Nic (00:01.225)
Welcome back to the show everyone. Today we have one of our favorites, another fellow VC and vertical AI enthusiast of course, Nick Tippman. Nick is the founder and managing partner of TipTop VC, an early stage VC focused like you could on vertical AI. He was a founding member and CMO of Greenlight Guru. If you guys haven't heard of Greenlight Guru, incredible story.

Bootstrap business grew to great scale and had a wonderful outcome to PE, which we'll hear more about today. And Nick translated that operational operating experience into investing. So he's now been backing Vertical AI founders at the earliest stages. And he's going to tell us a little bit more about his view of the world. And very excited for it. Nick, welcome to Verticals.

Nick Tippmann (00:55.318)
Yeah, thanks for having me, Nick. Super excited to be here with you and Luke. Longtime listener and excited to join and have a great conversation here today.

Nic (01:05.915)
Absolutely. The more Knicks the better.

Luke Sophinos (01:07.46)
It's the battle of the VCs today. I like it.

Nic (01:10.399)
The battle of the Knicks that love vertical AI. yeah, well, you know, we love our niches here. Well, Nick, do you want to kick us off with maybe you can do a better job in your background than I did? And yeah, just tell us a little bit about what you're up to at TipTop.

Luke Sophinos (01:16.155)
Yeah.

Nick Tippmann (01:17.48)
We're getting super niche already.

Nick Tippmann (01:34.53)
Yeah, for sure. I mean, you did a great job, so I add too much, but I like to think of myself as a founder, operator at heart, turned investor recently, grew up starting businesses, started a couple of companies after college before Greenlight Guru, raised some money with one, got into an accelerator with another, and bootstrapped a media company to a small exit after that. But then as you mentioned, most recent and relevant was part of the founding team, CMO at Greenlight Guru.

We were a B2B or are a B2B vertical SaaS for the med tech industry, help med tech companies get approved by the FDA and stay compliant once they got to market. Yeah, as you mentioned, we had a great ride there. As you can imagine, it's technically the first employee wore many, hats over our eight and a half year ride there. So always led marketing, led rev ops as well throughout that time period. Once private equity came in, got a lead product internally for a year.

was leading strategy and corporate development before I left as well. So had a broad range of experience at Greenlight that then translated over well to starting TipTop Ventures, which is a early stage vertical software and vertical AI specific fund leverage in my background where we write smaller collaborative checks of 100 to 400K alongside other leading investors like Euclid.

In addition to my background, I've also built an LP collective of primarily 95 other exited vertical software founders from places like Toast, Shopify, Procore, MindBody, Service Titan, et cetera.

Nic (03:13.21)
Awesome. What, you know, I know we want to get kind of deep in the, in the playbooks today, but I just wanted to kick off with, with this question. What, I guess what drew you to, you know, investing professionally and like, that start with angel investing or, know, particular insight? I know there are a lot of folks who are angels and maybe thinking about doing venture full time and like, how did that pair with your expectations?

Nick Tippmann (03:41.614)
Yeah, for sure. I think I have a little bit of an unconventional path in the venture, but I don't know that there is a conventional path when you talk to most folks in it. But I like to say my journey started back my senior year of college back in 2011, 2012. I was studying entrepreneurship back at the University of Indiana.

As I mentioned, always very entrepreneurial growing up, always grew up wanting to start a business like my dad, like my grandpa. When I got to my senior year, I realized that the question I went to college asking is how do you start a business? The better question is how do you, that I was looking for is how do you start a high growth venture-backed scalable tech startup? And being a kid from a small town in Indiana, a lot of those words were pretty foreign to me. But if you remember back,

Nic (04:32.766)
What?

Nick Tippmann (04:34.51)
Yeah, early 2010s, was kind of the beginning of VC blogging. And so some folks that probably on this show will recognize the names, people like Brad Feld, Fred Wilson, Mark Seuster, Paul Graham got turned on to their blogs pretty early. And I had this realization, was like, wow, well, three out of these four founder entrepreneurs all went on to become venture capitalists.

it probably seems like a pretty interesting gig here. So he became fascinated with the whole venture side of things, getting into tech and then operating and starting companies after school there. And then over the last 14 years in my professional career have been operating with the mindset of let's go be a successful founder operator. And then long-term I've always wanted to get into venture and raise a fund and build a firm. And I could get a...

into a little bit more into the why, why now and what got me so excited and excited about vertical AI as well. But that's a little bit of the backstory there.

Nic (05:39.294)
Amazing. Well, congrats on everything so far. Well, Luke, I know you had some ideas on, know, kind of directions to take it on the playbook side. Where do you want to go?

Luke Sophinos (05:52.123)
Yeah, let's do it. So I want to go super deep today since we got the two VCs who see deals all the time, you know, have a good basket of portcodes and seeing what's working, what's not. So I want to go deep on a few things. I think the first one to start out is the SaaS markets are crashing, right? They're getting absolutely hammered. What caused this really is a few of the AI companies. I think it was specifically

Claude, right? Claude launching some workflow tools and people are basically panicking and saying, you know, SaaS is dead. So I want to, I want to just kind of zoom in with this group for a second and talk about, okay, if that's happening, you know, what's next is, software dead specifically is vertical software dead? And, know, how are you guys balancing this? How are you thinking about this? You know, where is the world going?

next year, five years from now, 10 years from now. So maybe tip and kick us off and let's make it a debate. I don't think we're all gonna agree with each other here. So it should be fun.

Nick Tippmann (07:02.83)
Yeah, I would love to kick it off here, Luke. I do want to pause just a moment and we may have to cut this in the podcast here, but I don't know if you guys heard that, but I had a big, like, loud screaming sound in the background the whole time you were talking there. there, Nick, when you're coming on, it's a ton of...

Nic (07:15.613)
you

Luke Sophinos (07:20.932)
It says, yeah, it's his AC. We'll cut this out. They'll mute his thing when I'm talking, so it won't matter. Yeah, let's just all go on mute after we talk, just to be safe.

Nic (07:24.998)
Nick Tippmann (07:30.85)
Okay.

Nick Tippmann (07:34.958)
All right, right on. Sorry, apologies about that. I didn't know if it was gonna be, yeah, something ongoing or going forward.

