Ep. 166 - Rethinking SaaS Go-to-Market in the Age of AI

Guest: Jim Curry, CoFounder at BuildGroup

Is AI really transforming your SaaS business—or just adding noise?

In this week’s episode, Jim Curry, co-founder of BuildGroup and former Rackspace exec, discusses how AI is transforming both the products and operations of SaaS businesses—and why most companies are missing the opportunity.

Jim shares how BuildGroup's operator-first, long-term approach to investing gives them an edge when helping SaaS companies scale, especially in today’s AI-first environment. He explains why AI should be seen as an infrastructure wave (not just a flashy product feature), and how CROs and CMOs can practically apply it to go-to-market execution.

We cover:

• Why AI gives incumbents an advantage (if they act fast)
• How to use AI to reduce CAC and improve GTM speed
• Where traditional SaaS GTM playbooks break down in an AI-native world
• Why demos and SDR workflows are ripe for automation
• How BuildGroup partners deeply with founders post-investment

If you're a SaaS executive thinking about where to apply AI for real impact—without waiting for a full replatform—this conversation is packed with insight.

---

Not Getting Enough Demos?

Your messaging could be turning buyers away before you even get a chance to pitch.

🔗 Get a Free Messaging & Conversion Review

We’ll analyze your website and content through the eyes of your buyers to uncover what’s stopping them from booking a demo. Then, we’ll give you a personalized report with practical recommendations to help you turn more visitors into sales conversations.

And the best part?

💡 It’s completely free.

No commitments, no pressure—just actionable advice to help you book more demos.

Your next demo is just a click away—claim your free review now.

Transcript

Welcome to SaaS backwards, a podcast that helps SaaS CEO's and GTM leaders to accelerate growth and enhance profitability. Our guest today is Jim Curry, co-founder at BuildGroup, a venture capital firm based in Austin, Texas that invests in emerging AI enabled SaaS companies as they start to scale. Hey Jim, welcome to the podcast, 

Jim Curry: Ken, thanks for having me. I'm excited to be here. 

Ken Lempit: Yeah. And we're really happy and excited to get your perspective as an investor in the space. But before we dig in, could you please tell us about your firm and a little about your background? What got you here? 

Jim Curry: Yeah. I'll start with my background. 'cause you, you called me an investor which I technically am.

But I tend to think of myself as a recovering operator. Maybe it's a recovered investor. I'm not sure how to describe it anymore. But I started my career in operations and tech. Spent a long time at a company called Rackspace where I ran product and strategy for us. Some of your audience may not remember Rackspace, but in the early days we were a leading web hosting company.

One point we hosted 5% of the world's websites, e-commerce sites. And then Amazon came along and the story was written from there. We were an early cloud player too. I'd be happy to talk more about that story. I learned a lot from that. In terms of big transformations, like we're currently going through now with AI but I did that for a bit.

I started an open source project called OpenStack which has been very successful in the private cloud space. So while Rackspace wasn't successful in the cloud space OpenStack was leading provider of private clouds in the world today, and by a lot of measures. Just behind Linux in terms of popularity with developers and code.

And I'm still marginally involved with that. I don't do as much with it anymore, but I spent some time running the foundation and, and also had to commercialize it within, within the Rackspace operating environment. But when I left Rackspace, I guess this was 2015 now since it's been almost 10 years ago, not almost, has been 10 years ago.

You know, I wasn't a hundred percent sure what I wanted to do. My partner, Lanham was my boss for a long time. He'd been the CEO. And we had a lot of conversations about how we can contribute back, like what's the right next step for us in a career. And I think we decided a couple things.

One is, we wanted to have the Rackspace experience again, but we wanted to do it 10 times over if we could, right? So how could you go and do that with a limited lifespan? second is we think our greatest strengths are as operators and contributing on the day-to-day working of a company. So how do you put those two concepts together?

That's where we came up with the idea for BuildGroup, which we wanted to structure a little bit differently. So I'll talk a little bit about how we've set it up. BuildGroup is a permanent fund, which means we don't have a typical 10 year investment window. And so we can make an investment, hold it for as long as we want.

Why is that? It's because most people think in very short term nature in this business, they think about, you know, trying to build things in double, double, double, double and sell. We think much more about how do you build things that are gonna last, they're gonna have a legacy. And to be honest with you, even if you were gonna try to sell something in four or five years,

You're likely gonna achieve a better outcome by investing in creating a vision for a long-term company, right? So for us, we wanted to be in a situation where we could distinguish ourselves in the market by being really focused on long-term and really helping companies go through that. Also, if you think about from an investing perspective, I've always felt the 10 year horizons pretty arbitrary.

You know, you imagine buying a stock today and having an arbitrary date to sell it 10 years from now. It doesn't make any sense. You would sell a stock based upon market conditions, company conditions, all those kinds of things. And so we thought that was a better way to go. Second thing is we wanted to run a really concentrated portfolio so we could actually do what we do best, which is engage with companies as opposed to spending all of our time sourcing and do that type of work.

