How to Survive and Thrive in a World of OpenAI

I’m happy to be here. I love London. It’s awesome to actually do this in person and see you in three-dimensional Harry.

I think it was either a a professional video games tester, but I kinda knew that was in my career, or at least definitely wasn’t in the eighties. Sports journalist. I loved soccer and I loved writing. So I really wanted to, like, write about soccer. I just thought, like, if I could do that, that would be, like, the genuinely the easiest job. I used to actually do it. I used to, like I’d watch a game and I’d write down my match report and I’d get my mom to read it and she’d like, that’s very good.

But, yeah, it was I I think I love soccer watching it, and then I literally love writing as well. I’d always read all of the cover to cover every sports newspaper, everything just trying to get a sense of what’s the language, how are you supposed to talk, all that sort of stuff. Thankfully, I didn’t end up doing that because I have friends who are sports journalists, and it doesn’t pay as well as I thought of mine.

But

It actually started life inside a different product. We, had a different product called Exceptional. We wanted to talk to our customers, so the logo for Exceptional sat in the bottom right hand corner of the product. And one day, a little speech bubble came out of it, and it like, hey. Probably something like, hey. We’re sorry that we fell over or the server wasn’t working last weekend or something like that. That was a much better workflow to have that bubble pop than the old world. The old world, if you have if you some of your viewers might be old enough to remember, you’d send out the mail to say, like, hey.

We’re you know, here’s some preannounced downtime or more likely some apologizing for some downtime we didn’t know was coming. And then you get a shitload of replies in your personal email. It was crap. And then with Intercom, it was like one day, it was like, why don’t we just talk inside the product? Another sort of inspiration for us is we were working out of a coffee shop called three FE in Dublin at the time. And the dude who was building three FE, this new recurring software business recurring subscription business, if you like, like, where you’d be selling coffee every day to the same people, He was talking to people.

He was, you know, hey. I’ve got a new blend in. Do you wanna try it out? Or, like, hey. Is your latte nice? Or whatever. He was getting, like, real time in context feedback. But for us, we were felt really, like, divorced from our users. In our heads, we’re like, we’re sitting here in Dublin, Ireland. They’re all in San Francisco. It was a single street in San Francisco where we had more customers than we did in all of Ireland, to give you example of of the gap.

But what we realized pretty quickly was it wasn’t a unique problem to us because we were in Dublin. Everyone who runs software businesses wanted to have that personal connection. So we started to, like, iterate on this idea of, like, maybe users should be able to talk back. Maybe you should choose who you send a message to. That ultimately became what we call, like, the customer communication platform Intercom. Today, we call it, like, an AI first customer service platform because it’s kind of been ten years, and we’ve matured our thinking a lot.

But,

Yeah. You’d be an idiot not to Why why would you not be doing that? Yeah. I’m sure coffee shops are AI first too as well.

Picture of, like Yeah. Yeah. Yeah. We’re gonna end up talking with AI for a while now, so I’ll just start off. I think it’s largely bullshit if the basic tech stack hasn’t actually changed that much. Let’s say you’re a project management tool, and, like, hey, what what is project management? It’s like tasks, deadlines, files, approvals, decisions, communication, whatever. It’s not, like, really obvious to me that AI is going to take away a lot of that because, like, you’re planning a project, and, like, oftentimes, these are the things that really have to happen, whether it’s like you’re building a house or you’re you’re managing a rebrand for a client or something like that.

There’s a lot of, like, human decision and judgment made. It’s not obvious to me where the AI kicks in. So if you’re, like, an AI first version of Basecamp, I still think that looks like Basecamp, and maybe they just find areas to sprinkle in AI to speed up little bits and pieces here and there. Maybe. I think there’s if you take, say, the area of customer support, everything about customer support is gonna be affected by AI. For example, summarizing an issue is a very common workflow that has to happen.

Replying to a customer based on the fact you found in a knowledge base is a very common thing that has to happen. Consolidating different articles to to give a conclusive answer to something, being conversational, these are all things that AI can do. So I think if you were to build a customer support solution today, you’d build it totally different than, say, when Zendesk was started in 2007 or 2008 or whatever. Like, I think the world of, like, ticketing and structure and state and, like, you know, we’ll get back to you in seventy two hours and all that, that’s all gone.

To your question about AI, I just think it’s bullshit when actually the job is identical and you’re just sprinkling on, like, salt and pepper a bit of AI here and there. I think it’s true when it’s like, shit, we need to reimagine this entire space.