Luke Sophinos (07:38.416)
No, they'll tell her, yeah, it's annoying.

Nic (07:43.101)
Can you guys still hear it by the way?

Luke Sophinos (07:45.98)
Yeah, so just mute, Nick.

Nic (07:49.274)
Okay, I'll mute.

Luke Sophinos (07:53.597)
All right, Joe, Joe's our editor. You can start again and then Nick, you can kick off your answer.

Nic (07:59.763)
Yeah, yeah, Nick, you start. I'm gonna go turn off my AC while you go.

Nick Tippmann (08:08.258)
Right on, Luke. Maybe you could set me up with the end of the question there again. Apologies.

Luke Sophinos (08:11.888)
Yeah, we'll run it back. I'll wait till he gets back.

Nick Tippmann (08:16.418)
Right on. Sorry, while you were saying that, I was thinking about the background music there, but yeah.

Nic (08:25.724)
Alright, good to go.

Luke Sophinos (08:27.804)
I'm gonna redo the intro. All right, so yeah, so I wanna go deep today on a few things. Since we have two VCs here, I think it'd be helpful to really unpack what's happening in the public market. So SaaS is dead is the proclamation. We've seen SaaS stocks get absolutely hosed the last, I don't know, call it week or so, two weeks at this point. All of this triggered by...

Nick Tippmann (08:32.536)
Thank you.

Luke Sophinos (08:57.532)
Claude and their product introducing some workflow tools. And so I want to hear from you guys who are betting really on the companies of scale and five years from now, 10 years from now, 20 years from now. And so this obviously is very impactful to all of our listeners. How are you guys thinking about what's going on? How are you thinking about what legacy SaaS?

Specifically, legacy vertical SaaS companies look like 12 months from now, five years from now, 10 years from now. Like what's going on in the market? Tidvon, let's start with you.

Nick Tippmann (09:35.202)
Yeah, for sure. Thanks for setting it up. And it's certainly been an interesting time in the markets over the last couple of weeks and all the talk about the SaaSpocalypse and is SaaS dead. And my take is I definitely believe the worries are certainly overblown. And if you think about what some of the root causes you mentioned, Claude releasing, cowork. So are people worried that?

Folks are going to build their own software and they're not going to need to buy software anymore. Is the seat based model going away and therefore the growth engine of these companies going away? Is quad code and quad co-work just going to be able to do all workflows and all processes and all work? So there's a couple of different headwinds coming at SAS right now. I think you got to pull those apart a little bit.

But I think my general takeaway and sentiment as a whole is SAS isn't dying, but it is evolving. And we're seeing, whereas the motes, we were pretty confident in our motes, in our system of records, our data gravity, our workflow gravity. And now in the age of AI, a lot of those motes are looking a little bit shakier. But I think this is really just setting up for...

a battle of epic proportions between AI native upstarts and the legacy SaaS incumbent providers. And as a whole, as a venture firm, we're definitely making our bet on the AI native upstarts being able to come in and disrupt these legacy providers. That being said, I don't think the AI native upstarts are going to make a clean sweep across the board. The incumbents have a lot of

of built-in advantages that if they're able to move fast enough and adapt that they can leverage and lean on. Just starting with the fact they have the relationship, they have the trust. If they are the dominant system of record, they have the data, they have the workflows where the AI native upstarts have a lot of advantages in their ability to move fast and their ability to build products that

Nick Tippmann (11:51.564)
Maybe some of the incumbents are thinking are cute or tools and dismissing them as point solutions, but not really understanding they have a broader vision. And if they're able to position them properly in the workflow, they may earn the right to own more and more of those workflows and the data gravity downstream over time. And so I think it's really the trillion dollar question of who's going to win these AI native upstarts or the incumbent SAS.

But I think this flash SaaS crash is a little bit overblown as you see some companies being all the way valued basically to their terminal velocity of cash flows basically being zero, which I don't think the value for the majority of software companies is going to zero. But certainly there will be a lot of disruption and this is.

the most uncertainty that we've seen in software markets over the last couple of decades and really since the last platform shift from on-prem to SaaS and cloud. But Nick, I'd be interested in your perspective and how you're thinking about this.

Nic (13:00.99)
Yeah, mean, look, I mean, there's always a lot more baked into public market movements than a narrative as simple as, know, AI is now going to eat these companies. I think, you know, we've still got some, you know, persistent inflation going on. There's pressure there. I think that, you know, something like only 70 % of SaaS companies, you know, hit guidance in terms of revenue growth. So like, I think there is some factoring in of

of both near term and then maybe due to AI long-term growth prospects. I don't know if you look at just median public SaaS multiples, they're not out of the band that they've traded in over the last couple of years. So I feel like you can't look into these swings too deeply sometimes. mean, ServiceNow drew down 10 to 15%, but they've had

near 100 % retention over last year. Sometimes it doesn't 100 % make sense. But I think another one that was hit pretty hard was monday.com. While that business, kudos to them, it's been a really cool story. I do think that it somewhat mirrors the archetype of the sort of business which people see as being increasingly exposed in the near term. If you think about

outside of the surface level stuff of creating your own website, what are people doing with cloud code? It's like, hey, create a dashboard for me, create a Kanban for me for my daily tasks or whatever, right? And now that's a far cry from we've got all the functionality of monday.com, but like people I think are just starting to now make the mental connection between what this could mean. And the biggest takeaway for me is that the most exposed

are these generalist, horizontal, kind of universal solutions, right? It's kind of funny because we saw this rise of, vertical SaaS is exposed. And I always thought to myself, why vertical? Like it's the least exposed.

Luke Sophinos (15:11.172)
I never understood that either. That was Dave Friedberg on All In and I had to answer questions about that for two years.

Nic (15:14.822)
I think honestly, yeah, same. And to be honest with you, I think Dave is a pretty smart guy. I think he kind of meant it in a different way, which was like niche-y, like something that's been around doing a millionaire for 20 years sort of thing that could be easily replicated. I do think, especially kind of generalists, like task tracking, stuff like that. mean, even just playing around with

you know, Claude code or open claw or whatever else is coming out. You can kind of start seeing how maybe, maybe you don't need to reliance on that sort of UI anymore. but I think it wouldn't shock me if, in another six months we trade back up to where we were in the public markets.