So when we raised our fund, we raised three 30 million. We invested that in a total of 13 companies, but really only 8 of them are any size. So the partners are able to stay very, very involved with the companies day to day. Which, you know, has this pluses and minuses. We can talk more about that later.

But for, for us being able to have the, the bandwidth to be involved, to know what's going on, what was really important for us. And then the last thing I'd say about kind of our model, We really are focused on, from a strategy perspective, the initial focus was on workflow SaaS companies that had really good data assets.

So if you go back to 2015, you know, workflow SaaS was already an established category for investing in. People understood that that was a great category to invest in because if a human or machine depended on it for their job, tended to be sticky. Gave you a great opportunity for land and expand typically in the right scenarios.

What I was really focused on in the early days of data science was how companies were taking this unique data, asset that only they owned out of their workflows and servicing it back into the products to change 'em from just a workflow product to think about as more of an insight and action product.

And so that's where we put our initial focus. I mentioned AI in our initial pitch deck, and I think I said it was probably 20 years off, so I was, I was wrong by a lot. But it just so happened that strategy played really well into AI. AI in a lot of ways is obviously just an evolution of data science.

And so the companies that we had worked on for this were really well prepared for that. And a result, we've had some exits early on in the AI space because we had companies that had built really good products that could leverage AI. Which going back to the original point about having a, a permanent fund, no one ever thought we'd exit anything.

And I've always used that as proof to say. Well, no, what I said was, we'll never force an exit because we have an arbitrary 10 year horizon. But if a company thinks it's the right time to exit because someone's offering them a, an unreasonable price or they feel like the opportunity is gone for them and they want to get out, we're not gonna stop that from happening.

So that strategy's played out really well for us so far. But but yeah, that's, that's a little bit of our firm. We're based in Austin, Texas. Austin's changed a ton in the last 10 years. We don't really have a regional thesis. In fact, we've only had. One company based in Austin, Texas is a company called Anaconda.

They're they're a great company. A lot of them have been based in East Coast, West Coast, Toronto, Canada. So we don't have a focus here. Austin's a great environment you know, but when we started this, it was much smaller. And it really wasn't more enterprisey focused.

It was much more kinda almost consumerish. But that's changed a lot and I'm hoping we can do more deals here in the future. 

Ken Lempit: That's, that's a really great background on the firm and almost sounds like a Buffett like approach to investing, you know, as opposed to a transaction mindset. 

Jim Curry: Yeah. By the way, thank you for saying that.

'cause I actually just got back from the Berkshire Hathaway annual meeting week and a half ago, which I am a huge of his and, i've been wanting to go forever and I just felt a calling to go this year and I could not have timed it better. But yeah, I mean, I often describe it as, you know, in the end we wanted something that was permanent and lasting.

And one thing I'd say about the fund is it has recycling built into it. So fund will be around forever continuing to kind of invest and build. But no, I'm a big believer in the Buffett model of sort of both permanent capital and long-term compounding. He is obviously a great example for that and, you know, we're trying to do our best to bring it to the, this space.

But yeah, no, thank you for bringing, bringing that analogy up. I'm a huge fan boy. 

Ken Lempit: Yeah. I, I think neither of us are alone in that. I'd love to talk about though this like the founding moments here where you and your partners, you know, moved from being operators to uh, investors to being venture capitalists.

What were the, surprising things that you learned or uncovered as you made that shift and, you know, has that impacted, you know, how you and your partners are working with founders? 

Jim Curry: I would tell you the first couple years to work with us was probably not great.

as you can imagine, if you go from being an operator to being an advisor. It's a real big change and I think for us initially, we had to remind ourselves our hands were not on the wheel. We're not, we're not there to drive. And in fact, we're not even there. I don't like the word coaching.

'Cause I don't like the idea of when people say coachable founders we're there to be a partner. And ultimately they're in charge and we're there to help them get better results from themselves and from their teams. And that's a hard transition. And I would say that anytime you're looking at working with a VC, that's a former operator.

Maybe avoid 'em for the first couple of years and until they figure it out. So I think that was a difficult thing to do. But I think that what was surprising to me about the industry was how unengaged, especially these earlier stage, so we invest, or I didn't mention the stage, we invest, we do early growth.

So typically sort of three to 8 million in is where we get involved. And there are a lot of really good early stage firms in there. But they're not that involved with the companies. I, I know this from the Rackspace days. We had great investors. I have nothing but great things to say about 'em. We had Sequoia, we had Norwest really good people, but they're very, very busy on their portfolios.

This is kind of the downside of the power law outcome for early stage firms. They have a lot of things that in their portfolio, they can only spend so much time on each. And most of what they do is focus on interactions at board meetings, introductions, networking, that kind of thing. And so for me it was unusual and I found it interesting to watch how people responded for us to say, Hey, I actually want to get in here and really try to help and try to engage and try to figure out that right model for engagement on how to do it not only with the founders, but with their leadership teams behind them, and then ultimately with the rest of the board and the investors.

