207. Why Domain Credibility Matters More Than Ever ft. Jake Saper

Jake is the general partner of Emergence Capital. Dust is dying.

[Jake Saper]
I think the pricing model is definitely dying. We’re now in a world where AI can do much of the work itself. And so selling a tool doesn’t make as much sense anymore.

[Josh Schachter]
So we don’t want to sell the fish pole.

[Jake Saper]
We want to sell the fish. AI native services, that’s going to be the winning term. What do I look for in an AINS company?

[Josh Schachter]
You’re listening to Unchurned, brought to you by the Gainsight Podcast Network. Subscribe to our sub stack at unchurned.gainsight.com, where we go deep on every episode, like how one post sales team at CloudBeds built over 150 AI agents. That story and more at unchurned.gainsight.com.

Hi, everybody. Welcome to this episode of Unchurned. I’m your host, Josh Schachter, Senior Vice President of Strategy and Go-To-Market Development.

And I’m here at the offices of Emergence Capital with Jake Saper. Jake is the general partner of Emergence. I’m super excited to have you on the program.

Thank you so much. I know you’re a very busy man, but thank you for being here, Jake. Stoked to be here.

You know, you’re like a celebrity. I mean, you are a celebrity. You’re a celebrity, but you are a celebrity within the halls of Gainsight.

And I will tell you why. Okay. We had our Pulse conference in Vegas in late May.

And there was a video where Chuck, you know, our CEO, who, you know, was going out there. He’s talking about AIMS as a big transformation for the company. And we did this promo video to kind of get things like, you know, amped up in the beginning.

And it had some cuts to you. Okay. And it said your name like five times.

Jake Saper, Jake Saper, Jake Saper. And with the whole marketing team, we love Jake, but we cannot have his name five times in a Pulse keynote video. So that’s kind of the going inside joke within Gainsight.

[Jake Saper]
I’m stoked that I had an appearance in Vegas. As we’ve discussed, I’m a musician. So like, this is my sneaky way to get some sort of fame in Vegas is through my AIMS work.

[Josh Schachter]
So I’m going to telegraph it. And this is how I get people to commit to things, right? So next year when you’re on the keynote stage with us in Vegas, you’re doing the Fireside Chat, the live Unchurned in Vegas.

We get Chuck on the other side. What’s going to be your walkout song?

[Jake Saper]
Well, I think that where I thought you were going to take this is like, will you perform? Oh, shit. I thought that’s what we’re going to do.

I will perform if Nick does it with me.

[Josh Schachter]
I mean, listen, Nick is still very much a part of Gainsight. I know. He’s the soul of Gainsight.

Yes. He also, you know, he’s trying to be sensitive to like, you know, not being there and making the Nick fest of these things. But if there was ever going to be a good excuse, Nick to be back at Pulse, it would be for that.

So thank you for bringing that. We’ll bring that to his attention. Bring it to Nick.

Okay. Tell us a little bit about what you do, about Emergence Capital, the focus that you guys have. Yeah.

Just to kind of set context for folks that, you know, maybe you’re not a celebrity. Sure. Yeah.

[Jake Saper]
So at Emergence, we aspire to be early experts in emerging business models. Hence the name Emergence. Firm’s 25 years old.

When the firm got started, the big technology shift was moving from on-prem to the cloud. And the emerging business model that was enabled by that shift was SaaS. When the firm got started, it was so early that that term wasn’t defined yet.

So if you look at our founding papers, it was described as technology-enabled business process outsourcing or Tebbo, which doesn’t have quite the same ring as SaaS. So we got the name wrong, but we got the idea right. So the first investment was in Salesforce.

And the idea was, how can we… Your first, your first VC, your company’s first VC was in Salesforce? 2003.

[Josh Schachter]
Oh my gosh. Talk about the right picks.

[Jake Saper]
Well, I mean, this is the idea of being an early expert. Like we had a belief that this concept of on-prem to the cloud was going to enable an entirely new way to deliver and capture value in technology. And Salesforce was the kind of early but emerging poster child of that.

So we made that investment. And that investment begat lots of other investments in horizontal software companies. We did Box.

We did Yammer. We did Zoom. Zoom was actually the first deal I led diligence on when I joined the firm in 2014.

And that early wave allowed us to be early experts in that emerging business model and capture our unfair portion of that from a venture perspective. The next big wave of technology that happened was the idea of verticalizing this idea. So instead of selling software to every industry, what if you sold a CRM just to a pharmaceutical space?

Which back in the late 2000s, that also kind of felt heretical, just the same way that in the early 2000s, selling software in the cloud felt heretical. Because it’s like a tiny TAM. It doesn’t make any sense.

Have you guys invested in like Pipedrive or? We invested in Viva, which is actually now the world’s largest vertical software company. The biggest thing to come directly out of Salesforce, right?

Indeed. Indeed. This is why being an early expert in a business model gives you unfair access to the subsequent business models that are often connected to the previous iteration.

So we invested in Viva. We’re actually the only venture investors in Viva. My partner, Gordon, is still the chairman of the company.