Luke Sophinos (16:05.37)
Yeah, so I want to, I have kind of strong views on this. I think if you're just, if you zoom out for a second and just think of a few different kind of pieces of software that we have today. So we have system of record. So your system of record could be vertical or it could be horizontal system of record being CRM, you know, EMR, you know, in the healthcare markets.

know, SIS in the education markets, right? So just core system where all the data lies. A step above that, you have a whole host of either point solutions or like very broad horizontal tooling that doesn't meet the definition of system of record. I think broad horizontal tooling would be like a DocuSign. It would be like a Monday. It would be a Trello, right? Things like that.

And then above this now you have a new category, is AI, right? So this is Claude, this is ChatGPT, OpenAI, et cetera. I think this is, and you can actually see this, we'll put it in the notes here, but OpenAI had a blog post on their new enterprise solution called Frontier. I don't know if you guys saw this. Basically, it kind of gives some takeaways into I think what they're thinking, which is,

the AI layer is going to eat everything except for the system of record. Like, I don't know if we can pull this up, Nick, I can look for it here. It's open AI frontier posts, but frontier launch. Literally the system of record is like the last remaining thing that they're not, at least in this blog post, it felt like they're not going after. And their whole view was like, hey, we're gonna build agents that cover everything above that data set. I think that's important to note because the system of record is

Nic (17:38.238)
Yeah, let me see if I can find it.

Luke Sophinos (17:59.587)
oxygen for AI. there's no, there's AI can't perform without the data set. I think if we steel man that for a moment though, the inverse, here you go. So it's your system of record here at the bottom. And then it literally has open AI taking everything above. And I think the agents piece is interesting where they have your agency, the internal company could build them. They have open AI agents and then they have third party agents kind of app story-esque, right? But none of these things work without

the data set. Back to the Steelman point though, I do think I have read a couple of blogs where people are saying, hey, well, technically an agent could just kind of go into the system of record and pull all of that out. And then you don't necessarily need that. have some views of this, but I want to at least pause there and get Tippman's view of this. And then we can go a layer deeper.

Nick Tippmann (18:54.038)
Yeah, for sure. I think it's a really interesting diagram and aligns well with our thesis and views here at TipTop and where we think the value is going to accrue in the stack. Maybe 18, 24 months ago, I saw another graphic. I forget where I saw it at this point, but I thought it was a good way to describe kind of the different layers of the stack. And they had foundational models at the bottom, kind of small language models.

infra tooling, then apps on the top, both horizontal and vertical. And our core thesis at TipTop is that the value over time is going to accrue in that vertical application layer. And then what you see with this graph is basically saying OpenAI is going to do everything but the system of record. And so we understand that that proprietary unique data that these

vertical solution or really any system of record, excuse me, for that matter is going to continue to have value. What's gonna matter less is the UI. What'll matter more is more the UX, the experience, the outcomes more than interface stickiness per se. yeah, so I think that's where we're gonna see the value occur. then back to just real quick to double click on one of my points from my firm.

previous comment of the impending battle between the incumbents and the upstarts to quote some data from TideMarks, our friend over at TideMarks, excellent 2025 vertical SaaS survey. In 2024, they surveyed all the SaaS company, think there's like an N of 250 or so, and it about 31 % had released an AI product and it was about

just under 40 % said they were going to release one in the next 12 months. Now in the 2025 survey, the products was up to 55 % of SaaS companies had launched an AI product. So a little less than half of those folks that said they were going to launch a product in the next 12 months actually did. And if you just extrapolate that out a year or two, that's by the end of this year, we'll be looking at 70 % of SaaS companies, 80, 85 % by the end of next year.

Nick Tippmann (21:16.896)
And so this is where the belief that this battle is taking place across every industry and every vertical. And yeah, it's the trillion dollar question on who's gonna win and why.

Luke Sophinos (21:30.458)
I think we all can agree that my view is the most dangerous place to be is, is in a, obviously this is the verticals podcast. So people are going to knock me a little bit for saying this, but I do believe the most dangerous place to be is in point solution land, even if you're vertical or in, kind of broad horizontal app, like all these sales, sales enablement tool, like all this stuff is not where I'd want to be. I've seen so many VCs retreat into, into the vertical, you know, world just cause

How likely is it that these big companies are going to really kind of move into, you know, that space? It's, it's, it's interesting to think about it. It's a tough piece. All right. I want to just quickly show, visualize to the bloodbath in the public markets, because I think it's, I think it's helpful. And then I got two Nick's here. So Nick from Euclid, if you want to, if you want to kind of opine a little bit on this.

Nic (22:12.54)
Yeah.

Luke Sophinos (22:29.722)
So this is Thomas Tungas from Theory Ventures. He's a great blogger, been a great blogger for a long time. He visualized where is everything getting hammered. why are we, the three of us are disagreeing with the public markets in our previous statements. So like, what are they seeing that we're not? Like what's the steel man on this Nick? Why is vertical getting hit the hardest?

Nic (22:53.677)
you know, it's funny because I've, I've actually seen other breakdowns of this that should vertical is getting hit less hard. So I think there's like, there's honestly some questions of exactly. Yeah. There's some question of how he's breaking it up here, but, know, I think, I think there's also honestly some correlation with company size rather than necessarily whether it's vertical or horizontal. you know, on average.

Luke Sophinos (23:05.596)
It's moving every second. Yeah, for sure.

Nic (23:23.357)
Public vertical SaaS companies, and this is on average just because you have a lot of public vertical SaaS has been around for a really long time are smaller. generally mid caps and small caps get hit harder in downturns, right? So your big, big players, your big software players, which I don't know if he's including like Google, presumably he's including Microsoft, but certainly he's including Salesforce and Service now.

are big and they're just not gonna draw down as much. So honestly, I think there's some correlation versus causation here. I honestly think if you look at some that have been hit least hard, and you can look at like Moody's or some of these other like Fin serve players that...

have really strong vertical motes, they have actually been somewhat insulated from this downturn. I don't know, I'm trying to think what are, I don't know, where does Vivo fall on that chart, where does Pearl Car fall on that chart, and how much of that can we really attribute to their model versus just the size of the company? That's probably my take.

Luke Sophinos (24:44.518)
Tiffman, anything you'd add there.