 We found ways to do that and, and oftentimes I think I could sum up my role it's almost like a super executive chairman in some way. And I don't like the term executive chairman 'cause I'm not the boss. But I do feel like a big part of what I can do is help drive alignment. One things that the CEOs don't always have the luxury or never really have the luxury of doing in a fast business, early stage business is really lifting their head up.

Jim Curry: They're focused on the day-to-day. I know I've been there. They're just really kind of grinding. And so what I really try to do is be the holder of the North Star, which they've defined, here's the strategy, here's the execution plan against that strategy. How do we stay true to that? And keeping everyone aligned around that.

And I find that when we do that, it's actually relatively unique. The other thing we do I would say I found is the level of diligence of the stage is actually relatively low. I think what most people bet on is category and team, right? And ultimately I'll tell you, those are two great things, right?

If you get tailwind behind you and get team, that's really important, but you can't really understand the business and you can't really be helpful in the business unless you actually understand how it works. And so I, I do like to get in there and spend a lot of time that is actually in some cases has manifested itself in me or my partners having sort of a temporary day-to-day role in the company.

I, One of our companies, I was head of product for a period of time. I was head of sales. I ran FP and A to really kind of help understand it. So, fill a gap for the company, but also learn how the company works. So, we got really involved in that, and I think that because of that, our model feels very unique to people.

But I also tell them, you have to be prepared. It's different, right? If you're looking for an unengaged person and you just want some capital and someone to show up four times a year. Don't take money from me because I'm not gonna be happy because I enjoy business. I actually describe to people this way investing, eh, it's okay.

I love business and I love business problems, and I love business strategy. So getting involved in that's really important. And that's where I think as a firm, we can generate the best offer. So I think that's been the, the biggest difference for me. And I do think we've come a long ways in the 10 years we've been doing this and gotten much, much better at it.

And in, in some ways. I almost feel like we started with sort of a a job shop, professional services model that we've really evolved into a highly productized model with a much more efficient way for how we find a company, how we engage with 'em, how we diligence 'em, and then more importantly, once we get a investment closed, how we really work with 'em day to day after the fact.

Ken Lempit: Sounds like something that ought to be documented out there. You know, it is.

Jim Curry: Well, we have a, a lot of extensive stuff here. We're actually, you know, we're actually starting to put some of this into what I would call some open source frameworks that we're gonna share. I'll give you a simple, a simple thing I do because I mentioned the North Star.

Every company, I ask them to write their strategy down every year. And doesn't have to be long. It can be a couple pages, it can be five pages. But literally, what are we trying to do this year, longer term and, and really articulate. And I do have a format I like to use for that. Once that's done, we use that to decide what is the operating plan for the year.

And does that match the strategy? Are we putting resources against the strategy? Are we doing stuff that doesn't make sense? I always find that people are doing stuff that's off strategy, even in really small, early stage companies. And then the last piece of it is, are we spending the money to support that as well?

So they're related to operating, but they're different. And then throughout the year, whenever someone like me or anyone else in the board or a company has a new idea, I tell the CEO dust off that plan. Look at it and say, is it consistent or not? And if it is, send a note back saying, great idea.

We're gonna incorporate this in our strategy. Or send a note back saying, Hey, Jim. That's a really interesting idea. By the way, I've reattached our strategy. This doesn't look like it fits. Are you proposing to change to our strategy? And I have actually found this exercise is really good because it gets everybody anchored and gives us sort of an objective

template to work off of, to make sure everything we're doing is on track. Even early stage companies can get off track really, really quickly and start doing things that don't make sense towards their short or long-term strategy. So a big part of the focus, that's just one simple thing that we do every year with them, and I think it makes a big difference.

Ken Lempit: That, that sounds a lot like the EOS method to me. 

Jim Curry: Yeah. I'm an EOS. I'm an EOS and I'm a big believer in it too, so, yeah. No, it's very similar to that. 

Ken Lempit: let's kind of move on. We have a lot to talk about, and that was really cool, kind of setting the table conversation. I, I wanna talk about AI, and it's gonna come up a couple of times as we, as we go through the conversation.

But when we were prepping for the episode, you talked about AI as a feature, not a product. And I'd like you to walk us through your perspective on how existing SaaS companies can effectively leverage AI. To transform both their workflows and the insights they generate from the data. And by the way, I've always been a guy who believed proprietary data was the coin of the realm.

Yeah. So definitely on the same page there. But yeah, kinda share with our listeners, especially those who might be operators, founders themselves, your perspective on those two things. 

Jim Curry: I often describe AI as an infrastructure wave. And we've been through a number of these waves. The SaaS wave, the multi-tenant cloud wave the open source wave and mobile data science and, and what does it mean to be an infrastructure wave?

To me, it means that everybody gets to take advantage of it, right? And so when people think about investing in AI, I think a lot of what you're seeing now, the 70% of venture capital dollars, they're flowing into kind of the pure infrastructure, enabling the layer stuff for a great extent. Not all, but a lot of it's flowing into that area, right?