And that company has grown to become the largest vertical software company. Of course, we made a bunch of other investments in other industries as well. But Viva is the largest now across all industries.

If you fast forward to today, or really starting in 2023, obviously the underlying technology shift around AI is going to be much larger than even the on-prem to cloud. And as a result, it is engendering a bunch of new emerging business models. And so true to our DNA, we are striving to be early experts in these business models that are emerging.

And so that requires making some bets in the fog. We are in the foggy era right now, where it’s not yet clear. You’re literally there in the embarking era.

We are. And today it happens to be the weird day where it’s not foggy. But it actually is very appropriate that we live in San Francisco, to use that fog metaphor.

But if you think about, you know, in 2019, making SaaS investments, the hard thing there was picking the right team, less like, is this business model going to be the right business model? SaaS became, over time, became the consensus business model. We’re now in an era where SaaS is no longer the consensus business model.

And the question is like, what comes after SaaS? And what I and we are trying to do. It’s just dying.

So I think that it is very clickbaity to be like, yes, SaaS is dead. I don’t… I know, I’m trying to get it out of you.

I don’t think SaaS is dead. But is it dying? It is dying in…

I think the pricing model is definitely dying. The Percy pricing model is dying and will go away. But the enduring value that software providers play in certain industries, particularly regulated industries, I think will endure.

As an example, like, I don’t think that the pharmaceutical industries went to vibe code their software. In part, because there’s a lot of complexity. In part, because there are regulatory requirements where you have to have, you know, something that works in a certain way to be trusted with really important data.

[Josh Schachter]
But what about the companies that are large, you know, behemoths, the sales forces of the world, the gainsites of the world? Yeah. These are, you know, that are super course critical, mission critical, not necessarily in…

I mean, there’s regulated verticals that we all play in, but like, by nature, they’re not in regulated industries. What happens to those? Like the more classic thinking of SaaS.

[Jake Saper]
So the way I would think about those is their chances of survival are directly related to how well they’re able to shift what they deliver towards outcomes. And the zooming out here, the concept is like, we developed software, not because we as human beings ever had an innate desire or need for software. Software was always a means to an end.

So it’s just like the wheel, like we developed the wheel to literally get somewhere. We developed software to solve problems as a tool to help people solve a problem. That’s what Salesforce exists for.

That’s what gainsites historically has existed for. We’re now in a world where AI can do much of the work itself. And so selling a tool doesn’t make as much sense anymore.

So the way these companies survive… So we don’t want to sell the fish pole.

[Josh Schachter]
You want to sell the fish. We want to sell the fish. Exactly.

Thanks for telling us about the AI services playbook. And I’ve been told this is called AIMS for short.

[Jake Saper]
I call it AIMS just because it’s way faster to say than the mouthful of AI native services. To be clear, I named it AI native services a couple years ago. And I don’t yet know if that’s going to be the enduring phrase as historically we have not…

We’ve been good at calling the industry, but not good at naming it, given Tebbo and SAS. Do you like the name? I do.

I’m biased in favor of it. The reason I like it is, A, I think it kind of sounds good as AIMS, even though it takes a little while to get used to. The other reason I like it is I think it’s very descriptive.

So before we called it AIMS, people were using lots of different terms, but one of the popular ones was services as software, which I hate. And I hate it because it sounds like something, but doesn’t mean anything. Well, it’s very sloppy.

It’s sloppy. It’s like not precise. It’s kind of a mess.

And yes, it uses the SAS acronym. So it’s cute, but it’s kind of nonsense. AI native services is much more specific and descriptive.

Every word there is doing work. And it’s not just trying to fill a legacy acronym. Every word doing work is also sloppy.

But like in this case, well, I mean, as you’ll know, I have an em dash in the title, which is a little bit risky from a- No, that’s a single hyphen. That’s not an em dash. Well, it can be interpreted, like in some cases people do it with a two.

[Josh Schachter]
You’re right, I typed it with the one thing there. At some point, I’m guessing that you considered, tell me if I’m wrong, that you considered it being AI managed services. Am I not?

AIMS?

[Jake Saper]
Well, I am of the generation where AIM was a core part of my childhood. So I’m curious, what was your AIM screen name?

[Josh Schachter]
Oh, AOL Instant Messenger. Oh, ShaHockey. My last name is Shactor.

I’m an avid hockey fan, so it was ShaH. Are you a hockey player? Yes.

Are you? I grew up in Texas. But you went to Yale, and they had a tremendous hockey team.

I grew up in Texas. I’m a singer. Yeah.

[Jake Saper]
I played baseball, but I wasn’t as good at baseball as I was at singing. And I played both through high school, and then eventually the baseball coach came to me and was like, you have to quit singing because it overlapped with the baseball season. And I went to a Texas public high school.

He just assumed, like, of course I would quit singing. But I was like, I’m better at singing, and I kind of like it more.

[Josh Schachter]
And then you did acapella.

[Jake Saper]
I quit baseball, and I actually submitted a CD to college to get in. I mean, I don’t know what role that played. And then yeah, singing became like a defining part of my life.