Nick Tippmann (24:47.2)
Yeah, my God's preface by saying I'm no public market experts by any means. And I also have seen some conflicting data as well, Nick. But I think.

One takeaway or one thought with that conflicting data is that we may have just seen that vertical market solutions have a little higher variance than our others, which would maybe give credence to your point on they're just smaller mid-cap companies and having higher variance as well. My gut take, but this really isn't a macro take or would be trying to put a macro take on a micro or company specific is that some of these

vertical software solutions in the public markets are older at this point and are more legacy and are potentially more susceptible to an AI native upstart. And I think that even could have been a thesis four years ago, pre AI, just with the next wave of cloud solutions and even the first wave of cloud solutions becoming more legacy that it left room for a more modern

better user experience, faster, more aggressive company to come in and have some vulnerability in the legacy vertical SaaS in general. So I'm not sure if that's flowing through at all, but that's one other theory I'd throw out there.

Luke Sophinos (26:12.614)
So we'll take our public markets hats off and let's just talk about it from the customer lens for a second, because that's obviously where the value has to be created and that's the input and the output is the stock price and the charts and blah, blah. But from the customer lens, I've seen a lot of writing recently on, people even gonna use software tools anymore? Are agents gonna be using the software tools?

Like if you fast forward 10 years, you know, and I get the easy answer is, it's going to be vertical specific, but like, what's your guys' take on agent? You know, is there just an interface that we talk to and chat with? And then the agents go, go actually leverage like software tooling to take care of, you know, business workflows. Like where, where are we headed?

Nic (27:06.781)
Yeah, think I mean I can start. I kind of have two. Two thoughts here. I mean one is that. You know, I've always felt and I think I probably said this on the pod before. You know you you can't you can't fly a 747 through a chat box. You know, like human nature is such that like. I mean, I get it. It's very intuitive one text box. You know, like it's works well for.

you know, kind of context like an open evidence, right? But the reality is it's not the most efficient way to engage with every form of work in the world. And, you know, the same can be true of voice. mean, you know, we need visual ways to consume information. So I don't think UI is completely dead, nor do I think that honestly it makes sense that we're going to be in a world where all software is only this.

system of record decision data layer and I'm gonna use Claude or ChatGPT or Gemini to interface with all my software. just doesn't, it doesn't make sense. I already have trouble keeping track of like my projects in, you know, within those systems, right? So I'd say that's one caveat, but I do agree with the point I, know, other Nick, I think you made earlier, which is, you know, and we've written about a little bit too, which is that, you know, I think

the UI and the interface layer, if we want to call that, is becoming more fungible. And I kind of see two potential ways this could play out. One is that something I've called Inception software. like embedded in these solutions, you can have their own version of a wrapper or a lovable.

specifically to spin up UIs specific for every customer, for every user. There's no reason why the world has to look like Salesforce does where you get incredible feature creep and every button in the world because it has to be everything to everyone, right? I should just have what matters to me in my role at my company. And I think that's obvious. But there's no reason why platforms shouldn't facilitate custom experiences today.

Nic (29:26.749)
You know the second which I think is probably a little further out. I think of is bring your own UI You know like Today that's already somewhat happening on the margins like you know people creating their own dashboards based, you know from Whatever your addio API But you know if you just draw that forward there may be a world in which you don't as much have to worry about it I do think though. It's not just the data it is also

what we've been calling this decision layer, what some people in Silicon Valley have been calling the context graph. It's not just the data, it's not just the intelligence that comes from the labs, it's the context around how to make decisions for your industry and your company. I think that is like the meat that I've been most excited about. Nick, I don't know if...

Luke Sophinos (30:23.526)
That's interesting, like the permissioning, like the industry specific permissioning being a key driver of what the AI is responding to.

Nic (30:35.664)
Yeah, exactly. You're to get from LLM's general intelligence and I don't know, will they have certain areas where they go somewhat vertical? Maybe, but you need a layer that is ingesting first party data, is ingesting third party industry data that understands, if you're a GC, that understands what a takeoff is and what happens after that.

you're an owner of a medical clinic, right? What it means to be HIPAA compliant in your conversations and what it means to treat a patient, I don't know, post procedure, right? So I think this like context graph decision layer, like it's just the meat of what vertical SAS is today. It doesn't necessarily require the traditional UI.

Nick Tippmann (31:32.15)
Yeah, I definitely agree in that. Well said, Nick there. I think the UI will matter less. Things like authority, permissions, audit logs, deterministic workflows, rails to make sure those workflows end up being deterministic will matter more. I also was thinking about this talk and thinking about the concept of UI versus UX and where I think

Maybe the UI is mattering less, but the UI or the UX could matter more. Outcomes are mattering even more than how much time you're spending in the platform. used to be measured on how much of a worker's day is in the platform. Now a better unit of measurement might be how many tasks did you take off that person or the individual's jobs plate that day?

And it doesn't matter if that agents living in Slack through API is a headless platform embedded in solutions like you mentioned. But I think the winners will be thinking about things like audit logs and logging everything first, having human in the loop by default, but getting the ability to own more and more of the outcome over time.

you need to build in that predictability and the trust. And I think that's gonna be particularly important in the long tail of these industries that a lot of us are looking at.

Luke Sophinos (32:59.58)
So I got one more theoretical and then let's move into the tactical for listeners and kind of what's working within your guys's portfolios and playbooks. So last theoretical is on outcomes versus subscription and seats. Like I think my fundamental view right now is everybody should be experimenting with this right now is I think the best, like the perfect model in my eyes for most vertical companies is

Nic (33:00.123)
Yeah.

Luke Sophinos (33:29.752)
some low level subscription base mechanism with an outcomes piece on top of that. Like I really liked that model. I'm curious like how you guys view that. Obviously like the one side of the spectrum is like, it's only gonna be outcomes. And like the only way to make money is outcomes and all SaaS is dead. then the other side is, hey, we're just sticking with our subscription fees. That's a very hard question for probably a lot of our listeners to answer if they're locked in at.

10 bucks a user a month and they're like, if I go to five bucks and then price on outcomes, like, am I just gonna cannibalize my revenue? I'm curious kind of where you guys think the, where it's headed, where consensus may land. That's a tough one.

