Sila Health raises $27M to automate healthcare prior authorization with AI — Pavle Asparouhov and Delian appear together

uh I think Delian's hopping in as well. No way. That's great. The big brother to pump the little bro's bags and tell him to clean his camera. Yeah, clean up the camera. Got it. Good. But you got the both of Sparrow hops. This is the first time we're ever making an appearance on anything together, so you know. Fantastic.

Not the last. Welcome to the show, guys. Uh, break it down. What's happening? What's the company? How much you raise? What are you doing with the money?

So, the company basically the root of it is, uh, health care providers in the United States spend a ton of time just dealing with the paperwork from insurance companies. And it really both distracts from patient care, so clinicians are just able to serve less patients.

And when something goes wrong clerically, it can actually lead to like a delay in patient care. So depending on like some of our customers like our autism therapists and it's like that delay in care for some of those kids can really lead to adverse outcomes.

So what we built is basically AI that merges basically like moves data from point A to point B because fundamentally it's just it as much as you know maybe other curly hair technologists like to say it's it's it's not it's not as much of a systemic issue as it is a uh you know technological issue and the we're just really like insurers set up reasonable processes and have an appropriate check and balance in this workflow where you as an employer don't want to be paying like unreasonable premiums.

You as a patient don't want to be getting care that's not medically necessary. But the problem is is like there's a competitive space between all these insurers. There's like 10 different insurers that a individual clinician might work with. So there's 10 different processes they have to manage.

So it's really about plugging these two parties, making them just like work together better. Um, we sort of raised $27 million across our seed and series A, 5 million C, 22 million series A, and where we're taking that capital is we really just want to double down on our core product set.

It's basically like expanding into more specialties and really we want to go a layer deeper into the extent that we help clinical staff today. we really just help with like the clerical parts of things.

But as we sort of sit between these two parties, we really start to understand insurer guidelines like we understand why often might reject something and we just make sure that the insurers are able to get clean streamlined data that is past all the basic validation.

So that's where we're bringing the capital in hiring engineers sales operations is really trying to scale the machine and expand on our core product set. So Bane Capital did the deal. Did the deal get done at Fogo de Chiao? You got to let me know. Did the deal Oh, Delian. Oh, that was a Dian question. Go ahead.

No, no, that's for you cuz I saw you posting about it like two years ago. To cow. No, the uh you know, I it it was part of the close more so than anything else. That's what put you over the over the finish line. That's great. Yeah, right at the finish line. No deal.

Uh we were fortunate enough where uh this actually this deal didn't hit the market. Uh we were for able to close it from our New York office. So fantastic. Fantastic.

Say which, you know, probably kind of covered, but I think for the broader audience, you know, that are maybe not super familiar with healthcare, the way the providers have gotten paid has just kind of fundamentally changed over the past like 10 years where basically 10 years ago provider provided care and then went to insurance companies and said, "Hey, here's basically, you know, sort of what I did.

I'd like to get paid, etc. " And sometimes the insurance companies would look and be like, "Whoa, this is like really out of whack.

" like this is not what I wanted to pay for and so you have these kind of like you know sort of misaligned incentives and so polo companies focused on this like new workflow on basically prior authorization.

So it's basically like now you have to go to the insurance company first but because this sort of like new workflow and it slows down care right because before you just go to the doctor they would give you the care and then you you deal payments.

Now the insurance company is set up in the pre-work but then because that if that takes too long now all of a sudden you're like slowing down the patient experience which is also directionally like good. You don't want to be hit with like an unexpected medical bill, right?

Like that's a lot of the times when you hear some of these like unexpected medical bills happening, it's because you don't want the check happening after the fact. You do want it happening up front, but like with our technology, you're able to do that upfront check without necessarily delaying the patient getting care.

Yeah. If it takes like 30 days and this person needs like some cancer treatment, right? Like you want to be able to like, you know, get this stuff like approved quickly. And so there have been like with the roll out of this new process, there's definitely been some patient backlash being like, "What the hell?

Like the doctor gave me my care plan. I would like to proceed. " and I'm like waiting on my insurer to approve my ability to even go get, you know, care in the first place, you know, before the bill. And so it being fast is like super critical.