I traveled the world doing it. Like the core thing, I now have three daughters, and like one of the key messages I try to teach them that comes from this experience is like, do the thing that gives you specifically joy and not what the society you’re in thinks is cool. Because also that changes over time.

Like society thought that baseball was cool when I went to a public high school in Texas. When I went to college, weirdly Yale’s this alternative universe is we’re saying we’re singing is cool. And so you have to follow your own North Star, and that’s what I’m hoping my girls do.

[Josh Schachter]
Okay, before we get back, oh, before you go. Why do I like this term? So this morning I woke up, went to the elevator at the hotel and got a message from WhatsApp that they are now allowing you to have usernames on your accounts.

Oh, that’s interesting. So yeah. So I was actually just thinking about this topic this morning.

So what was your thing? I was hoping we could get away from that.

[Jake Saper]
Because when I asked the question, I was like, okay, this is going to get back to me. Everybody’s going to guess your password. Yes.

No, no, it’s not my password. It’s just embarrassing to say out loud. So my AOL instant message name was fatcat600.

And the fatcat was P-H-A-T-C-A-T-600. Well, it was the late 90s. It had to be.

It had to be. For like all the Gen Z-ers listening to this, that was a common phrase back in the 90s that I wonder, given that we’re like reliving the 90s from a style perspective and a music perspective, like, are we going to bring P-H-A-T back into the mainstream? And maybe this podcast will be the thing that does it.

Maybe. It depends on if you repost it. So, okay.

Why do I like it? So the AI native services is doing a lot of work. So the word native is really important.

It’s not just saying, you know, trying to distinguish people who are kind of, you know, redoing this, you know, from a later perspective, but it’s actually distinguishing it from the AI roll-up strategy, which is actually where the majority of venture capital dollars have gone into in this. So there are a few firms that are doing a strategy where they will roll up a bunch of accounting firms or MSPs or what have you and try to join them together and then put AI into them. My and our belief is that the likelier way to create an enduring, you know, kind of iconic breakout company is to start from scratch, build an AI platform, and then go to market with that approach.

There is a third approach that I think is quite interesting and is relevant to this podcast, which is what Gainsight is doing. And Gainsight is the pioneer in doing this, which is moving from being a SaaS business to being an AI native service. Gainsight is the first major SaaS company to do this.

And Chuck and you all deserve a ton of props for moving in this direction. And you will be the company that the rest of the SaaS industry looks to to say, can I do this? And if so, how?

You guys, the Lord knows, you guys took some walks. So if we did walk and you were talking about this stuff, hate the scene. We’ll start conversing.

So Nick, so Chuck came to the office. I didn’t have a ton of context. Nick just said like, hey, you guys should meet.

I think you’ll really enjoy chatting. And we start walking. And I had no idea what Chuck was thinking in terms of what his plans were for Gainsight.

And as we’re walking down the Embarcadero and as we’re passing the seals on Pier 39, Chuck had asked me a lot of questions about AI native services. And then he kind of drops on me. That’s what Gainsight’s going to be.

I was like, what do you mean? Gainsight’s a software company. What are you talking about?

And he was like, no, I’m already putting the wheels in motion. Like we are going to do, we’re not just going to sell software to do customer renewals. We are going to do the renewal and we’ll get paid for doing the renewal.

I was like, wow, like what is really cool. And this is like part of why this whole like being early experts and emerging business model thing is cool is like when you go down this path, you never really know what you’re going to expect. So when I started working on this in 2023, it’s the first investment I made.

I didn’t have any conceptualization of like this was going to be an industry. I made an investment in a company that was an AIMS business. And as I was in there, I realized like, oh, what they are doing is selling the outcome thing is more than just one company.

This is going to be a broader trend. Which companies? It’s called Mechanical Orchard.

And they sell a service that does mainframe migration. So they work with very large companies who have mainframe software that was built in the nineties that when Fat Cat was used more regularly. And this company uses AI to shift those workloads into the cloud.

And I learned a lot from that experience and was like, this is bigger. So I wrote the Death of Deloitte piece in 2024. As you know, that’s when I got the cease and desist letter, which told me like, I should keep doing this.

I’ve hit a nerve. And then since then we wrote the playbook. We’ve invested in a lot of companies, et cetera.

But throughout that journey, I really wasn’t thinking what are the implications of this for SaaS businesses? And Chuck was really the one who kind of opened my eyes to say like, oh, maybe this is actually a post SaaS play for a SaaS company. So we went for that walk.

Chuck kind of shared this with me. And we’ve been in close conversation ever since talking about the how. And I’m gonna learn, I’m learning from you guys as you make this transition.

What’s simultaneously cool is I’ve had a number of conversations. So by the way, don’t tell others about it.

[Josh Schachter]
Yeah, exactly. It’s a secret. Let them think.

Can you just fail? Please take over all the time and we can release the word about it.

[Jake Saper]
Well, I think even if you talk about it, it’s so nuanced and the how of it is so hard that I think a lot of people are not gonna do it well, even if they try. So I’ll give you an example. I was in a board meeting a few months ago for a SaaS business.