Nick Tippmann (34:19.18)
Yeah, I'm happy to jump in here first. I think it's a big debate and you're seeing a lot of experimentation going on between outcomes and seats. I think of it in many ways though as a continuation of pricing models that we've already seen in some ways. A platform for seats model is great. The platform gives you some base model and the seats is there to be a variable component.

And then maybe in the late 2010s, early 2020s, we saw transaction or usage based model, API keys, usage from places like Twilio, et cetera. And now you had this big debate on recurring versus reoccurring fees, which has kind of gone away in the age of AI. it's just everything is reoccurring at this point. But when you peel back the curtain, maybe not so much. So to wrap that, I don't know how much is changing.

Luke, I agree with you that you should have probably the best that I've seen today is some sort of base platform model with some sort of variable component on top of it. And then when we get into the variable component on top of it, I think that's where the big debate and a lot of the testing and hypotheses are going on right now around how much of the outcome that you can own and how much you can charge for that.

Luke Sophinos (35:39.783)
But sorry to interrupt you. was gonna say, by the way, think, I think founders too think that outcomes just simply means like a business outcome. Like we lifted your revenue. I think a lot of people are getting lost at like outcomes could just be the AI generated a PowerPoint for you. It could be like very micro outcomes. It doesn't necessarily need to be like this monster 12. It takes 12 months for us to see if there was an outcome. Like I think a lot of founders and CEOs panic over that. Cause that's what they attribute to outcomes, but it could just be like very micro.

things like we generated a PDF, right? Just one piece of color I

Nick Tippmann (36:14.796)
Yeah, really well said. Just to double click and to wrap up the thought here, because that's exactly where I was going in the outcome and maybe some more practical or actionable takeaways of tips. If you're ultimately like an AI, AE, and at the end of the day, you're selling a new closed one deal, like there are a ton of micro conversions that happen in there all the way from booking the first meeting to when they didn't show up for the discovery call, that you had to reschedule that meeting to.

have an AE to have a demo and there's maybe 10, 20, depending on how you break it down, 30 micro conversions in there. And so don't start with, I'm gonna sell you a new deal and every deal for you is 20K ACV and I want 10 % of that. So it'll be 2K per closed deal. Start with maybe what's the value of a new booked meeting for you or what's the value of a missed call? If we could answer one missed call for you and get one more meeting booked, what is that value?

So I totally agree with you that you should break down the outcome into smaller value units and kind of test from there. And then as you get confidence in your predictability and ability to deliver those outcomes, then you can take on more of those and capture more of the value that's associated with the outcome.

Luke Sophinos (37:29.072)
Yeah, well said.

Nic (37:31.198)
Yeah, I think there's always a perennial battle between two things here. One is more of like a basic human psychology thing, in which essentially the buyer, they want visibility and simplicity. They want to know what they're going to pay for something. They don't want to be surprised by some vendor costing them five times what they assumed they were going to pay. And I think

In experimenting with AI, we've probably all experienced that. It's like, oh, sweet, my anthropic bill is $400 this month. I guess I should turn that off, right? So on the same side, any startup, any investor in a startup would love to be able to, kind of like traditionally in SaaS, say like, OK, well, as long as our churn isn't crazy and we continue to do well in sales, we know what our renewals are going to be. And so we have visibility into how the business is going to perform.

And so, you know, like, I think there's that one psychological side that will always inherently push people back to having at least as you said, as you both said, like baseline, okay, you're gonna pay this and that's gonna, maybe it'll get you tokens, credits, it'll get you to a certain level. And then if there are overages or we over deliver on what we promise, then we will take a cut of that. So I think that is, I think I'm broadly in agreement.

that that is probably where it will net out. I would just add one thing in though, like.

If you are offering, as a startup, as an AI startup, if you are offering a service that does not have a strong moat or is not otherwise integrated into a platform that the company is reliant on, then ultimately your pricing power will get competed down to the marginal cost of delivering that service. If you are an undifferentiated AI service, it totally makes sense that you go to

Nic (39:36.166)
you go to a business owner and you say, you have this service now, this guy creates a PowerPoint deck for you and he charges you $10,000 a year to do this across your whole business. We're going to do that for a thousand dollars a year. Of course he's going to say yes, right? But eventually there's going to be 40 other players doing the exact same thing. And so they will compete away that margin and it will come down to whatever the input cost is. like, I think it's really important to think about like,

What, as you think about pricing, like what is your pricing power? How are you gonna build a margin over time? It's great to grow quickly early, but I think we're entering an era, and I'd love to like Nick, get your thoughts on this, where across a lot of verticals, we may have five, 10, I don't know, credible startups doing five million ARR offering the same, more or less, service, answering your calls at night, know, doing customer intake, scheduling.

What happens? You know, probably not they're all going to be winners So I think that defensibility question is important for pricing

Nick Tippmann (40:48.524)
Yeah, totally.

Nic (40:49.309)
What do you think about that, Nick? As you look at, because we face the same thing, as you look at a category where it's like, wow, this is a great team. It makes total sense. You're selling the thing. Growth looks good. But I've met seven others doing the same thing. How do you think about that?

Nick Tippmann (41:14.326)
Yeah, I mean, I think it's one of the big questions that comes up in nearly every opportunity that we're both probably looking at, right? For going after a vertical market, you definitely want to understand who are the legacy competitors, the incumbents in the space, but also a great risk and nearly any vertical worth going after is who are the other five to 10 great credible teams with great backgrounds also going after this same problem.

And I think part of this comes back and gives credence to the belief that like, I don't believe this, but it certainly was spread around a lot over the last year or two that speed is the only moat left. I think when you are competing in a specific vertical and there is somewhat of a land grab type of moment, then speed really is a determining factor that if you can go grab that land,

develop stickiness, have the audit logs, the compliance, the context of the organization. Well, then maybe that gets you running enough ahead of some of the other players in the space that you can create a big enough a gap. And I think where it comes back that I always believe brand has been important, but I think brand will be even more important over time, particularly with the enterprises where

trust and credibility will be a big part of the solution. So I don't know that there's a perfect answer out there, but things like network effects, brands that have been around, ecosystem lock-in, partnerships, trust, will all go a far way. And I also think that's why domain founder fit matters so much. And particularly with...