Same thing happened in the open source realm. What I'm most interested in is how do you take these technologies, these big tectonic shifts, and incorporate them into existing products, right? So I did this at Rackspace, we were legacy client server. We put up servers for people. We maintained it where they were all dedicated environments.

Then cloud came along and we had to transition really, really quickly to being a software company and building a multi-tenant cloud for people, right? Big change. So I could go through a lot of examples like that. In this world, this is a really interesting change for a lot of reasons. Number one, oftentimes incumbents don't have an advantage.

I think incumbents have a tremendous advantage right now with AI. And so if you think about it I talked about our thesis, right? Data enabled SaaS workflows is what we used to call 'em. Now call 'em AI enabled SaaS workflows. If you own a workflow, you own really valuable data. I mean, I could go through any example, Marketo, HubSpot, Salesforce, all the stuff we know on the go to market side.

They know tons and tons about not only our data, kind of static data, but how we process that data, the workflow around that data, the team interactions on that data, incredibly valuable, valuable stuff, right? AI is not valuable without data. We all know that, right? It's really not valuable without proprietary data.

Otherwise anyone can get access to it and start utilizing it. It also requires expertise. Like being able to finely tune and train these models to serve an outcome that you are familiar with is also really important. So, SaaS companies, own data. That data is proprietary. Those SaaS companies, if you've been in business for a while, you probably know a lot about your users, whether it be in the marketing space, manufacturing space, maintenance space, you name it, you're an expert and you're gonna know how to train those models.

We had a company called Case Text. They were a legal workflow company for helping lawyers do research and draft briefs. I invested in it five or six years ago when no one would look at it, and what sold me on it was the CEO in our first meeting over pancakes, I asked him to, what are you working on?

He turned it around and said, well, it's an early iteration. As you're doing your research, it starts pulling citations over into a brief, making suggestions. This is where I want to go with AI. I said, this is really interesting. So tell me more about your AI team. His AI team was all lawyers. They were all experts in how this worked.

And I said, well, where, tell me how you're, what you're really focused on right now. And he goes, we have eight years of data on how lawyers do this, how they do their work and do their research. And so when GPT-3 came out and GPT-4 came out. These guys were really prepared to take advantage of it, and their product capabilities exploded.

So I think that for existing companies that have really good workflow and data assets, putting AI on top of 'em is a tremendous advantage. I think when you look at the world of Agentic AI startups. I think that's a really hard space to win. If all you're trying to do is win, 'cause you've got the best trained LLM, you're gonna be competing against dozens, if not hundreds or thousands of competitors.

And no doubt, many of them will win because they're gonna be funded with lots and lots of capital and have the opportunity to run for a long period of time to build the data sets they need. But the reality of it is. These companies today that exist, that already have this with these valuable data and the workflows are already in a position to start doing this kind of work.

And so we're doing this with our existing companies today, which is let's take AI and really focus on how do you incorporate that into the product to deliver a better outcome.

You asked the second question, which is about the operation side, which to me is also really interesting. And so I can tell you how I'm starting to think about that. So a lot has been written about these early stage from scratch startups that are scaling without any people. And I, it's absolutely happening.

We're, we're seeing it in a lot of the stuff that we're looking at today. I think that the time is here where firms are gonna have to start getting serious about looking at the operational side of their business and starting to decide where can I at a minimum, augment humans with this capability, if not replace humans.

And I think that this is actually where the startups will have an advantage. They don't have the legacy teams, they don't have the legacy cost structures, so they don't have the data necessarily, but they're able to scale from scratch with an operating model that works. We are in a, a position where I think it's gonna be very common to see both fast growing startups that are not burning or have very limited burn.

It's just a unique position to be in. And so we have been working with all of our companies to take not only the stuff that we do with AI on the product side, but really working to incorporate it into their business. A big focus for us, is on the go to market side. Go to market is all human interaction and AI is great at mimicking and in many cases today, bettering human interaction, of course on the backend.

I think there's a lot of interesting stuff there to do too, and I can talk about that. Coding's obviously an example, but a big part of what we're trying to focus on is the go to market side where I think the opportunities, especially in the next 12 months are gonna be tremendous. 

Ken Lempit: So on the go to market side, you're talking about for your portfolio companies?

Correct. Enhancing the efficiency of their own go to market. 

Jim Curry: That's correct. Yeah, we have been going through and looking throughout the funnel at opportunities and, and let's be clear. The market right now is all over the place in terms of the maturity of these technologies.

It's also all over the place, in the willingness of customers to accept an interaction with a chat bot as opposed to a human. But we are seeing areas that it works. So, you know, you can start up top of funnel with how you're basically getting leads in, how you're parsing them, how you're doing.

That kind of work is all interesting. But I've found a couple areas I'm actually most interested in. One is we've been spending a lot of time looking at demos. Demos are one of those things where we have a lot of dropouts in a pipe, right? So the example I always give, like to give people is you have a, you have an SDR, they've got a, a really interested lead, they're ready to get a demo of the product.