I think that’s true. And when that happens, I’ll support them saying it’s AI first, but that hasn’t happened yet. Right now, we’re we’re firmly in the, like, salt and pepper category in my opinion. But you are correct. Like, the job of, like, you know, managing a project involves, like, reading email, checking tasks, looking at GitHub commits to see if something got approved or not, checking off tasks, and corresponding with a calendar and seeing when things are and and be using all that information to take action or to request action, then, yeah, if somebody builds, like, the the project megabot that can do all that, I will celebrate them while investing in their company.

I’ll I’ll call them, like, an an AI first project management solution. We’re not there yet.

One area where we have been fortunate is we were all in on messaging from the very start. That’s where chatbots live. The idea of a chatbot is actually the intersection of two megatrends, AI and its messaging. That’s what makes a chatbot. So I think we were in a very fortuitous position to have a complete sort of CS platform and have as our core mode of interaction, actual messaging back and forth. So I think we were we were in pole position to do it. There wasn’t a substantial rebuild.

Our first AI product was built in 2016 called ResolutionBot. Bots kinda went through different eras, if you like.

and all

Precisely. Yeah. And Facebook had their big demo of Yeah. You know, all that. We basically went from, like, bots that were just literally if this then that logic, you know, very IVR style bots. Like, if Harry says one, then we take them down path one. Right? The next generation of bots were like, what I call it fuzzy AI, which is like, we’re gonna take Harry’s message, and if it looks like he’s asking about pricing, we’ll give him the pricing answer. Right? And that generally had broader coverage because it was an attempt to be conversational, but actually, in a lot of times, it just pissed people off because the fuzzy logic wasn’t good enough.

And you still have to go and build all of these if this, then that rules. You just had the magic AI inferred piece. Where we are today is, like, generationally different, which is the LLM era, which is we know what Harry’s talking about, and we can infer it. And we can actually work out the specific question and find a precise answer. I think that’s the sort of era that we’re living in today. And so

Oh, it’s a thin layer on top, so it’s not very valuable. There’s a big difference between a feature and a product. If you wanted to say build a chatbot that does customer support, that’s pretty easy to build. It’s like half a day’s worth of work, and you’re sitting and you’re wholly dependent on OpenAI or Anthropic or whoever it is, you could go for it.

There’s a massive difference between that and say something that, like, a large company would actually deploy to its customers, which, you know, ingest knowledge bases, requests, refreshes, reports on itself, reports on customer satisfaction, you know, can parse multiple conversations, can understand customer context and feed that in, knows the difference between a premium user and a free user and gives context to where answers, whatever. Like, does a wave upon wave of extra stuff you have to build in. Like, that’s before we get to, like, permissioning, controlling, and targeting, and making sure that some users don’t see bots and some users only see bots.

Like, all of that is, like, the extra stuff you have to build to turn what’s, like, a thin how would you say, like, science fair tech demo Yeah. Into an actual, like, fully commercialized product to the VC argument of, like, a thin wrapper. I think if it is a thin wrapper, it’s a bullshit argument. But if you take, say, like, some people are building on top of, say, Dolly or something like that. Right? Literally, it depends.

You can build a canvas style product on top of Dolly where you’ve got like digital asset management, you’ve got controlled and approved images, you’ve got workflows that support all that, you’ve got embed functions, or you can build the thin wrapper, and there’s just a massive difference between those two. There’s also such a thing, I guess, as a very thick wrapper, And the thick wrapper is is like when you’ve actually solved the user’s problem end to end fully in a way that OpenAI never will.

all aspects of users’ workflow. You have to fit into their tech stack and integrate where you need to integrate. You need to, like, solve for, like when I say all aspects of workflow, I mean, like, in, like, the hypothetical, like, Dolly wrapper. Like, you need to solve feedback feedback on various images, or maybe there’s bit of a collaboration on the client side where they wanna discuss things. Maybe there’s an approval workflow, and then you’re into, like, asset management after that. Then maybe there’s hosting, maybe there’s embedding, all that sort of stuff.

And then there’s a reporting on usage. You know, if it’s a marketing product, maybe, like, does it correlate or, like, does it test well with, like, conversions or does it not convert when we use it or whatever? Like, that to me is, like, all of the shit that you would want around a tool. If you’re gonna say, like, this is kind of about AI first, if you know what mean. And I think if someone goes and builds all that and makes it make sense, the dude around the corner who’s got, like, this in wrapper, like, he’s got a lot of shit to build.

You’re then back into what would call, like, maybe the classic B2B arms race, maybe. Those who really understand the user workflow are the ones who actually win this, not those who kinda, like, straighten their glasses and say, academically, that’s been possible for seven months now. It’s like, great, mate. But, like, people actually need to buy real software and use it for real things.