Luke, you've asked about playbooks and what's working. Like one of the playbooks that I've seen work well across our portfolio and across industries is just a community led content driven educational based playbook. And if you're able to make your founder CEO or someone on the executive team, the figurehead of that and really lean into that and build the trust and have IHEMBS and.

Nick Tippmann (43:32.718)
get the industry believing in you and believing in your vision so that they're not just buying what you have today, but they're buying your 12, 24, 36 roadmap as well. Yeah, there's some of the ways that I think about it. No clean answer again right now, but yeah, some of the different components.

Luke Sophinos (43:50.395)
I think one, it's a really good thread. Just, I think one thing I would add and let's just look at, let's just look at history, right? Cause history repeats itself is like, we had the same thing happen in the cloud era, right? You had, there was probably a hundred different toasts, right? And there was one winner that got whatever it is, 50, 60 % market share. So if we're in a similar moment in time, you know, and they were battling like hyper legacy, like on-prem stuff.

And so that's not the case today. But I think what it really came down to was go to market was such was so important. And like, you know, I think when you when you hear about like TOS early stuff, they were feet on the street. They were like geo focused. They were building relationships in like very targeted, like small geographies. I think around Boston is where they started. And you had a bunch of other people that were running like product led growth because that's what was the cool thing. And like, that's what

everyone said to do because the economics look better and, you know, in toast one. So I think go to market is actually, everyone wants a product led growth motion, right? Because it's typically the cheapest. But I don't know, maybe like go to market outside sales becomes a stronger mode in this era. It's interesting to think about.

Nic (45:14.609)
Yeah, mean, I look, I think I think distribution matters more than ever. And that's something that's that's unique in a lot of ways. Well, it's not it's not unique just to vertical software or vertical AI. But I think distribution can provide uniquely powerful advantages in vertical markets, right? Because they can be they can be harder to access. You know, they have different partnerships, ecosystems and, you know, can be more difficult to unseat.

So yeah, like you, like I think you said it well, Luke, like a lot of the breakouts we've seen so far, because, you know, we've got to remember, like, I don't know, we're all talking about cloud code, cloud code, wasn't a thing a year ago, right? Like these things are moving very quickly. The breakouts we've seen are, you know, a lot of these PLG, like open evidence style things. And it kind of makes sense because like, look.

Luke Sophinos (46:10.347)
That's only thing that can grow this fast.

Nic (46:12.823)
Exactly. That's the only thing that can possibly grow this fast. Now, that doesn't necessarily mean that in 10 years, that's going to be the best company versus ones that

Luke Sophinos (46:22.406)
Evolution IQ, how many evolution IQs are happening right now in the enterprise?

Nic (46:26.425)
Exactly. we, you know, for listeners out there, this is what two, three episodes ago, you know, we talked with Mike Saltzman, founder of Evolution IQ. you know, a couple years to a, you know, nearly billion dollar exit had zero revenue in their first, what year, 18 months.

Luke Sophinos (46:42.652)
Yes, 35K ARR in year two.

Nic (46:46.533)
Yeah, and that's because it was like, look, we're selling into insurance companies. We got to get this thing right. So they worked with one design partner for an entire year making almost nothing. And that's very different from like, hey, let's create this new thing, see, get it in as many hands as possible. Look, I personally think you could extend that to a lot of these companies. They're like, hey, look, I did zero to three in a year or whatever it is.

I mean, yeah, that's like absolutely, you know, I will not argue that that's incredible, but like, are you gonna keep growing? Are you gonna retain that revenue? Also important questions.

Luke Sophinos (47:29.734)
So this is a good segue. So let's, we've talked pretty theoretical. Let's get tactical. I want to try to zoom into some playbooks in what's working in this era that listeners can hopefully learn from and experiment with their own companies or with companies that are involved in. maybe Tipman, let's kick it over to you. Like, there any kind of playbook? And what I mean by playbooks is like a few actionable steps that you saw create.

You know, really interesting growth, really strong growth. Obviously Nick was very early in GCAI. So congratulations, that business is incredible. Like I'd love to, I'd love to hear from both of you guys. what, what are some interesting kind of new AI age, you know, steps and actions that, that you've seen really work in today's world?

Nick Tippmann (48:22.252)
Yeah, for sure. Happy to kick it off and get a little more tactical on the go-to-market front. I wish I had a great answer for you and that there was this silver bullet answer, but this is where maybe I'm the boring old operator that like this is just about disciplined execution and setting up the process the right way and running it properly. I think

What's more important than ever, and I always thought it was extremely important, was advice that I was given five, 10 years ago as a rev ops leader, just how important revenue operations and data are to your entire business operations. And now I'm also pulling in product and financial data into that. So what I see the best companies, they instrument their companies and their go-to-market engines early.

and they think about their go-to-market engine and rev ops just like a product. They have a backlog, they have priorities, they have, like we did at Greenlight Guru, had weekly demand gen meetings where the owner of inbound, outbound events, AEsourced, upsells, renewals, onboard, everybody came in and had to report their number on the week and...

give a red, yellow, green if it's on track, and if you're a yellow or a red, then what was the immediate action that you were taking? And you better go take that over the next five days. And when we are back here all next Tuesday, reporting on what action you took and what outcome you had on that result. Now, I think what the data, focusing on the data and the rev ops and the enablement early allows you to do is pretty cool. It allows you to be much more efficient and to get more

SQLs per SDR or more, a higher quota per per AE. It's making the whole ship in the machine more efficient. But I think the silver bullet is really not a silver bullet at all. It's just back to the practice and focus on the fundamentals and the discipline and instrument your company early so that you know what is working and what isn't working. So when something is working, you can press the gas and really push on it, whether that's

Nick Tippmann (50:36.974)
community-led marketing, founder brand on LinkedIn, doing educational webinars, partnering with industry associations. All of those could work or none of them could be working, but if you're not measuring and building the data back into your brain, then you won't know and you won't know where to push the press on the gas or where to cut.