You hand it over to an AE to go and process that lead. They give 'em a Calendly link, and the next time they can do a demo is three days away. You know? And again, the example for me would be if I called a buddy of mine and said, man, I've had a really rough day. I'm ready for a beer. He goes, that's awesome.

Let's do it in three days. Well, I may, I may want it, I may not want it in three days. So the ability to capture that person at that time is really important. But the ability to also make the demo work is really important. So there's actually some pretty good tools now. For cutting custom video demos.

We're getting closer to a world in which you can stand up, tear down, and customize real demos. We're not quite there yet, but it's an easy way to start experimenting within our pipe. The second area that we're really working on is SDR replacements. And SDRs are interesting, right? And by the way, you should be clear, I never think people are going away.

I'm a big believer that there's always gonna be a role for people in these processes, but I'm also a big believer you gotta push the bounds on this because everyone else is gonna do it. And you need to get ahead of it because it's not, you're not gonna be competitive otherwise. So a lot of what we look at with our SDRs is where can we go that we couldn't target with people because the costs are too high and try to run a fully automated process.

And if you think about SDRs as an example, SDRs are challenging in my experience, in a lot of ways. Number one, they have an incentive problem. They're humans. They wanna get paid. We typically pay them when they hand off a lead to an AE and an AE agrees to accept it, right? So their whole thing is just, I want to get an AE to accept this lead.

Now there's lots of ways you can get an AE to accept a lead that may result in actually not being qualified really well. The second thing is they tend to be very junior. They tend to be people right outta school. Maybe they're naturally gifted at sales. Maybe they're quick learners, but even in a great scenario, it's gonna be four to five months before they're relatively up to speed.

And then you typically face a 50% or better turnover rate on them. And that's just replacing the ones that you need to replace 'cause they're not good. Then you have these guys that are in the job for 12 months and they're ready to move on. So you have that problem with them. You can't really, because it's a human led process, you can't really stand up and at a time you can do a couple right.

And kind of build from there. If you flip that over to agents, you don't have these problems. Agents don't have an incentive problem. They're gonna process things according to the rules you give them. They certainly don't have a training problem. You can train them up really, really quickly if you've been really good in particular, keeping customer conversations recorded and all that kind of stuff.

And actually the interactions can be better. Like I've been noticing that certain people actually prefer the interactions. I think they know where they're talking to an AI agent, but the information, if you've really trained them well, can be good. So that's an example of it. And I think ultimately what we're trying to do is lower the marginal cost of growth, but we're also trying to bend the curve of growth and grow faster. And we have the luxury with our portfolio companies, a places to experiment with this with. And I'm actually relatively surprised at how many companies are not going all in to experiment with this because I think it's just like with open source, when that wave really took off it was clear that you're gonna have to figure out what your strategy was for adopting it.

It's the same here on your operating side of things. You're gonna have to figure out how to use AI in your business and rather than do backend, I would encourage people to really look at the front end. I think it's an interesting place to really spend your time. 

Ken Lempit: Do you have a couple of recommendations of software products?

Jim Curry: No. No, not yet. ironically, I've talked to a couple of the companies, I've talked about pricing models. So for example, one of the SDR replacements I've been saying, you know, we're not gonna pay you based on interaction or some other model. I'm gonna pay you based on, on conversion. The same way I'd pay a person, right?

Because I need to make sure the economic's work. But with some of these, I'm actually gonna work on some case studies and I'm happy to come back, Ken, and walk you through those and we have 'em. like I said, most of the technology is pretty early trying to skate to where the puck's gonna be in 12 months.

And I also think. 95% of the companies I'm talking to are gonna be out of business within 36 months. You pick your category , I could probably send you literally, I could probably send you at least 30 alternatives in it. And you know, none of them really have a natural right to exist so far in my mind.

So also what you have to, you have to be ready for and, and I think it's actually relatively easy in this world. You have to be ready to replace things pretty quickly. And the way we're handling that right now is we're gonna run three at a time and just try out three different complimentary solutions.

Figure out which one we like best, and then just be in a position where if we need to move, we can move. 

Ken Lempit: Makes a lot of sense. I, I think the uh, interesting thing, and you, you tried to slip it in there on me was the, uh. impact on pricing model. Yeah. Where you're saying to the SDR solutions, I want to consume your service the way I might incentivize an actual employee, and they might have other pricing models in mind.

I think coming into those situations. 

Jim Curry: It's interesting, you know I think much of what we know about the kind of predictable revenue SaaS go to market model that has been in place since Salesforce really kind of you know, initiated it, it's gonna change. I don't know how, like, I think there's a lot of things, a lot of things that we're talking about are basically, well, look at what I'm doing.

I'm taking a legacy model and I'm applying new tech against it, and I'm, I'm actually bifurcating those model, I mean actually dividing that model up into very traditional ways, assuming had people, that's not necessarily the right way to do it. And I think we're gonna find that out. But the same thing's happening on the software side of things.

