also believe that too?