The other thing that I wanted to mention is, you know, so Pavle co-founded with these two guys, Sagar, who was previously at True Work, but then the CEO is this guy Jeff Morelli. It was actually my high school buddy that, you know, has known Pavle since he was, you know, sort of 14 years old.

Um, you know, the, you know, sort of first time they met, I think, was like up on the ski slopes in, you know, sort of Utah where I peer pressure smoking some weed for the first time when he was way too young to smoking weed.

Um but allegedly allegedly allegedly allegedly allegedly uh you know officially according to regulators Pavl has never smoked marijuana. Um, but you know, I think that provided a good bonding moment.

And then Jeff actually came and worked with me on my first company, which was not in the prior authorization space or like financial services, but it was actually in the healthcare space. That's right. Software for autism therapist.

And so he this pre-existing understanding of like the field, you know, autism therapists, what their workflows were and we were more focused on like clinical workflow software. But it all kind of ties together of like, you know, Jeff and I used to work together and, you know, had been friends with high school.

He had known Pavle since he was, you know, sort of 14.

Paul crushed it at ramp and like you know being in the first 25 employees and helping build out their like you know sort of build pay product which has been phenomenally successful for them and so actually both were in like this like co-founder dating process like whatever it was like you know sort of two years ago and I was like guys like you guys should 100% consider you know sort of you know working together given the overlap of interest and so in some ways what they're working on today is like the perfect marrying of those two backgrounds of like you know financial you know sort of services and what Pavl has done and then Jeff's background in like you know sort of go to market in healthcare marry the two prior is like this crazy background trend and you know the Asperero magic can you know make space factories and it can you know uh you know make patients lives and clinicians lives way easier.

Uh part two hopefully Sila ends up way better than Night and Gale ever was. I love it. Uh you got talk about uh so we had Lulu on yesterday. She was talking about the golden ratio ship to YAP.

I think it was a had to have been a pretty uh intentional decision to wait and announce the company and two separate financings at the same time. Uh Pavle, you're a generational poster, you know, uh you know, potentially poster of the year if you really get back into the game. Yeah.

Uh maybe talk about that decision to not, you know, be posting 10 times a day about the company when there's probably an argument that maybe that would have helped in different ways, but clearly you made an intentional decision to just keep quiet and and focus on building.

Yeah, I think if it had meaningfully blocked the business anyway to not be talking about it, then we absolutely would have launched. But I was just like, you know, early hiring is predominantly a network and doctors aren't on Twitter.

So, it wasn't necessarily that we weren't like talking about what we were doing, but it was we weren't talking about it in front of like a tech audience. I think also just there's a degree of focus that can come from downstream of not being in the public eye. I think it's limiting in a lot of capacities.

I think as we're scaling that was sort of the decision now and I think part of it is just you know I think we had sort of an opportunity when we closed the A earlier than we anticipated to really just come out and like you know I'd say like a bunch of things get lost in especially in like AI hype cycles a lot of things get lost in the noise where it's like even like healthcare AI it's like is this even real what's going on here but it was I think it was really compelling to us to be able to come out of the door and be like hey we operate in like 45 plus states.

We work with hundreds of insurers. We have tens of thousands of patients. It felt like that was more powerful. And I think it's sort of like I I think it's really important to me that like I work on something very tangible with like less I I I didn't want I don't want to be a hype guy.

I'll hype my stuff up, but I want to hype it off the back of something like very real. And I think we're coming today with very real business progress to show the world. Uh, has your hiring criteria changed? You're on record saying that you have two hiring criteria.

Uh, you need to have that dog in you and you need to be nice with it. How has that evolved recently? It's a perfect framework. I don't know why you why why why would Can you unpack it a little bit for us? Break it down. What does it mean to have that dog? Sort of. We we we sort of sit someone down.

We we we just dig through every life decision they've ever made. And then after that, we we sit all in a room and we go through the two avenues of like, okay, let's all start. Everyone go around. Do you think they have the dog in them?

What's the what's what did we learn about them that show that they have the dog in them? And then um are they nice with it? That's just, you know, we sit down, you know, we hire everyone that they're supposed to be nice with some thing.

We don't hire like a ton of like, you know, we don't have a lot of generalist strategists around. It's like very we have specific skill sets.