They're, they're doing the same thing and they're trying to price it with a legacy model. So in the, in the middle, we're trying to come up with something a little bit different. And in my mind, anytime we're talking about something that literally is a replacement for human labor, I need to make sure that my measurement of that human labor, it's actually better.

And not by some small percentage, it needs to be markedly better, right? So one thing I'd say to somebody like, look, switching over to you know, all, you know AI based SDRs, if that only results in a 10% improvement, I'm not happy with that. Right? I wanna see a real dramatic improvement in it. You know, that's within the, the margin of error for me and whether I'm not actually measuring it or not.

So that's what we're trying to figure out right now. So there's a whole lot of stuff in our own adoption of this that we're, we're experimenting with it and trying to figure it out. I've got a dedicated product team of about five people right now that have been working through with this to try survey in the market, make sure we have a good idea of what's out there and then really just try to put 'em in production.

And luckily we have some great CEOs who are really interested in trying this stuff with us. 

Ken Lempit: I wanna drill down just a little bit more with you in this area. What are the, like the areas of improvement? You talked about the SDR and that's I guess to book the demo or the meeting depending on your sales motion, right? What other areas and metrics do you think are ripe for early adoption here? 

Jim Curry: We're, we're experimenting with more, like I'll give you a simple example, AI enabled dialers that are better at parsing, theoretical, and actually what I would describe, you know. There are already tools for dialers that actually help you sort your leads.

They're better now with AI, is the way I describe it. So some of these things are not like um, revolutionary replacements, they're just evolutionary, right? And that this is an example of AI being leveraged on existing technology and delivering better products. So that would be an example of it.

SCR replacement's a completely new thing, right? Because you're talking about an agent that acts like a human right in the, in the process. So I think that's completely new, but we are looking at everything from how do you source leads and much of what I should give some clarity. I don't do a whole lot of enterprise and investing, so I don't really go to people who spend a lot of time outbounding for enterprise. A big part of what we work on is SMB and midmarket. Ideally with a lot of product led growth and ideally with a lot of inbound, right? So a big part of what I try to solve is how do I, I more efficiently sort through inbound to get to the right people to talk to. But we also do outbounding and I would say there are some tools that we've been playing around with.

Are actually a better at sourcing leads for us than the, than the typical thing because they do more they do more human-like research is what I would describe it as, right? So again, you can get them to go answer questions that some of the existing tools cannot. And by the way, you can go program this stuff yourself.

Like I, I've written GPT's to go out there and look for customers. And I'm not an expert at doing that. So I would describe the results as not as good as what I get from a third party. But I, I would describe it as like. It's noticeable that you can build these things that have a level of intelligence that we've not had in previous tools before to do go into that kind of work.

So we talked about SDRs, we talked about the handoff to the AEs. The question then becomes like, how much of an AE job can you actually automate? I think there are obviously exchange of documents. I think there's responding to RFPs in a lot of our cases, it's responding to, you know, SOC or other kinds of information.

There's all that kind of stuff. Anything that's information intensive is gonna be easy to go and kind of replace. And those are the kind of tools we're playing around with. We're also looking at what's available, being released by the, the existing vendors we're working with, right? So to a certain extent, there are gonna be things that come out from Salesforce and others that are good enough at least in the short term for us to use.

And so that's where we're playing around with it. But anywhere we have a human in the loop we're trying to think about how do we. Reduce the number of humans required in that function, and then ultimately, do we need them in the long term? To me, the most obvious one to replace, especially with our model, because we don't do a lot of outbounding, the SDR is the easiest one to replace.

AE is a lot harder. And so that's kind of been where our main focus is. 

Ken Lempit: It makes a lot of sense. And you know, we had the founder, co-founder of Syft AI mm-hmm. On the podcast a few weeks ago, and the ability of their tool to consistently surface, you know, likely leads by doing the human style research, you know, allows you to expand your reach in a way that you probably couldn't afford to, you know, if you were staffing more and more SDRs.

Jim Curry: It's a great, yeah, for sure. Yeah, no, I mean, I'll, I'll give you little things that even I do, like, I have a little chat bot that whenever I'm traveling I'm like, can you please go through my CRM and tell me who is in that region? Super simple. I get a report from it every time so that I can make sure that I'm going through and maximizing my meetings. Now could I go through my CRM and just sort by? Yes I could. But then you add other characteristics to it. Are any of these folks potential investment targets, can you prioritize it by who are potential investment targets? Who are the most important targets? Can you sort 'em by ones that indicated they were gonna be raising a series A or series B at this period of time?

You can go through and give it kind of parameters, but this is where like once you program A GPT with what, what you're looking for. It actually is really good. I've actually done a situation where I've had an SDR agent and I've had a customer agent and I've had the customer interact with and I've listened to it.

I turn on the audio so they can talk to each other. I've essentially had them train each other well, I shouldn't say I'm train each other. It's really the customer training, the SDR. But these things can train themselves. And it's really interesting. It's all based on the quality of data you give it, right?

The more data you give it, the better. But these things are really good at doing that. It is the number one thing I always find that's interesting about this is people don't try the stuff out and you know, if you're not in go-to market, if you don't do go to market for a living, that's okay.

