Intercom raises $250M in debt to fund its AI customer agent Fin — now approaching $100M ARR with 8,000 customers

is the super intelligence cloud. building AI supercomputers for training and inference that scale from one GPU to hundreds of thousands. And without further ado, we have Owen McCabe, the CEO.

Uh a huge supporter of TVPN as well. Welcome to the show. How are you doing?

Great to see you.

Great to see you again.

Great to see you both. Also,

it's been too long,

man. You have you just have a voice for for radio,

live streams, podcasts, all these.

I think it's it's incredibly uh calming. I love it.

I love it.

Okay. What would you like me to say? I could record a meditation for us. We could

kind of start out the hour that last. Do

you have time?

ASMR funding announcement.

Do you have time to meditate these days?

It's kind of the only thing that keeps me sane.

That's good. Uh, well, you have some new funding for new project. Break us down. What happened? What What's the plan? How much did you raise? Take us through it.

Yeah, so we raised $250 million.

Let me hit the gong. Congratulations. I'm very excited for that. Did that gun get bigger since I was here?

It did get bigger. It did get bigger.

It would be cool if it got bigger every every single season, right?

Fast take off. Just wait. There's a fast take off.

There's a fast take off size over here. It's growing exponentially.

Except this was actually I think the largest we could find in the US. So, we got to go.

There's always bigger. I kind of have gun envy right now. I wish I had gone.

We'll send one over.

Let's talk about that later.

We'll talk. We'll talk.

So, $250 million. What's the plan? Yeah. Yeah. So I mean we've been building this thing called Finn for 3 years now. Service was just this obvious place where AI was going to you know blow up and us and everyone else started to chase it. We were about a year ahead every everyone else

and we started doing you know customer queries and then more complex problems. Originally was answering questions for customers and it's fantastic. You know the typical response rate with a human was 1 to two days. you have to deal with the beautiful imperfections of humans. Unfortunately, these bots are just far more perfect than humans and they answer in seconds.

And of course, they're cheaper, too. So, we we started to say better, faster, cheaper. These things are really better, faster, cheaper. So, that was a tremendous success. Finn is the biggest service agent in our category, nearly 100 million or 8,000 paying customers. You know, some of the biggest logos in the game. Thank Thank you. from Enthropic to

yeah,

I don't know, Snowflake, the cool guys like

Poly Market and Invent and everyone else are on Finn.

Um, so wonderful.

Um, but where does it go next? You know, clearly it's just so obvious as each of these categories develop that there's just more to explore. There's none of these categories are in any way fully baked.

And that's one of the interesting things about kind of all of these AI spaces that we're still kind of pulling the thread and seeing what's going to show up next. And the next big play for Finn is not just answering service queries, but really getting deep into every stage in the customer life cycle. And so we call that the customer agent.

And this money will primarily bring the customer agent to market.

Okay. So when I want great customer service for a company that I'm running, I want yeah, you know, search the knowledge base, answer the obvious questions, handle the refunds, track. That's part of the customer journey. Uh are there other touch points that are unlocked by AI along that journey?

Absolutely and completely. Yes. From day zero on the website when the customer has very basic questions when they get more interesting when the business wants to start to qualify and understand who that customer is. When the customer actually then wants their handheld when they sign up or they want to actually engage and learn and meet salespeople, they want to book a demo or they want to talk to a virtual salesperson. I mean that's just the very top of the funnel alone but to get into the application if it's B2B there's all the onboarding all the detailed um help required to get someone successful if it's e-commerce it's actually finding products and getting it into their cart and bringing them back for more shopping. So there is really the literally the full customer life cycle of work to be done by agents

and uh customer service was just the little sliver. I want to know what's happening off camera though. Oh. Uh, we we just have our next guest walking in into the studio. So, I'm just They're less interested than me.

Yes. Uh, I I I I want to talk about uh the some of the best uh uh customer service agents that I've hired have

uh sort of wound up, real people have wound up in sales roles almost like they have a fish on the line. They're reeling them in saying, "You don't want to cancel. We actually this product does exactly what you want." and uh that the fact that it was delivered one day late, we're happy to refund you, but

it seems like the underlying issue is that you actually want this product instead of this product. Why don't we get you over here? And that's this like sort of uh different skill set that comes out in the best customer service agents. How are you thinking about actually driving incremental sales through customer service agents? Yeah, I think one of the interesting things is that a lot of the work in these functions isn't entirely separate. We just needed to draw lines between them when we were running human organizations because you can't possibly have one person who's outstanding at service, sales, marketing, success, etc.

Y make the SAS market maps too complicated. You were all

totally

a pain in the ass. But um unfortunately that pain in the ass is coming with AI because I just don't believe you're going to have these separate sales agents and marketing agents. You're not going to have these things fighting with each other having different context and information, different memory in different parts of the customer life cycle, different messengers. Not going to happen. U there's going to be vendors that will provide one seamless solution across the customer life cycle. Um and that's a big thing we're passionate about. And one of the very fascinating things we've seen, we've been building this for quite a number of months now and we're going to launch the first part of it next month. One of the fascinating things we've seen is that when there's an underlying relationship with the agent and then there's actual connection in the moment and then there's clear intent, um the the context is primed for selling and encouraging the user to do something next.