And I said to the team like, hey, what if we think about taking the software we’ve been selling and using it ourselves to deliver the service that our customers were using the software to do? And the reaction I got was like, so you want us to hire more professional services people? And I was like, no, no, no.

Professional services in SaaS exist to deploy the software product. That is why PS exists on a SaaS P&L. What I am describing is doing the work itself, not selling or deploying the software for the customer.

It’s actually using the software to do the work. And so even in the companies that I work with, there’s a kind of a mindset shift because this is kind of like, it is really, I think for a lot of SaaS operators, there is a crazy kind of mindset shift that is required because what has been burned into SaaS executives for the past 20 years is services are evil, software is good. And so everyone has always tried to push that.

Because of margins and because of recurrence. Those are the two reasons why software has had such high gross margin or such high multiples over time. Obviously, the core underpinning of the AIM’s thesis is that AI will enable us to deliver services at a high gross margin for the first time.

And therefore it’s venture backable. Therefore, you know, this is an attractive direction to go. But historically that hasn’t been possible.

And so SaaS operators, you know, services is the forbidden word. And so the idea that your board member would come in and say, we should think about pivoting to become a services business is like a mind shock that I think requires a big transition. And even once you get your head around it, the how of it is really, really hard.

Like, how do you think about repotting yourself and, you know, changing what you are delivering to customers? In most cases, I think similar to what Gainsight’s doing, you’re not going to stop selling the software that you’re currently serving. It’ll likely be, you know, a new business unit that you open up.

[Josh Schachter]
And over time that starts to, you know, something could be the biggest headwind that a company like Gainsight would face with transforming into AI. I think it’s cultural.

[Jake Saper]
Like, I think if you’ve built a company over 10, 20 years that, you know, has been built to be something that builds and sells software, the idea that you can transform yourself to be someone that builds and sells a services platform, it just, like, it’s different in every way. The way you build these companies is different. The way you go to market is different.

Although the legacy players, ideally, have the existing customers they can cross sell into. But the way you deliver the product is different. Like, historically, obviously, like selling software, even if you’ve got an FDE team that’s out there deploying the software for you, is very different than selling some sort of outcome, taking responsibility for it.

It has huge implications on your CFO, right? Because if you move to an outcomes-based model versus a seat-based model, you’re moving away from predictability, which is scary. And there’s this, like, fear that the market’s going to punish you because, you know, your recurring revenue line is going down for this, like, kind of uncertain outcomes-based thing.

I think what’s kind of interesting about the position Gainsight’s in is that when you have a private equity sponsor, you may actually have a bit of a shield to pull this off. Now, most private equity firms are probably not going to be super stoked about this because most private equity is still thinking about, how do I optimize the business from a cashflow perspective to flip it? What I think is cool about the situation Gainsight is, obviously, the private equity overlords have blessed this approach and are saying, yeah, let’s put some revenue at risk by moving into outcomes-based because we think, ultimately, this will create more enduring value for the enterprise.

So, like, ironically, it’s possible that, like, being housed with a private equity firm could allow you to make this shift better. I’m skeptical that most private equity sponsors are going to understand and support this the way that yours is. But it’s possible, like, if Gainsight’s a screaming success, it’s possible this will be the example for other PE firms and potentially the public markets.

Person behind it, what is AI? So, an AI native services company or an AIMS company provides an existing service better, faster, and or cheaper than an incumbent by using AI to do it. So, you have to be doing it quantifiably, often exponentially, better, faster, or cheaper to disrupt that incumbent.

And you are moving the model towards outcomes-based pricing over time. Some AIMS businesses, some services businesses are already outcomes-based in the way they sell. Many are not.

Many are more labor-based where you’re just kind of paying for labor. Just the way that per-seat pricing doesn’t make sense in the AI era for SaaS, labor-based pricing doesn’t make a ton of sense in the AI era for AIMS. And so, these businesses have to teach their buyer to move towards an outcomes-based pricing model.

[Josh Schachter]
When you invest in companies that are practicing AIMS, do you have certain criteria for margins, profitability? Are you allowing this to grow over time? Yeah.

[Jake Saper]
When you say practicing AIMS, it makes it sound like a religion and makes me feel like a cult leader, which I’m like, I’m cool with, but like, it’s like, yeah. That’s a personality, Jake. Fair enough.

So, what do I look for in an AIMS company? So, one thing that I look for that’s a little different than I think what I looked for in the SaaS era is some sort of domain credibility. In software, since you’re selling a product, you can be anyone and the buyer may not care as long as the software they’re using is really awesome.

[Josh Schachter]
In AIMS, you’re selling- It’s like the Y Combinator. You can be a really smart 28-year-old and you can start a business at SaaS. 28-year-old, like 18-year-old.

28 is like, come on, you and I are both boomers at this.

[Jake Saper]
Yes. So, in AIMS, you’re selling yourself. You’re selling a service.

You’re going to someone saying, I am going to deliver the mainframe migration that you need. And so, you need a certain level of credibility. In the case of that company, Mechanical Orchard, the CEO is a guy named Rob Mee who started Pivotal Labs, which became the most successful technology consulting firm in the Valley.

He took it public. So, the guy has a lot of credibility around doing heavy tech consulting work. They had a super cool model.