And so just making sure the process sort of shows like I think it goes back to like that tangibility is like don't get me wrong we do have generalist work to be done but we always hire someone like on the back of some specific tangible skill set that we can track for. That's great. That makes sense.

Talk about the moment you were working on integrating GPT3 into like what you were working on at RAMP was like was that like like mind-blowing at the time?

Did you feel like you had discovered something that uh you know discovered a broader opportunity and you just said I need to dedicate my life to creating business efficiency using artificial intelligence? I guess just like talk about that talk about that moment. I I think it was a few things.

I think like it was Jeff my co-founder his just like family is on like his mom's a nurse, his uncle's the CO of a hospital system. So just had this like really native understanding of what the problem space was I think at the time. We were basically deploying LLMs in the context of like pulling out vendor contract data.

And I think what I found in my time ramp is like top tier company, best of the best. I I I think when I was sort of looking at my domain and like just the area of fintech, if you go if you go talk to like a finance person about their like technology stack, they're pretty happy with it.

And and so part of what it was I wanted to be able to operate in a space where the technology we were delivering was truly like revolutionary in nature. Something like 10x better not 10% better. So I think healthcare was this like awesome moment where it was like hey healthcare data is fundamentally textural in nature.

Like you can't tell me what's wrong with your knee with a bunch of like codes. You need to just like verbally explain to me what's wrong with your knee. And so the sort of combination of all these things is like actually LLMs are basically able to unlock that 10x experience.

And there's the funny thing that happened is that like the last big techn technological wave like web apps which is sort of like web apps and like SAS tools and APIs which is was what a lot of the successful successful startups of like the late 2010s were.

Um it didn't hit healthcare in the same way because frankly like a web app isn't like for a physical therapist a web app isn't like 10x better than like a filing cabinet because you're already in person.

and there's already these physical like notes and what you're finding with LLM is you're able to really drive like both technological waves in so you're able to do like we're delivering a ton of value with the same sort of technical skills that I learned in developer ramp on top of just like this AI blend is really able like you can see some of the quotes on the silahalth.

com of like we've just become like immediately this art artery and like game changer for a lot of our customers which for me is just like really engaging gets me excited to get up and work on the problem every Okay.

One of the frameworks for looking at these startup ideas is find a industry that's highly fragmented and low NPS. Can you talk about the previous structure of the industry uh prior to you launching? Yeah. Well, what I'll say is that's a great structure if you're an analytical guy and not a vibes guy.

So, we just like kind of like the problem and people would talk to us for a long time about it. So, that's how we land on it. Yeah, totally. It is like a real uh it is a real fragmented market where frankly there was not meaningful solutions on the market that could solve this in a holistic way.

There were little tools here and there that you basically had to in this the fragmentation maybe is a little different in that the tools were all fragmented and you would still have to basically employ someone to go through and use those tools and run the workflows.

Where we operate is we're just like an end toend solution. You plug us in there's no tangential tools. your able to get it.

So like patient comes in, you give us their insurance card, we tell you exactly what clinical documentation we need and tell you like, hey, you have patient cleared for care, get them scheduled without having anyone on their end necessarily there to manage the process.

So I think that's how we think about a lot of the consolidation that we're able to do. Do you have any advice for founders out there that are seeing your announcement today, all the money you've raised and and want to copy? Go ahead. post more on Twitter.

Generic advice is that like not generic advice, my advice if I had to give the world is uh uh I think Keith has said it, but I'll sort of reiterate as someone that's lived it is like work at a great company before starting your own.

I think that has been like massively helpful from like basically like inception we had like seed funding. It allowed us just to take a longer time horizon how you're thinking about it. It just really shifts the model when you're talking to candidates. You're able I'm able to point like hey here's the track record.

And I think that's something I reflect on a lot is like that Keith advice. Go work at a ramp was a great company. Had an opportunity to work there and I think it's just made the first like year and a half of entrepreneurship a lot easier than it would have been otherwise.

Well, hopefully you provide that for the younger versions of you out there. We got we got some young we got new ples. Fantastic. Uh last last question for you. Uh is there a certain milestone that you want to hit before you cut your hair? Is it 100 million ARR? Is it 500? Is it a billion?

You just It might just be the look forever. Growing it down to your waist and beyond. I think it aligns like look it aligns with whenever I can start tweeting whatever I want. Okay, there we go. There we go. I want to see the hardcore buzzcut P.