If you're on a website and you got a super annoying popup that says, "Hey, John or Jordy." In fact, they're not even going to use your name. They're going to say, "Do you want to learn more about our whatever?" And you're like, "Get out of here." It's like not the right time or place. Whereas, if you're actively discussing, it can ask you questions back and it's so much more effective. And so, we've seen just these things, the service agent when it starts to do sales, for example, and vice versa, they're far more effective. And so the real magic from AI across the board and of course this agent world is going to be when the agents can do things that humans can't.

Yeah. One one thing that stands out is like right now you have this spikiness in dialogue as a customer of a business, right? You might have a question coming in and it like creates a moment and then it goes away and then you get maybe transition to someone else and you're talking to real human. you talked with them a little bit and you get passed to someone else and like there's something about creating I think the customer experience of having this continuous conversation with the organization that you're a customer of that I think it'll be super powerful because it just means that you know a split second away the entire time. Uh I I I was talking uh with some guys that own a number of like big car dealerships recently and I was talking to them about AI and they weren't fully sold yet because they they use like chat GBT but but they haven't had anything spread across their whole organization and I was explaining to them how there's so many different moments where like a customer lands on their website and they want to know if a car is available but somebody gets back to them too slowly and like they it the light was like finally going off for them. Do you think that there's a number how many industries out there have been kind of resistant to even

using software from Silicon Valley to date that are going to kind of take a second look at all the different tools and things like Finn now that they maybe have had a cool experience with AI in their in their personal life and they're basically saying they're they're willing to kind of look because maybe the the historical SAS platforms they tried them but it wasn't quite uh magical Yeah.

Yeah. There'll be so many spaces, so many verticals that um some smart entrepreneurs will benefit from bringing sexic technology to um that aren't immediately obvious to those of us in Silicon Valley. You know, we tend to sell to ourselves. It's really insular. So, all the names I just mentioned there like Enthropic and stuff like they're all the cool brands in our world. However, there's like car dealerships that are, you know, I don't know, maybe it's not as big as as anthropic, but they're pretty damn big. And

depending on the car dealership, I I would put car dealership above Anthropic and Snowflake personally if it's like a Ferrari dealership, Lamborghini dealership, like just on the cool side, we got to give it to on the aura factor. Sure,

for sure.

Maybe not on on the economic factor. Yeah.

But there are there are so many industries that just don't have our attention. And so there will be, you know, there'll be companies like ours sell into many verticals, but really, you know, we'll focus on all of the verticals that are adjacent to ours.

And there's, you know, tens of billions of dollars of opportunity just in that alone.

But there'll be people who will say, "We're going to make agents for dentists,

car dealerships, or certain government bodies, maybe even for tax bodies." like there'll be people who will do this because so much of the effort is not just in building the technology. It's actually creating the trust as you say Jordy with the actual people who will have to deploy this technology and make it work for them with their legacy systems. So there's a lot of very unsexy work that will be rich and rewarding for some people who bring it to these places.

Talk about the decision to raise debt instead of equity. Not something you see.

Yeah, I mean Yeah, I know. I I just realized I had this kind of blind spot for this as an option for for capital and there's just so much equity opportunities out there. We raised a bunch last year for the secondary, but when you're in our dynamic where you're accelerating as quickly as we are, like every every year we we're doubling our our rate of growth pretty much, there's no way we're going to be able to achieve our the valuation that we're pretty confident we'll get next year.

Yeah. And so any equity raise is going to just be massively dilutive. We calculated that raising this in debt is going to cost us about a tenth um the kind of price to shareholders that it would cost in equity. Wow.

So this is you know the incumbents have disadvantages.

They tend to move a little slower. They have older businesses they need to wrangle with but they have access to phenomenal amounts of debt. Like 250 million is a relatively small amount of debt for the size of our business. Um, and I think that you're going to start to see a lot of these older SAS companies as they pivot to AI and they actually find new growth start to lean on debt as an opportunity um to invest rather than,

you know, super dilutive equity.

And and honestly, I just think it takes a little bit of maturity to take that route. There's so much pressure. It's not just maturity, it's it's confidence

to know your business. Yeah.

Well, I you know, I like to think so, but what I the reason I say maturity is that there's so much temptation to go and attain that latest hottest valuation. Your employees want it. Your previous investors want it. Media wants it.

But if you're able to tell them

we don't need it,

actually it's way cheaper

for us to get a lot more capital with far less effort.

Um then you can actually thread the needle in a way that I think is going to work much better for big companies like ours. gong over here rings just as loud for debt as it does for equity. So, don't let the

I didn't You guys are hiring a ton. You said uh you're you're bringing on 650 new people this year.

Yes. Yeah. So, that's gross. I think net it'll be a little less than that. It's a phenomenal amount of hiring.

Um I fear it. Um I celebrate it with great caution. There are just countless stories of the hubris of CEOs who boasted about how many people they've

hired and it's all ended in tears.

So, it's actually not a point of pride, frankly.

Um, I think small is beautiful,

but we are growing incredibly fast and we need all the help we can get.

And you've and you've been at this long enough to know the impact of what you know 100 people, 600 people, what what impact that will have on your organization. So, I I have the full faith in you.

Thank you.

Thank you so much for taking the time to come chat with us.

Great to see you.

Congratulations.

Good to see you both.

Keep doing your thing.

Amazing stuff.

Byebye.

Let me tell you our Railway. Railway is the all-in-one intelligent cloud provider. Use your favorite agent to deploy apps, servers, databases, and more while Railway automatically takes care of scaling, monitoring, and security. And let me also tell you about Sentry. Sentry shows developers what's