I remember like a decade ago. Pivotal Labs was the thing. Yeah, the annoying thing about Pivotal is the first time I met them, I was on the board of this company called Drone Deploy, which provides drone software to construction workers, energy, etc.

It’s actually the very first board I joined 11 years ago. And a couple years in, they came to the board and said like, hey, we’ve got this mission-critical new product we want to build and we’re going to hire this group, Pivotal, to do it. And two things annoyed me about that.

First was the price. I was like, this is insane. Is it $150 bucks now or something?

Way more. And they did pair programming. So, they often had like multiple people.

And the idea was like, they’re teaching you while they’re doing it, etc. But it was very expensive. And the other thing is, I was like, why are we outsourcing our hardest product to someone else?

We should do that. And they’re like, no, these guys are the ninjas. These guys are the SWAT team.

They know how to do it. And they’re to take all the knowledge. So, they built a public company.

And Rob is super credible. Anyway, he has domain credibility. That’s part of the reason I funded the business.

But I think in general, I’m biased towards investing in AIMS businesses where the early team has domain credibility. It doesn’t necessarily have to come from the founders. It’s great if it does.

But if it doesn’t come from the founders, the best AIMS companies find a way to get early domain credible talent on the team so that the prospective customers can trust them.

[Josh Schachter]
Whether you’re Gainsight or whether you’re a new up and comer, you have to have some kind of a fiscal fiduciary responsibility. In the SaaS world, it’s what? 80% margins.

So, what is it for you as an investor in the AIMS world? Eventually, you want to get there at 80%. But you’re certainly not starting there today.

So, what’s the cutoff for you? Or what’s that boundary look like?

[Jake Saper]
So, I think about how to measure your progress towards becoming a successful, high-margin AIMS business with three layers of metrics. So, the first layer is bespoke to every company. But I basically call it like, what’s your North Star AIMS metric from a product perspective?

When you think about whatever service you deliver, if you were to break down renewals, for example, there’s a bunch of steps to delivering a successful renewal that historically have been done by humans. And if you were to break down that SOP, that standard operating procedure, into its most granular parts and figure out how can I build AI to do each of those parts, you can set a metric to say, historically, it’s taken us four hours to respond to this type of thing. But now with AI plus human, it’s taking us two and a half.

And next quarter, I actually want to set that to be two. And you can track those improvements time over time. And you can figure out it’s not always just time.

There’s other ways to measure those North Star product metrics. But basically, you create a dashboard and you track it quarter over quarter and you assign those KPIs to individuals in your company, which is a really important point. Because otherwise, if it’s just this kind of loose thing, these things, the metrics don’t actually get moved.

So, the first is like, you have an AIMS operating dashboard and you’re tracking it quarter over quarter and you can see the progress. That’s first layer. The second layer is what I call your annualized revenue per service FTE, which is a lot of mouthful concept.

But basically, the idea is how much leverage are you getting out of each service provider? So, each renewals person or each coder in the case of the mainframe migration one, or each accountant or each lawyer, depending upon what your service you’re providing is. And the idea is over time, you should be able to extract more revenue per service FTE, per person that you’re delivering.

So, like a CCO talking about CSMs and coverage ratio. Yeah, it’s not dissimilar. Like it’s a ratio that AIMS business need to be tracking.

And you as an investor, as a board member, as a CEO, want to see that leverage going up over time. And then lastly, is gross margin. But the idea is those two first layers are the building blocks to get to this terminally awesome gross margin.

I think when you’re doing early stage AIMS investing, when you’re moving to be in AIMS business, understanding getting to a high gross margin initially is going to be a bumpy journey by definition, because you’re going to do a bunch of initial investment in AI products. And that’s going to be really expensive. And over time, you’re going to figure out how do you optimize token spend so that you’re still getting a great outcome, but it’s cheaper.

And that should ultimately drive up. What do you want to see?

[Josh Schachter]
This is for all the founders out there that are spamming your inbox right now. I shouldn’t say that. That are trying to get in front of Jake Saper to pitch him an AIMS business.

Six months into their company, what do you want to see as their gross margin?

[Jake Saper]
So I can’t tell you a specific number six months in, because often the company is six months in. If they gave me a gross margin number, it would be wrong. And what I mean by that is how they calculate it.

It’s pretty messy. The way you calculate gross margin in any business, but certainly AIMS, you need to ensure gross margins is revenue minus COGS. So the COGS and AIMS business is the human labor that goes into delivering that service.

It is also the tokens that go into delivering that service. And it’s any sort of web services or other types of things that go into it, et cetera. But allocating the token spend and the human labor costs between COGS and R&D is a delicate but really important distinction because some of those tokens are going towards developing a broader platform.

Some of those are going towards just serving that specific customer, in which case it’s COGS. In the former case, it’s R&D. That’s probably too much accounting for you.

But as the business is built- That’s what you know as well now, right? The token spend stuff? For sure.

So there is a huge thing happening right now that I can do a whole other podcast on around open weight models. Open weight models are really, really important for AIMS. So open weight models are having a moment. There was a huge letter that we coauthored with NVIDIA and Microsoft.