Luke Sophinos (50:57.318)
So this, so, so I like Nick's, Nick's approach is like day at a time, right? Like I remember when we really started measuring channels and testing cha- and this changes by the way, like one, you could have a channel kicking ass and it stops kicking ass in 60 days or 90 days. And I always tell founders, I'm like, literally when you're just starting out, make a spreadsheet of a couple of channels, you know, test them rigorously and then whatever working.

go deep as deep as you can and hold on to it as long as you can, because at some point it's going to stop working. Just to push a little bit on you, how are your company the fastest growers in your portfolio? Are they product-led? Are they sales-led? It felt like you were kind of doing the, you're speaking to a combination of everything. Where are you seeing, at least broadly,

success coming from? Is it in PLG? we're just at the beginning of this AI era.

Nick Tippmann (51:59.608)
I'm happy to take, I definitely have to, I mean, just the physics of sales, PLG does lend itself to the fastest grower. That said, I think then the proof is in the pudding really in what is your second act and your ability to run the layer cake strategy, continue to grow ACV and actually be able to retain that revenue. But not...

Not every company that's growing quickly in the portfolio is PLG. We have some that are more traditional system of records. Actually one in the portfolio that you just had on your last podcast, Darren Finke, who just announced his round for Sante earlier today, had a really unique, interesting angle that he shared on your podcast.

around being able to rip and replace legacy tools and using AI to do that. So I, yeah, I thought that was really interesting. Yeah, hammering the phone, just old school hammering the phones and then having a unique way to rip and replace. And I've got one like more traditional events, like they're going to the random plumbing conference and the random HVAC conference and.

Luke Sophinos (53:00.252)
He's hammering the phones.

Nick Tippmann (53:18.03)
putting boots on the ground and shaking hands and getting out there with folks. I got a port code that's running a forward deployed model that's going much higher, six figure ACVs to start. And so they need to go in and deploy, but 80 % of what they're building is replicable across customers. And so they're able to grow quickly with these larger ACVs. So yeah, not all PLG, but PLG definitely does lend itself to fast growth. But in that scenario then really comes down to your second act.

Nic (53:49.798)
I've actually, I've found that, I think I would argue more broadly that like, PLG is generally less common in vertical. I'm not saying it can't be. mean, obviously you've got open evidence. There are certainly some fast growing legal AI businesses, but you know, in general, a lot of buyers, especially once you look below the enterprise because they,

don't think this is true of Harvey, do not care that it's AI. They want solutions to their problems. And in a lot of categories, some of these blue collar categories we're talking about, it could be a turnoff. And so I think what I've seen is more traditional selling cycles. And then once you start to become somewhat of like a known quantity in the space, then you can get more PLG. like that start.

has to be more so selling something that matters to the buyer in your vertical. And that's probably not selling AI. It's probably selling, hey, I'm going to bring back revenue. I mean, we talked about some of these, but like scheduling can work very well because it can recapture that lost 10, 20, 30 % of your revenue that was due to cancellations or inefficient scheduling, right?

voice AI intake can work very well because it can bring back revenue from loss leads because you just don't have enough people in the office or you've got to sleep, right? So those sort of like revenue connected, very discrete wedges we've seen work well. I don't know that that one's a surprise or a secret, but maybe like two other things I'd point out. One is,

Pretty consistently, the best growers in our portfolio sell way ahead of product early on. And I think it's just, they come to understand really what the problem is, and they find that problem that is so hair on fire that people are willing to make the bet even in advance of having the perfect product. In some cases, even in advance of the product, they're like, I just want someone to solve this so bad or to get rid of this old solution so bad.

Nic (56:10.682)
I will pay you to do it right now. And so, I think that comes down to honestly customer discovery and kind of having that vertical insider domain expertise. And then one final one, like maybe a really tactical one since we've been maybe more theoretical is kind of leveraging that understanding of the customer to...

give them something like almost like a takeaway that empowers them. I think this was, who was it? I'm trying to remember who is either in our portfolio or in a collective, you know, had this strategy in which they would basically do an analysis for the customer and present them. Like in this case, it was an IT oriented product and legal, and they would actually just give the legal team at the law firm like,

a complete breakdown of their current system, right? Kind of like just to demonstrate like, hey, we can already bring you value even though we're not customers. I think AI can be really well suited to do that. So I thought that was just like a cool tactical hack that I've been seeing.

Luke Sophinos (57:25.68)
Yeah. And back to Darren, just to brag on him a little bit, like he literally has on live demos, then download a file and he shows his system populated with their historical data set on a live demo. I was like, that is so genius. So, I mean, there's, there's a lot of creative approaches here. I love it. we're, coming up on, on time here. Let's, let's kind of shift a little bit into rapid fire.

if we can. I'm curious, Tipman, so this is like our little rapid fire segment, we'll throw five or six questions at you. But I'm curious kind of for your take, like, so first question is where are you seeing graduation metrics? And maybe it's, you know, from seed to series A. Is it like five million, is it 10 million? And then Nick, you can jump in with a few rapid fires on your side.

Nick Tippmann (58:22.126)
Oh man, this is a great question, highly debated topic and probably has the never fun answer of it depends. Particularly in the early stage and particularly today when you do have the possibility of these one to tens, 10 to 30, one to 100 in a year and those options are out there.

Luke Sophinos (58:30.447)
Yeah.

Nick Tippmann (58:46.374)
And VCs know they're out there. It makes for a tough middle ground for a lot of really great companies that are building solid, durable businesses, providing real value to their customers that just aren't finding the venture dollars that they once have. I saw...

The data analyst, Peter Walker over at CARTA, who always puts out tons of great data. I'm sure most of your linceners will be familiar with him and CARTA's data. He actually just repurposed some data. I think it was from Silicon Valley Bank earlier this week that showed the average ARR at pre-seed, seed, series A and series B. And the dispersion from the bottom quartile to the top quartile.

was unbelievably high. I thought it was pretty crazy, but if I remember properly, think the bar for seed is pretty low, which would say it's still a lot of bet on the team, bet on the market, bet on the size of opportunity around 220K ARR. And then for a series A, whereas traditionally, maybe a couple of years ago, you thought the bar was get to a million in ARR, that certainly has gone up on this data set, which I believe was from the last year.

2.5 million in ARR was the median for a series A to get done. But I think bottom quartile was something like as low as 750K and top quartile was like five or six. So it was vastly, vastly different. And obviously that data didn't take into effect growth rate as well. But I would say A's somewhere between one and three in general these days and seed somewhere.