That's going to be It's a It's primarily like you want something pretty quickly when they see me be like, "Ah, that guy that guy works with computers. " Yeah, everything's computer now. Uh it's great it's great having you on. You're our new healthcare expert. Fantastic. We'll have you back soon. Thank you for having me.

Congratulations. No, it's been it's been uh I remember I remember I came over to Will's office. It was just you sitting there by yourself. Uh you told me roughly what you were working on. It's fantastic to see all the progress and I'm Yeah. Thank you. Thank you guys.

As Joe As Rogan would say, it's an honor to uh you know cover uh your fundraising announcement. Yeah, exactly. Thanks for having me on, guys. Really appreciate. Cheers. Bye. See, and see you, Delian. Later. Couldn't get a word in edgewise, but it's so funny to be like, "Yeah, little bro.

I'm coming on your fundraising announcement. I'm going live. " I love it. That That's the nature of the call-in show. He's just in the text message group. Can I hop in just to hype up Pavle? Pavle find.

No, I mean I can't I can't uh Pavvel's smart enough to know that he should leverage every single advantage that you have in life in having a big brother like Delian who can be your hype man who can help you avoid pitfalls, help you make sure you're actually focused on help you avoid bad investors, etc.

I mean, there's another thing that that he's really good at leveraging and that's eight nights that fuel your best days. Turn any bed into the ultimate sleeping experience. Go to t go to eightsleep. com/tbpnb $350 off your pod. Uh that fuels your day. I saw Andrew posted a hundred sleep score. Oh no. Devastating.

I don't think I'm anywhere near there. I woke up at like 4 a. m. unintentionally. I actually I actually did put up a 100 last night. You did? Which is crazy. Oh, you prepped this, didn't you? Okay, so uh we have somebody joining the show. We have Exactly. And don't don't say any names.

So this is somebody named carried no interest. They are a prolific poster on X. You're going to hear their voice. You won't see their face. Quite the following. And uh he's going to come on and talk about the the situation with all these AI SDRs, all these companies saying they're going to automate outbound.

And then very exciting, he's going to dox himself in like two weeks live on the show. Live on the show. So I'm very excited for that. But let's bring him in. Space reveal. Let's bring him in. He's coming down. Well, in the meantime, we can talk about Wander. Find your happy place. Find your happy place.

Book a wander with inspiring views, hotel grade amenities, dreamy beds, top tier cleaning, and 247 concier service. It's a vacation home, but better, folks. Go to wander. com. Use code TBN. However good you think it is. Could be better. It's better. Could be better.

We've taken you on a whirlwind tour of the world today. We've taken you to China. We've got East Coast. We got carried. No better place than wander. Welcome to the show, Carrie. Great picture. I've arrived. By the way, my hair is wonderful right now, too. I want to show it off so bad.

But you sound I mean, are you doing a Danny DeVito impression right now? You sound kind of like him. You like that? No. Just a lot of just a lot of smokes, boys. That's all. Uh I've almost said your name 20 times already. Uh but uh but anyways uh for everybody this is this is no interest.

He is a private equity investor skewing towards software former head of AI at a billion dollar PE fund uh and doing a bunch of interesting stuff in SAS and uh looking at businesses as well that are using AI to sort of reinvent themselves. But uh you were chirping on the timeline earlier today about 11X.

Wanted to just have you on. I I don't want to pile on 11x too hard. Uh but I do want to talk about the broader sort of like sales automation space. Uh what's going on with AI SDRs? So here's my hot take, right? Like some some software companies will be AI first and succeed massively.

I think that the notion of the AI SDR is flawed permanently, whether it's 11X, whether it's some of the ones from Y Combinator. And there's a few different reasons. Like the first reason to me is that cold email is functionally like an alpha game.

You are constantly trying to take advantage of a bunch of different quirks about cold email so that you can send 5 to 10,000 a day. And when you abstract all of these quirks away, I know there's a bunch of cold email wizards that will agree with me.

You actually lose some of that alpha when you just say, "Hey, here's this tech company. I'm going to let you run large scale cold email. " Very tricky, right? Very hard. That's the first piece that I just don't love.