My partner Gordon helped write it on Friday supporting the rise of open weight models in contrast to the closed weight models, which has been the dominant ecosystem so far. What open weight models do is it allows all businesses, but certainly AIMS businesses, to have tokens that are cheaper and have tokens and models that you can customize with your own data and obviously keep that data private because you can host that. You don’t need Anthropic or OpenAI or whatever to host it.

The big boon for AIMS businesses is if token spend is basically the labor for your business and open weight models can make that token spend exponentially cheaper than it was even a few months ago, you can deliver your service even more cheaply and you can customize the models you’re using to deliver that service with the data you’re gathering from performing that service. And this is like a sort of nerdy, but important point. That’s important.

That’s your defensibility. Well, I think the defensibility actually, yes, but I think the defensibility is in the data you gather from performing the service. So here’s like a hot take that I haven’t proven yet, but I think will be proven over time.

I think that ultimately AIMS businesses will be able to build more data-driven defensibility than SaaS businesses. And the reason I think that is because they own the full outcome. In a SaaS business, if you are just selling software, you often don’t get the full life cycle of like whatever that task is from the start of the task to like what is the business outcome?

Did the deal close? Did the, you know, ticket close? Whatever.

Sometimes you have that data, but often you have only portions of that data. But if you were responsible for the full outcome, you have all that data, which allows you to build an eval. And that eval set, which is ultimately what allows you to train these open source models to your use case, is the defensibility.

[Josh Schachter]
It’s my friend, Brett Fleener. Yeah, it’s great. Shout out to Brett.

Shout out to Brett. Talks about the judgment layer. This is similar.

You’re building that judgment. You’re building that.

[Jake Saper]
I think the challenge though is like, and I read Brett’s piece and that was really great. I think the challenge with the judgment layer thing in practice is that most founders today talk about this concept, but have failed to truly kind of operationalize it.

[Josh Schachter]
This is hard. Let’s talk about operationalizing it because I do have a question about that. Paint that scene as well.

Like what does human in a loop and reinforced learning operationally, what does that look like inside an AIMS, I was going to say service. I don’t know if that’s AIMS. Religion. Religion.

Yes, yes. This is the church of the big saber. Yeah.

Like what is that person doing? Are they offshore? Are they onshore?

Yeah. Are they hands on keyboard?

[Jake Saper]
I’ll give you a very tangible example. We’re invested in a company called Hanover Park. And what Hanover Park is, it’s an AI native service that does fund administration.

And what fund administration is, is basically accounting for private equity firms, private credit firms, venture capital firms. And the way we do accounting is kind of bespoke and wonky and super complicated, but really, really important to get right. Because ultimately we are financial institutions.

Historically, the service has been provided by big legacy services providers that are all humans. And they’re using legacy software to do it. So what that means is…

It makes a lot of money to do it. They make a ton of money. And it’s also slow and bad.

Because like if I want to do a capital call and I email my fund administration, fund capital is basically asking my LPs to fund an investment I’m doing. There’s a whole process of humans that go through, check Excel files, do merging for email files. It’s a whole complicated process.

If I want to distribute a stock and then distribute that to my LPs, it’s a long, complicated process. There’s a bunch… There’s thousands of other things you do with a fund admin.

The idea that you could do this instantaneously is like a mind-blowing thing for those of us that manage money for a living. And that is what Hanover Park is doing. Hanover Park built a…

First, they started by building a GL or a ledger that is very specific to this industry, which is really hard. They built the software layer and they built agents that specifically can plug into that software layer and do those tasks. And I think…

They come from this background, by the way, going back… It’s a really interesting question. No.

And this is like…

[Josh Schachter]
That just… That unvalidates, it disvalidates everything you just told us.

[Jake Saper]
No. So what I said was you want the early team to have domain expertise, not necessarily the founding team. It is great if it comes to the founding team.

Mechanical Orchard, it was there. In a lot of cases, it’s not. So what this company had to do was hire really senior, really credible people from the legacy service providers super early.

And that was the… I think a core part of their success because they could go to their first customers and say, we built this nifty software. Yes, I didn’t come from the industry, but all those people that used to be serving you, they actually work for me now.

So like… And they’re sitting on top of way better AI than they ever had in the last place. That being said, it is still critical for the CEO to go native, not to use the native term too much, if they are not from that domain.

So in the case of Chris, who’s the CEO of Hanover Park, when I first met him, I was skeptical. I was like, look, you… I mean, yes, you worked at Goldman Sachs.

You have like a finance background, but you don’t have a fund administration background. So why should I trust you to do this really wonky, complicated accounting stuff? So in our second meeting, I invited our fund CFO to join the meeting, which you never do in a pitch because it’s like, that’s back office, this front office, like whatever.

But I had Christy join and I had Christy ask Chris a bunch of super obscure accounting questions. Like, how do you book this for this type of thing? And I just wanted to see, like, has Chris gone deep enough to…

And they went so deep. They nerded out so deeply on not just like the accounting, but like the formatting in the spreadsheet for this specific… And I was like, wow.