I think that’s probably true, but I what I would say is, again, going back to the thin wrapper, thick wrapper, like, if what you’ve built is like the wealth management advisor bot, yeah, you’re dead. Right? Like, as in it’s gonna be like an actual one of the GPTs, that’s just gonna be available. Alright?

If instead you’ve built something that, you know, connects to your Robinhood or your eToro, connects to your bank account, monitors your wealth, your exposure, you you know, issues you case by case advice, monitors the markets for you, like, again, thick wrapper, like, those are the actual things you have to do to manage somebody’s wealth and can have a conversation with you about, like, and suggest actions and next best steps, that’s still, I would say, substantially beyond, like, what OpenAI would ever do. You have to bear in mind, a parallel here might be like, look at what iOS ships with.

Right? It ships with a basic notes app. If you want a really fancy one, off you go, and you can go buy Reflector, Bear, or something like that. Like, it ships with a basic camera. If you want photo room, you have to go buy photo room. OpenAI will hit some sort of minimum viable, like, kinda what’s good enough for everyone. You’re never gonna make money filling in in the gaps in a platform, the thing they haven’t gotten around to yet. Like, I described that as like it’s like you’re on a train tracks picking up, you know, euros or dollar coins or whatever.

There’s a train coming. It’s gonna hit you at some stage. It doesn’t matter how rich you get. That thing’s hitting you. Right? However, I think if you find an area where they’re not gonna go that deep, OpenAI is never gonna put, like, five other engineers going hard on wealth management and, like, you know, banking integrations. Yeah. So, like, that’s the thing. Like, you pick an area and you say, let’s do all of it. Let not let’s just do a little tech demo science fair. Let’s nail the use case.

I mean, it would suit me if we did.

Because then there’d be no pricing power, and then we could actually base, like, you know, we could base Fin on a far cheaper thing and would increase margin, you know, like the the Kyoto. Another alternative will be if there was an actual best one, but but, like, the best CS the best LLM for customer support requires hundreds of millions of customer support conversations. That would suit me too because we could train it uniquely. Why does Intercom not and I understand the heavy expenditure and the heavy r and d.

I mean, you kinda just answered the questionnaire. Like, why do we not train our own LLM? Yeah. Officially, I just said, like, no comment. Like, maybe we are. But, like, you you know, the the reason I love you so

The the the reason, like, the reason we might not is the reasons you would suggest, and the reason we might is the reasons I just suggested a second ago. There’s a dynamic there. You know? Do you think that they will continue to be commoditized? I I don’t know if it’s actually happened yet, if I’m clear. Right? Like, right now, it at least to like, when we look at, like, how best to do the Fin thing, which is to, like you know, and we see we see staggering results for Fin.

Some people are getting, like, know, like, very, very high percentage of total support volume resolved with satisfaction, above 50% of times. Right? Like so all we care about is, like, who can do Fin the best for us? And we actually torture test all of the LLMs. It’s not yet the case that they’re all equal. What were your lessons from torture testing all of them? So it’s really important that Fin stays on topic. So if you ask Fin Sorry. For those that don’t know, Fin is. Fin is our AI chatbot that sits inside the Intercom Messenger, and it will have conversations with your users and help them, answer the support questions so that your support team don’t have to.

We generally remove all of the undifferentiated heavy lifting that a support team might have to do. It’s a product by Intercom that uses an LLM. It uses GPT-four by OpenAI. Now, when it comes to, like, picking the model, the requirements we have are, one, is, like, trustworthiness. So, like, we need to be able to prompt it in such a way that it will not hallucinate. Knowing that that prompting will occasionally cost us a valid answer, but it’s really important to our customers that that all the answers Fin says are, like, high confidence.

A second thing is that it has to stay on topic. So we don’t want Fin taking opinions on current events. So if you go to, like, a bank website and you say, hey. Who’s the president and are they doing a good job? A bank’s customer support has no official policy in that, and Fin better not either. Right? So there’s a lot of, like, staying on topic that is really, really important. And then it’s, like, just depth and power of quality of conversation. So we have a set of metrics that we kinda look at, and we and we configure and tweak our prompts and tune as much as we can to get it as good as we can.

And right now, we are, like, in with OpenAI, and we use it on GPT-four. And it’s against that logic I say that

What did the test show you? It’s basically quality of conversation and, like, does it fail any of our, like, hallucination tests? Does it fail any of our trustworthiness tests? Can it infer its own confidence? Was it closer, or was there a way? It close and narrowing, and and, like, it’s also a work in progress. Like, all of these things are moving targets. Right? We have been moving very fast in this space. Fin launched, I think, in June. A lot happened since then. So, like, we haven’t even gotten around to maybe testing the latest and greatest of all the providers, which are increasing a number.