250K to a million and pre-seed, pre-revenue to 250K.

Nic (01:00:39.834)
Yeah, no, I mean, I think that I think that squares. mean, it's kind of funny because in this environment, like because of the increased growth potential of these businesses, there's more pressure on Cedar series a funds to preempt. And so on one hand, you're hearing about how series a bars are higher, which they are at the same time. think you're seeing sometimes companies get these big rounds.

Basically, if you just think about call it time to $10 million raised, that's, I'd say, shortened in a way for high quality companies. it's a unique market right now. Nick, how do you think that the ideal founder profile, the characteristics you look for in the best founding teams has changed, if at all?

in the kind of post-LLM era.

Nick Tippmann (01:01:43.478)
Yeah, I think it has changed and rotated a bit in my perspective of placing an emphasis on learning quickly and having seen greatness before over domain expertise. I think the domain expertise is still important and needed on the founding team. And you should have some unique insight as to why you're starting the company, but I don't.

always believe that that domain expert needs to be the founder CEO. I think it's actually probably harder to learn how to build a high growth venture back scalable AI startup that needs to go one to 10 million in 18 months to be attractive to the venture market than it is for a 20 year insurance brokerage agent that knows insurance workflows up and down to take that knowledge and go and spin and learn all the.

the AI and high growth stuff. And then also with AI, it's just been the ultimate leveling of the playing field. so where expertise traditionally has been higher valued now that everybody more or less started as a beginner with AI three years ago or so, really started a lower playing field. And so that slope or that rate of learning for a founder and their ambition and having seen greatness before I've been.

place a greater emphasis on these days.

Nic (01:03:12.924)
What do you think in terms of technical talent? Have your views changed there at all in terms of either founding, you need the CTO to be in the seat to invest? Do you need the founding engineer? Or does it vary based on the team?

Nick Tippmann (01:03:35.72)
We definitely want to have a founding CTO on the team. That being said, I do think formal CS technical skills, maybe slightly less important in particularly the vertical AI and niche industries that we're focused on where not a lot of the innovation is particularly on the software that they're building per se. And then with the rise of all the coding and AI tools, the knowing what to build and having the taste of what good is like is

I think trumping the value of having a technical skill set. That being said, also just coming from a go-to-market and operator background, tend to resonate better with go-to-market leaders and go-to-market and business background than technical founders. So I think on a whole, it hasn't changed who we target or who we focus on. We have technical and non-technical founders, but I think on a whole,

the importance of that is going down some and the importance of being native and fluent in all of the AI tools is going up.

Luke Sophinos (01:04:43.836)
All right, I'll bring us home with the last rapid fire question here. What is your AI stack? What are you using? And what's the coolest thing that you've done with AI?

Nick Tippmann (01:04:58.338)
Right on. Look at my VC stack is probably pretty standard for many VCs. I'm using Affinity on the CRM for great things about Adio and others as well. I'm using Granola from a note taking standpoint. Then my most exciting workflow is then I'm using basically a system of action or engagement that sits on top of my CRM called Velvet FS.

that allows me to ingest all meeting notes, deal notes, emails, and company materials and documentation into its own folder container. And then I can query and have different agents on that of go run the TAM analysis, go do a competitor analysis, go look at news in this sector, go look at fundraising. so then once all that data is in one spot and it's all pretty much automated, I can then just go run these automated bots.

do 60, 70, 80 % of the work that I think traditionally an analyst would have been doing. And that's how I've been able to scale a bit as a one man solo GP fund.

Luke Sophinos (01:06:08.7)
Not a good time to be an analyst.

Nic (01:06:11.836)
Yeah, well conversation for another episode, but that's for sure Yeah, no, it's a cool time. I mean I we've we built a couple things internally, but I I have been have you guys seen this whole thing with like Claude Claude code Sorry, not Claude code Claude bot molt, but now called open claw I've been I've been I've been playing around with that and it's

Luke Sophinos (01:06:17.072)
You

Luke Sophinos (01:06:35.419)
Yeah.

Nic (01:06:41.532)
Pretty cool, I'm not gonna lie. I mean, it's like...

Luke Sophinos (01:06:43.564)
You see you're texting it right? I haven't tried it yet, but you text it and then it takes action

Nic (01:06:48.348)
You can set it up. Yeah, I mean like I have mine connected to telegram You can set it up to whatever channel you can put connected to your slack or whatever you want but basically yeah, I mean you're It just is more it's easily connectable to really any communication channel, but then you know on the back end it has access to Whatever you give it access to and yeah, exactly

Luke Sophinos (01:07:06.75)
yeah.

Nick Tippmann (01:07:09.528)
Did you go buy a Mac Mini or did you hook it up to your real rig?

Nic (01:07:14.332)
I didn't really see the need for that when you can run it on a 10 year old laptop just fine, you know, but.

Luke Sophinos (01:07:21.66)
Can it send like love poems and words of affection to my fiance? Because I always forget to that.

Nic (01:07:27.644)
It definitely could. It definitely could.

Luke Sophinos (01:07:32.368)
Would she think it was me?

Nic (01:07:34.865)
you know, if you gave it access to all of your past poems, I bet you could get pretty close, but...

Luke Sophinos (01:07:39.612)
I've got a huge database of prior poetry.

Nic (01:07:44.668)
Just make sure she doesn't watch this episode. You'll be good.

Luke Sophinos (01:07:48.336)
Yeah. She'd be sleeping if she watched it. I know that for sure. All right. Well, Nick Tippman, you're the man. Thanks for being here. Thanks for jamming with us on verticals. was a fun episode. So we appreciate it and kudos on everything you've done and are doing and excited for where you go next in verticals.

Nic (01:07:52.444)
Hahaha.

Nick Tippmann (01:08:11.886)
Thank you, Luke. Really, really appreciate you having me on. Had a ton of fun talking with you and Nick here today. Hopefully there's some good nuggets and insights for your audience here as well. But yeah, I've really enjoyed it and thanks for having me on.

Nic (01:08:12.208)
Yeah, man.

Luke Sophinos (01:08:24.354)
Awesome. All right, till next time.

Nic (01:08:25.361)
Thanks, Dak.

Nick Tippmann (01:08:26.542)
All right, great chance.

Nic (01:08:28.957)
See you man.