And Chris later told me, he’s like, I interviewed 150 fund CFOs before I started this business because I needed to be able to have that depth of conversation because ultimately, I’m the face of the firm. Long way of saying, it is not absolutely… It’s not disqualifying if you don’t come from the industry yourself, but you need to hire people that do and you need to be able to go toe to toe with people from the industry.

[Josh Schachter]
I mean, the lesson there for me, again, like Boomer speaking here is don’t take shortcuts. Yeah.

[Jake Saper]
Chris, I think there’s also like a level of intensity you need to be a CEO in general right now. But I think specifically for an AIMS business, I will say this. I think, and this is like an interesting bit of frustration I’m feeling in AIMS right now.

I think a lot of people are rushing to start AIMS businesses right now, in part because people are talking about it and it’s exciting, part because Gainsight’s doing it perhaps, but in part because they think like, oh, I’m not that technical. This is a service thing. I should be able to do this.

And what they come to realize is that actually building an AIMS business is much harder than building a software business because you have to build a software platform and a professional services organization simultaneously, and the two have to integrate seamlessly. You have to build McKinsey and Stripe at the same time. It is so hard.

It requires so much intensity to pull off. And so the level of like no shortcuts is even higher than in software because you have to both serve customers. I started my career in consulting.

I know what it means to be a professional service writer. It is constant. It’s a barrage.

And also be building your AI platform simultaneously and be able to allocate resources between the two. The amount of sleep you get is basically none.

[Josh Schachter]
It still reminds me. You went to business school? It reminds me of a lot of business school students coming out and doing search fund.

[Jake Saper]
It is very similar. And with search funds, I actually think it’s a really well-suited thing for MBAs because like in that case, there’s not often like a technology transformation you are doing. You’re not building a software product.

You’re going in and improving the operations of some family-owned business in some location. This is not about that. This is not about trying to improve some legacy.

This is an AIMS business. You’re starting from scratch. You’re building a software platform and you’re building a services thing simultaneously, which is not to say that all the Stanford MBAs out there shouldn’t start AIMS businesses.

I would say if you’re going to do it, don’t skimp on the technology part of it and the technology co-founder.

[Josh Schachter]
You use the word barrage so I can only imagine you’re seeing a barrage of AIMS ideas. I get a lot of inbound, yeah. How much inbound do you get actually?

[Jake Saper]
It’s hard to quantify. LinkedIn is a strong source of inbound. And what’s interesting, I get more inbound on LinkedIn than in your email.

Yeah, I think it’s easier. Like my email is out there and people can email me, but I think it’s even easier. No one has ever accepted my connection request.

So I am not great about that because that is like, there’s a lot that come in every day and I only go through every day. I can go through it. I’ll do it after finish.

But yes, what’s interesting is like, while there is like a barrage and I do miss some great stuff and some great people, I had some amazing conversations since starting to speak more publicly about this that came about from kind of randomly. Like, for example, I started writing about this. I’ve known Nick for years, but I started writing about this publicly and Nick reached out to me and we had a few conversations about it and then he introduced me to Chuck.

And so like me thinking out loud about this has led to some interesting outcomes that I never could have forecast. I can imagine. Yeah, yeah.

[Josh Schachter]
That’s cool. Okay, so the barrage, back to the barrage. You’ve probably seen, there’s probably a little bit of Ains over ham, right?

Yeah. You’ve got a lot of ideas, a lot of people that want to disrupt accounting firm, whatever the case might be. Yeah.

What do you see? What are the, what’s the patterns that you see there of like, the overplayed Ains ideas right now and if you’re talking out there to founders, what’s one that you want to see?

[Jake Saper]
So I think that the biggest risk in AI native services is the concept of Mirage product market fit, which I talked about before, but the definition of Mirage product market fit is you’ve got revenue scaling quickly, you’ve got customers that are really happy, but the majority of the service is being delivered by humans. And that’s just a service. That’s not an AI native service and that’s a low margin business that should not take venture capital.

I’m seeing a lot of that right now. What’s interesting is like, an answer.

[Josh Schachter]
Going back again, going back to my margin question, it was a tough one to answer. Yeah. It’s not the most important question in the world, but if you’re looking at a pitch deck and they’re saying, okay, we’re starting this way, but we’re going to mature in 18 months, not six months, but 18 months is what our steady state, you know, margins are going to be.

What do you want to see that number at? I want to, what do I want to see the gross margin at? It’s not a services business that’s disguised as an Ains business.

They’re actually going to get to the margins. I think, I think the goal for an Ains business is to get to at least 70% gross margin.

[Jake Saper]
70%. I think that the reality is you’re going to see a lot of businesses in the 50 to 70% gross margin, which is still going to be a lot higher than the legacy service provider. Yeah.

The ultimate question for an Ains business, for the Ains category, is how will these be valued in the public markets, which I think will be a function of what their terminal gross margins are. And so obviously the higher gross margin and the higher multiple people are going to pay. Once you cross that 70% threshold, it looks like software.

