Speaker unverified: belief

Transcript context

…live in your product right now and what you're doing and you could store data against them. You could say, “how me all my premium customers,” and weird use cases like that still exist today. Concise— Yes, exactly. Let's just quote a lot of the reasons why or whatever, some sort of disease as well along the way. But yeah, so 2022 it was like, hey, Eoghan returned—Eoghan had left in 2020, I think—Eoghan returned, business had been in declining net-new revenue. He said, “We need to focus. We're going to focus on customer service.” And then a short time later, I think maybe like 10 weeks later, AI happened. ChatGPT launched, I think, on a Thursday. I think I had a call with Fergal, our Head of AI, on a Friday. I spoke with Eoghan all through the weekend and we made the decision, Eoghan made the decision on the Sunday I think it was, and we started working on the AI version of Intercom on the Monday and that was 2022. And I think, were it not for that, it's hard to say exactly where things would've gone, but certainly that's the reason I'm sitting here. And obviously people are naturally wired to be skeptical when they hear the AI version of Intercom. You have everyone and their mother out there saying, “We're an AI company now,” but you actually are an AI company now. And so describe what the AI version of Intercom actually means. The biggest thing it means for us is our product Fin. So we launched Fin in March, I think 2023. We launched a few little AI features. We were the first people to actually build anything on the GPT-3.5 and then we launched Fin in March, GPT-4 launch day, and Fin was basically the first chat bot that worked, is the best way you could think about it. What that really meant was it could actually have conversations and answer questions, and when we launched it, it was doing I think 25% resolution rate and that was crazy numbers. (00:06:59): Today it's like 65%, and today Fin's resolving about, I think it's over a million conversations a week. It's handled about 40 million actual end-to-end customer service scenarios to date. It's growing over 200% year over year. We charge a dollar per answer so you can work the revenue out, or 99 cents even. It's becoming just this AI sort of growth story inside Intercom, which is already a mature SaaS business into hundreds of millions of revenue. But I think when we think about, “What does it mean to be AI?”, it's like, first of all, what is the future growth of your business? And the answer is AI. And then over the last, say, six months, we've been going hard on being a kind of properly deep AI company. We're now at a point where we're using our own models inside Fin, we're using custom reranker or custom retrieval summarization, etc. And we're doing a lot of this work—we have an AI lab of 50 people and we really just have gone all in on the idea that, obviously, Intercom still has a help desk product, but the entire future of CS is clearly going to be AI, and that's what we're all in on. people and we really just have gone all in on the idea that, obviously, Intercom still has a help desk product, but the entire future of CS is clearly going to be AI, and that's what we're all in on. Yes, yes. Maybe there's the repeated pattern in tech where the enthusiasm for technologies comes before the technology being ready. And so people are excited about computer games before the wave of good computer games. Yeah, yeah. Or people are pitching mobile internet and you'll buy cinema tickets via the web. It's like … J2ME and all that. John: And there was this whole enthusiasm cycle for bots in 2017 I want to say, and the tech just wasn't there at all. It was a horrible experience for customers. It was also quite clunky to set up for businesses, and at some point I think everyone looked, there was a genuine question at times where it was just like, is a web form not just better? And I think in a lot of cases it was. Yeah, yeah, yeah. Whereas I think now people are starting to have the experience of—the classic thing is you're talking to a bot, it's like please, will you please just connect me to a human? Whereas now it's like, can you just connect me to a bot? We see that a lot. We see like, “Oh hey Jenny, sorry to bother you. Can you put me back onto Fin? It was actually doing a great job, it just thought it wasn’t. I asked a few too many questions.” But yeah, I think there's a general pattern we're noticing, which is that a lot of experiences are just better digitized because, partially because of human considerations. One of the reasons people go to the kiosk in McDonald's I suppose because as opposed to their actual counter is because they don't want to have to think out loud in front of a human. They're like, oh, “Give me a second. Do I want fries?” Okay, so you're seeing a lot of induced demand for people using interactive customer service. Which is interesting. It turns Fin into not just being a cost takeout, but also it's like how much better would your business be if everyone knew how to do everything they wanted to do? Let's presume you're doing those handcrafted answers—which you're not—someone is busily trying to close the case before they move on to the next one. So I think comparing it, I dunno, we have this thing where we expect our AI products to be flawless and we're totally tolerant of humans showing up only speaking one language and only working six hours or whatever. It's just, it's a funny contrast. hing where we expect our AI products to be flawless and we're totally tolerant of humans showing up only speaking one language and only working six hours or whatever. It's just, it's a funny contrast. Yes, yes. That's interesting, and it's interesting you talk about product onboarding here because I think of Intercom as you guys have a house view that product onboarding should be much better. And I remember a lot of the use cases you would talk about for the original Intercom—talk to your customers through the website—was that you can have personalized nurture tracks, and it's weird that you drop people into SaaS products and just expect them to be able to use them right, and you should see how people are using the product and then give them kind of specific steers based on their usage. And it sounds like you're coming to this vision again, which is: people should have better onboarding support, people should be nurtured along based on their use case, but now kind of interactively AI-pilled. Yeah, I mean we talk a bit about the city of what is ultimately a customer agent going to be. That's what Fin will be as it grows up. It'll just become this way in which customer conversations are handled. And obviously the most direct attack here is customer service. I think every single customer touchpoint can be improved by basically immediate accurate answers available all the time. Isn't an obvious limitation—right now, you require customers to come up with a prompt. And if you look at why TikTok is so successful, it's like I would never prompt for, “I want to see videos of planes landing low over the beach in St. Martin,” but it turns out that's what I want to see. It's reacting to what customers say. Exactly. A customer has to come along and type things into the box. I mean today that's what customer service is. It's still kind of like, here's my problem and then we'll solve it. I think for sure there's obvious directions this will go as like, hey, what does a good customer look like? And maybe we can honestly infer that as well.But certainly “people like you should do things like this” is definitely an understandable domain, and then I just think working at the right level of interruptive help. You don't want to be too naggy or too pop-upy. It gets quite grating. But I think if you can get the first message right, you can sort of say, “Hey, if you come here, you're always going to get the thing you should do next or the thing that looks like you're stuck on.” If someone's on, I dunno, the renewal page and they have an error message, we know they're probably going to open the thing and we know they're probably going to say something that's got a lot of the context already there, so we can work out the right things to say and do. I think that's pretty doable. It feels like you could do a lot around, you train a model on what the customer is seeing on that webpage at that moment in time and use it to feed the answer and things like that.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.