The other piece is like when you think about when you think about it to me there's five reasons that make it tricky. The first reason with these AISDRs, you're relying on someone else's cold email data. You can guess what happens to your gross margin when that occurs, right?

So, if you're an AISDR company, you have this big database of emails, you're mooching off somebody, you're going to take a hit on gross margin just there, right away, right? The second issue is that in my opinion, you have a high probability of potentially being embarrassed. So, and embarrassing the end customer, right?

Both, right? You have this AI that is sending out emails. Who knows who it targets? You know, that's a tricky situation, right? When you think about the AI utility relative to the magnitude of mistake, Kugan's law, I have to find a way to coin this. I really do. Let's do it.

When you think about that ratio, let's think about uh cursor really quickly, right? You get all of this instant efficiency from cursor. You get more code written. If you know anything about developer environments, if you push bad code, it should get caught in testing.

The magnitude of the mistake relative to the efficiency gain of the AI is very low. You get all this output. The mistake is caught before it goes to prod. Win-win for everybody, right? On the AISDR front, no such thing, right? Mhm. We're doing it live. Yeah, we talked about this. It's a good point.

Like there's an opportunity for a high degree of embarrassment. Y and I think that that that is actually leading to a lot of churn. So I've seen like behind the scenes on a few of these businesses, the churn does not resemble like top enterprise software companies. It simply doesn't. Well, yeah.

And one of the issues is you wouldn't want to use an AISDR on a very important account because let's say you're trying to close an account that could be worth $2 million a year. Well, is it worth any potential embarrassment to just spam them with a bunch of emails where it's like, "Hey, I saw you live in New York.

Have you checked out Central Park? You should see a Broadway show. " Right? Like it's like it's just not worth it. It's like, "Hey, this could be, you know, hugely creative. " Or at least keep a human in the loop. Yeah. Yeah. And I don't think it's an 11x problem by the way, right?

I think it's a entire notion of the idea problem. Here's another tricky part. Let's talk about and I you us three have talked about this privately but like you only grow to $100 million in ARR the way Cursor did by being self-s serve. Mhm. Right.

The and and some of these AI first companies are incredibly self-s served meaning your CAC is extremely low. Your onboarding time is very low. You can grow really quickly. That's a double double-edged sword for a bunch of different reasons, right? But now let's think about the AISDR.

Not only do you have this high potential for mistake, the onboarding is a whole thing. You need like a full customer success team. You need constant check-ins. The amount of opex you're dedicating to maintaining an existing customer with a bad churn rate, that is not the profile of an excellent software business.

Does that make sense? Yeah. Yeah.

I mean overall I feel like um the the risk to what AI is doing to these high growth startups is that we're seeing conversion rates go up a ars skyrocket but churn rate also goes up and that's the real uh underlying question here because we have so few months of churn data on this new generation of companies that look very different and have different switching costs and different installation costs and and there's a new hot model every two weeks and so people are bouncing around a And so there's like three buckets to me as like a very boring software investor.

You have your very high self-s serve, very sticky, low ACV, and a low CAC. Mhm. At his cursor, if I remember correctly, I don't know if they spent a dime on marketing, right? Their ACV couple weeks ago, they said they they'd never sp spent a dime on marketing, but I'm at some point that will change, I'm sure. But yes.

Yeah. And so think about that. You're high self-s serve. You have no customer success. You're very sticky. Your your contract sizes aren't big, but it doesn't matter because your CAC is low. Yep. That's 100 million of ARR in two years. Yeah. Right now, let's think about a business like sales like Salesforce.

Non-selferve, extremely high ACV, and very sticky. Also a good software business. Now, let's think about an AISDR software company.

No self-s serve, many touch points, worse gross margin because you have LLM calls and you're depending on somebody else's email and contact database, high CAC because again, you don't have this amazing self-s serve option. And as we all know now, bad churn.

That's what is what is in your view the future of enterprise sales broadly? Is it is it just back to basics, golf courses, car clubs, you know, you're just uh you know, broing down and you you know, you meet you you're you just become boys with the buyer. Yeah.

Le Man, the best deals in software will be done at lemon and F1 and all these different places. I think there's no substitute for that, Jordy, for a million dollar contract. There's no substitute.

What I do want to highlight is on the notion of the AISDR, the idea of LLMs automatically market mapping and targeting your personas but not sending the emails is obviously high utility, right? Like that is a good thing.