There's something that you do that's go to market. And all you needed to do is go to ChatGPT, start playing around with your own GPTs and see like how you go and do this stuff. And I've done this for a number of my friends who are in the non-tech space, like how do I use this to source leads? How do I use this to potentially interact with customers, potential customers? And it's really interesting to show people how to do it because. The tools are there, they're powerful. And they're very easy to use. And I, you know, if you and I go back 20 years ago, I mean you, when you had to program anything, right? Even going back to data science, where if you were a business analyst, you ultimately still had to depend on a data scientist to deliver you an artifact.

You really weren't interacting with the data yourself. That that's different now. You don't need that as much. So, yeah, no, it, it's a really interesting world and quite honestly, I have to make sure that I don't spend my entire day looking at new tools. 'cause there's so many out there and my credit card statement tells that.

Ken Lempit: Well, so, so you're a really hands-on guy and you have a concentrated portfolio. So it sounds like there's a lot of opportunity for you to engage with the founders, and I think you touched on this in the beginning of, of the episode, but I'm wondering how, how do you as an operating oriented investor know when it's time to like let your people run?

How do you balance that? 

Jim Curry: You mean with, with new technology? 

Ken Lempit: Yeah. Like how do you balance your desire to introduce. You know, tactics or technologies to your investment teams. 

Jim Curry: Yeah. With 

Ken Lempit: the need to let them run their own show. That's seems like 

Jim Curry: Yeah, that's a great balance. It's a great question.

So I would tell you that like, I never mandate anything and I wanna pull model, right? To me, my job is to create something for them that's really interesting that they wanna try and experiment with, right? And Ultimately this one's a little different, right? There are things that we encourage our companies to do because they're the right thing to do from a business model perspective.

Right. You know, CER and I can go through examples of that. They're just, I know work and I want to make sure the company adopts 'em. And my job is to help kind of lay the groundwork for them, understanding what that looks like. This is a little different in that I'm actually trying to learn as well.

Like I, I've never adopted and use these tools. I've never put 'em into practice. So some of it's actually selfishly for my own learnings and ultimately to benefit the companies we work with going forward. Luckily, you know, what I've found within our portfolio is if you are a technically oriented C-E-O i.e. you're a developer engineer, or if you are a product oriented CEO you tend to be very excited about this concept.

And so then the question becomes, okay, how do I really get this to be attractive to you to tackle? And to me it's finding an area of their business that they would like to solve, but they don't have the time to solve it that this technology could fit. So the examples I gave you, demos was an easy one.

We have a couple of companies that are very dependent on demos in their sales process. We have a lot of stuff fall out at the demo stage, not because of the demo, but because we just aren't great at getting 'em scheduled. We've a lot of resource constraints. And by the way, the last couple years through the market we've been through, you've had to be very, very judicious on where you spend your money. Another one was a situation in which you had a much smaller customer size that you wanted to target that are perfect for the offer, but because of their size, their churn rate, the cost acquisition of you're using humans, it's very hard to go and serve them.

But if you can run a fully automated process that does not allow them to ever talk to a human and allows them to sign up, then you can run something better. So think about that one as an augmented PLG, whereas like. I might find out about a product. I go to the website, it gives me some basic info and says, Hey, would you like to sign up for a small, medium, large?

And you do. The difference here is when you come, we're gonna give you actually something that feels like a human, that's gonna guide you through that process. Ask you questions, find out what you wanna see. Help you really work through it as if you were actually going through a more traditional, not just product-led growth process.

So we found a way to do that, implemented there and, and doing this, and we're early stage in all this, right? We're playing around with it. Another thing I'll just bring up that we're doing that's also a relatively easy sell is everyone wants to figure out how to use AI to do more, more, and quicker coding. The challenge with that, with legacy code bases is it's really hard to do, right? Depending on the language you've used, depending on level technical debt. You know, at best you're probably looking at 20 to 30% of your code being able to be written by AI going forward. But if you started from scratch, you can get, as you've seen, you can get results that are much higher than that.

So one of the things we're trying to work through right now with our companies is can you actually build an automated tool to map out an architecture and have AI actually lead a rewrite of it in a way that AI could actually write code going forward? And we're really early in that effort. 'Cause anytime someone would come to me and say, Hey, I really wanna replatform and I wanna do it all at once and it's gonna take 24 months, I would be like, Ugh.

No, like that's just painful, right? Of course I want to, if we have to redo 20% of the code every year, 'cause the technical debt, that's one thing. But a complete replatform is hard. I think you actually can probably use AI to do much of the mapping of a replatform. I think you can actually have it do a lot of the replatforming and then once you do that, you can let it take over.

And especially engineering oriented CEOs are very. Into this topic. And again, I think this is one thing we have to figure out because this is where a startup and that competes with our companies that are all in the teens of ARR or better face a risk, right? They've gotta get that cost down.

And so we're trying to figure that out in rapid order as well.