And I think there’s a higher likelihood that these companies get valued like software companies, which is why that’s what I want to see. Back to your question, how do I sort this? A lot of early stage, like super early Ains businesses are getting really fast revenue growth, faster than software, which makes sense because if you are selling existing service faster, better, and or cheaper, a lot of people don’t want to try your thing.

So the question isn’t, if I can sell accounting services 50% cheaper, is there a market for that? Of course there’s a market for that. The question is, can I deliver that service in a high quality, high margin way?

[Josh Schachter]
Which ultimately means, can AI perform the majority of the service? And if it’s a 30% margin accounting, AI based business, then it’s a great bootstrapped company, right?

[Jake Saper]
Don’t take venture capital, or have a credible path to increasing those gross margins over time. What’s one that you really want to see? There’s a ton.

So, you know, something we were talking about yesterday is this concept of an MSP. MSPs are basically kind of firms that help other firms with their internal IT work. We use MSP here.

We’ve used a few over the years. There’s a lot you could do to provide that service, I think, in a more AI native way. There’s been some roll-ups that do this yet, but I’d like to meet more teams that are doing this from an AI native perspective.

I’m also pretty interested in categories that have some regulatory angle to them. And the reason that is, is I think AIMS businesses are often, much of the interesting AIMS businesses will ultimately have to have a human in the loop in a terminal state. And this gets back to something you alluded to before we started, which is, is AIMS just an intermediary phase where ultimately, you know, the humans are in the loop.

Yeah, it’s a human service, but ultimately just the AI is going to do it. I have two thoughts on that. The first is that I think the value of the human is ultimately in guaranteeing and warrantying the outcome.

It’s a throat to choke, a literal throat to choke. Like ultimately, if you think of it, if you abstract away, let’s for a second, let’s just say that code is a commodity. Everyone has the same piece of code.

And so what is different from the thing I’m selling versus the thing you’re selling? If you’d be at a service or a software, ultimately, like the difference between, you know, bootstrapping yourself with Anthropic or AI or whatever, versus hiring a third-party services firm to deliver it or third-party software firm is that in the third party, you have someone who’s responsibility for updating the thing. And you have someone whose responsibility is like to fix it when it breaks.

And so I think that there’s a kind of a wild future where technology vendors look less like I sell you code or I sell you some service backed by code. Instead, it could look more like I am an insurance carrier and I am guaranteeing financially some outcome you’re hiring me to perform. Like think about Gainside.

I don’t know exactly how you guys are pricing the renewals product, but there’s a world where you move to a model where we guarantee a certain renewal percentage and we get paid, you know, accordingly when it flexes up or down. If we can beat our guarantee, you pay us more. If we lose it, you know, we get less, which is effectively an insurance product.

That’s a kind of extreme way to rethink about and move towards outcomes-based pricing. But I think that’s going to exist. And back to like why they don’t think this is a passing fad.

I think that having a human in the loop to guarantee that outcome is going to be really important. The second point I’d make is there are certain businesses and industries where you are going to always be required to have a human in the loop because there’s some sort of regulatory requirement. So a niche example, insurance is a good example.

Another niche example would be customs brokerage. So every time we import stuff, the tariff regime is obviously changing a lot and constantly changing. And there are licensed professionals have to go through a really intense licensing process whose job it is to assign the tariff to every item that comes into the country, including all the subcomponents of that computer, et cetera.

That is something that obviously AI should be very well suited to do, but you’re required to have this licensed person kind of in the loop. And I think that was kind of an interesting, you know, enduring angle.

[Josh Schachter]
My mind is flashing to like, oh, that’s interesting. If agents are behind the scenes in customs duties, then who’s going to be paying off that? Yeah, there’s less bribery.

Yeah, exactly. Another alignment problem.

[Jake Saper]
There is a really interesting like bribery angle in like weirdly in a lot of these services businesses where there’s like the mafia is involved inside. Like the mafia is actually still a part of our economy in a way that I think we don’t expect. But like I’m involved with some construction businesses that have some AI angles that allow.

Where you back mafia based? I back companies that sell to construction. Mafia also sells to construction.

So mafia is often like a competitor, weirdly to some of the companies, meaning like if like the mafia operates dump trucks and they charge per dump truck that gets emptied and you now have AI to contract the actual number of dump trucks to get emptied, they can’t fudge the numbers as much. And the accountability. So yes, there are really weird things you never expect when you dive into the AI world and the implications in the real world.

Jake Saper.

[Josh Schachter]
Emergence Capital who coined the term AI BPO.

[Jake Saper]
No, no, no, no. It’s not AI BPO. AI Native Services.

That’s going to be the winning term.

[Josh Schachter]
I love it, dude. I love it. We have one more thing.

I need a Sharpie. I need Jake to sign this book.

[Jake Saper]
There we go. I’m so scared of Sharpies because of my daughters. Oh, yeah.

All over the walls. Yes. Will you remind me what you are?

My name, I’m Josh. No, no, no. What is your AOL screen name?

[Josh Schachter]
Oh, Shah Hockey. S-C-H-A Hockey. Shah Hockey.

Keep Ainsing. We will keep Ainsing. Thank you, man.

This was such a pleasure. You’re welcome. That was fun.