So I think that the future of like really high ticket software sales is probably LLM assisted market mapping and like persona identification, right? And then the golf course, right? I love it. The only problem is that first step, that's not a venture backable company, gentlemen.

That's just a really good feature of Zoom Info or Apollo or any new market entrant, right?

What do you think about the cursor model for SDRs and so it's something I remember there was a company called Streak that was a CRM that plugged into Gmail and it was very self-s served and it was the idea was you're a small company, maybe you've just been tasked with sales. You don't even have budget from your boss.

You just want to speed things up. You plug into streak in your Gmail account and all of a sudden you can do some mail merge and some automation. Cursor for SDRs might look like some email generation functionality, but it's very much that centaur model where the humans working alongside the AI.

Could that be the next 100 million ARR company in the next few months? Well, yeah. And to be clear that the sort of like AI sales co-pilot probably has 50 companies coming at it, but I'm curious. No, I think we I think we're actually the first ones to ever think of that. Yes. Uh let's incorporate right now.

You know, I like to think about I liked about I I like to think about this Kugan's law. This also needs to be coined. I don't know what to call it yet. For an AI product, how many times per day does it call an LLM and derive utility? So let's think about cursor.

You're constantly coding, maybe three hours of deep coding work a day. You're hitting an LLM API constantly or at least once per 10 minutes. Let's just say great AI first software sticky. Let's go back to what you said with the notion of an AI SDR self-s serve.

To me, the utility is how many times are you really hitting an LLM per hour, right? I don't know. Right. And so to me, I I think that there is going to be a self-s serve AI SDR feature that is nice. Happen stance from why combinator did catch my eye. I don't know if you guys saw that. No, great tweet about it.

It caught my eye. Doesn't seem like a $100 million AR business to me yet. Right. We'll see. Good luck to him. You know, I think Happy Stance is very cool. I don't see it being $100 million of ARR uh this year or next year. Right. So, gauntlet's been thrown. Lap stance. You got to do it.

You got to hit 100 million and come on the show to prove carried no interest wrong. Prove me wrong. Last question.

Uh our mutual friend Jeremy Gon likes to talk about the sort of iron law of the business universe which is like if you grow revenue just shockingly quickly eventually you know you might fall back to earth or uh you you could potentially lose it just as quickly.

So what's your take on you know generally on some of these uh various the the cursors the wind surfs etc. Um do you think that revenue do you think they can get to a point where they sort of have a durable uh moat or uh are they going to just be uh forever uh relegated to you know extreme competition?

I get a lot of I I have a lot of VCs who I I talk to in my network that ask me that all the time. Right. Um, you know, here's the real question, and I think let's just cut straight to it. If Cursor hits 300 or $400 million of revenue, could they IPO? And with the share price be supported, right?

Let's cut all the way through it, right? And and and and I think that the answer is it could go the way of Slack, right? It's a double-edged sword. The and I think Jeremy's right, whom I love, shout out to Jeremy. Uh I think he's completely right. It's a double-edged sword. There's no way around it, right?

that as soon as you gain that that person as a customer, you could just as easily lose them. You know, I think that that there could be a Slack Teams situation that plays out with Cursor, right? That classic Slack is amazing.

You know, Slack is is is the trailblazer and all of a sudden everybody realizes Teams is just fine, right? Yeah. And and Microsoft just decides, you know, it's time to come for them. I think the same issue could happen with Cursor on a variety of dimensions. Um Yeah.

I I can't say with any confidence that Kurser is going to IPO. I do think it could get acquired for a like gang busters deal like insane, right?

I don't think it's an IPO worthy company given the churn rate and the the the potential for Microsoft or any any of the big kind of like tech distribution companies to to go at them. I think that it's a double-edged sword. I could eat all those words. They could get IBO next year. What do I know? Right. We'll see.

Uh we got to get on with Taipei. It's been fantastic having you. close out the show with us. Let uh let the audience know to go give us five stars on uh iTunes and Spotify or not iTunes, Apple Podcast. Apple podcasts. Five stars. Five stars for these guys. Five stars. Carrie said it. Thank you for joining the show.

This is fantastic. We'll see you tomorrow everybody. Hey. Bye. You're the man. Almost said his name