 Kind of the last thing I want to ask you. You know, you're meeting founders who may not be as up to speed as you are about AI and the impact on their business. What should they be preparing for, you know, the SaaS firms targeting the mid-market, like your investments doing?

Ken Lempit: Yeah. Like where do they need to be looking? Maybe they're not, and I know you've covered a lot of the waterfront here in terms of 

Jim Curry: raising money or in terms of operating their business or?

Ken Lempit: Operating their business. The fundamental changes with AI, it's sort of a, what else you got? Question.

Jim Curry: I would say, you know, first of all, I found it intriguing. I've actually been invited to and joined a couple of groups of basically CEOs that are trying to figure this step out. There's a guy named Jeff Woods, who's a friend of mine here in Austin.

He has written a book on harnessing AI really at the strategic level, how do you use it for strategic planning? And he's pulled together some groups of folks that some of us have been talking about operating strategy. So part of it is like um, you know, I'm an open source guy, right? So I remember the early days, you know, kinda the 2010 period of time.

When databases were exploding in open source and the level of interaction and communication, not, I'm not talking about. Writing of code, I'm talking about people discussing how to use this, where to use it, when to use it. It was really impressive. That's one thing I've always liked about open source. Some of that's going on in the world, AI too.

And so I always encourage people, get a part of a group and look at ways to do things. I've subscribed to a whole lot of different newsletters where people walk through how they're using AI in their workflows. I saw one the other day and I forget who wrote it, but the workflows they use to replace Canva and how they get a better result on it.

And I thought that was fascinating, right? So kind of reading through all that stuff. A lot of times people ask me, okay, so I need to hire a head of AI. I need to hire someone to go and figure this stuff out. I say, well, not necessarily. You know, I tend to think of, first of all, I think I tend to think about two things again.

Product and operations. I think of AI as not a separate function. I think it lives within product and lives within engineering. And I think first and foremost you have to have a head of engineering and product that gets their head around the importance of that, of the product and really wants to go hire someone to own that and figure that stuff out, right?

So it does fit within product and engineering. The biggest challenge is to make sure that you have a head of product or engineering who are fascinated by this and want to do it. Not all of them are but you need to have someone that wants to go and do that. You need to have a CEO. Who's very focused on the operating side of things, on how are we gonna leverage this going forward.

Only the CEO can decide to go and do this. And I don't agree with a lot of stuff that comes out around big companies of like, we're not hiring anymore. We're just using, I don't, that's not the answer. The answer is, we need to go out and try these tools, and when they're good enough to replace people, we're gonna do that.

When they're certainly gonna be good enough to augment people in the short term. And also let's starting for looking for marginal cases on the horizon that we can invest time in now where we could deploy AI and it might be meaningful to our business. The examples I gave you are one of those.

And then just read and then just play around with this. I'm helping my wife write a podcast series for herself right now, and we're using ChatGPT extensively. I've, programmed ChatGPT tutors for my kids. I have, my son is building a website right now and I programmed a GPT product manager that's been interviewing him for a while, and we're gonna turn that into a

engineering product. We're never gonna write a line of code's a goal. So there's a lot of ways, it's just like anything else, you gotta go try this stuff and then you start having aha moments. Got it. I know how this can work. And there's a lot of great information out there. Yeah, I mean, I, I honestly, I'm not sure I've been as excited about tech as I am now.

I wish I was, 24, not 54, because I think the next 30 years are gonna be absolutely amazing. And it's a lot of fun to work on it. And like I said, you know, the one thing I always wanna leave people with is, you know, incumbents often have a disadvantage in waves like this. You know, again, the, the SaaS wave, multi-tenant cloud, the startups were competing against people with legacy architectures.

That was hard. And a lot of those legacy players went out of business. I think the same could happen here if people move slow. But you have a natural advantage in the product to adopt it. It's the operation side where you gotta go quick or you can get caught by the emerging startups that don't have that legacy problem and can basically scale , with no burn.

I think that's a world we're heading towards.

Ken Lempit: Well, that's a great place to land our episode. Jim Curry, thanks so much for being on the podcast. If SaaS operators wanna pick your brain or learn more about your investing criteria, how can they reach you? 

Jim Curry: I'm jim@buildgroup.com. Anytime anyone wants to reach out, feel free or on LinkedIn.

Happy to chat anytime. This is a topic near and dear to my heart, and really Ken, I really appreciate you having me and, you know, talking go to market's, my favorite thing, so I really, I really do appreciate it. 

Ken Lempit: And that's what we do here. We love to talk, go to market. I think you've inspired a lot of thoughts for me, and I bet our listeners are gonna love this episode.

If you haven't subscribed to the podcast yet, please do so. Wherever podcasts are distributed, that's the SaaS backwards podcast. My demand generation and advertising agency for SaaS companies is Austin Lawrence. We're at

austinlawrence.com and I can be reached on LinkedIn/in/kenlempit. Hey Jim,

thanks again for being on the SaaS backwards podcast. 

Jim Curry: Thanks, Ken. That was fun.