You mentioned Mistral. There’s also Glamour. There’s Anthropic. There’s, like, Cohere. Like, there’s, a whole chunk of them, and it’s a bit of work to go around and constantly be trying to find out has anyone got we only really care about better. Right? Right now, we’re not in cost optimization mode. We’re just like, who’s got the best?

Yeah. Yeah. For sure. Well, I don’t know if I agree with the sentence, but I think it’s a really important attribute to be able to transition from one LLM to another. Because if somebody does unlock new power, you’ll wanna be able to use it very quickly. Winning involves more than just simply being agnostic about your LLM.

Yeah. All the way down to your homegrown cheapest one or Yeah. I I think even within OpenAI’s ecosystem, you can use GPT-four turbo, GPT-four, -five turbo. They cost different amounts, and, like, they’re good at their, like some of them are faster than others. Some of them are more accurate than others. So, like yeah. What’s been the hardest thing about kind of the integration with OpenAI? The moving target, and that’s not a bad thing. They they move incredibly fast. We can grow a lot of confidence in a particular direction, and then they can unlock new capabilities.

And we’re like, oh, I guess we should jump on this instead. That’s definitely one of the challenges. We have a great partnership with them, but, like, where our team moves fast, but they’re, like, throwing out new capabilities regularly. So Does it feel like unnerving being moved in that way by an external team? I’d say more exhilarating, like, you know, we are all in on AI, so we have to. We just have to be the best providers of AI for customer service. So it’s not really an option.

Like, if and when the world stabilizes stabilizes a bit, like, we I suspect we’re still in this s curve, we’re still in the spike. There will be a plateau eventually, right, where, hey, most of the main problem areas have been fleshed out, and we’re starting to get a bit more sober now. I think that’s when we can move into, like, a different mode of work. But for as long as, like, you know you know, our whole team watched Tuesday’s Dev Day by OpenAI, again, it was like, let’s get to work.

Like, there’s there’s new stuff here. There’s new things. Customers don’t have questions. We need answers. We need to have opinions. It might change the road map. I don’t consider it so much dependent on OpenAI. I consider it dependent on, like, I would say, the boom time of AI that we’re in full stop. I liken what we’re going through right now to be, like, the beginning of the Internet. It’s very big what’s happening to both our world of customer service, technology in general, and society in general.

I just think we don’t have the luxury of sitting back and watching. I think if we were like a massive, like, government sanctioned monopoly, we could be like, let’s just sit this one there for three years and see what happens. Whereas in practice, I think this is much more like, hey. There’s gonna be a lot of winners and losers, and there’s gonna be a lot of market share, for grabs over the next two to three years, and people are gonna be looking for who’s got the good shit.

That needs to be interim.

The way we’re working today still is that we have a central AI team led by who’s our VP of AI, and, they work on Fin. They work on the core, like, question and answer model of Fin. They also work on, like, things like what’s the best LLM to use, etcetera. And then they work on, like, what we call horizon two stuff, which is, like, this could be a total brick wall we run into, but we we need to try it because if it’s not a brick wall, we need to be the first ones to to realize that.

It’s very centralized, and they effectively have breakthroughs, and they go and partner with product teams who might own Fin or might own the inbox or might own the reporting engine, and they say, hey, we just realized, given this, we can do that. Does that seem like some of your customers would use? And they partner with the product manager, the EM, and designer, and they say, right. Here’s a new capability. Like, literally, there’s a new capability that did not exist before. Like, this like, is, literal breakthrough shit.

Here it is. Can you see a way to use that in your world? And most of the time, the answer is yes. So then we’re like, right. How does that rank and stack in the road map against everything else our customers want? That’s how we’ve been working for quite a while. It’s How much of resources resources are are allocated to r and d and AI today? R and d and Intercom is about, I between three hundred and fifty and four hundred people or somewhere in that space.

And then how many are AI? I wanna say, like What what percent is that? I’m just intrigued. Of the company? Yeah. We’re 900 people. So Wow. Yeah. Shit. Is that shit big or small? That’s big. Yeah. I mean, we’re like a technology company.

Yeah. Slightly less. We probably we’d just gone through a rift, and I guess AI hadn’t happened. You know? Like, AI, literally, like, we’re coming around to the anniversary at the time of recording. Right? Like, we’ve always believed that, like, the best product wins. We’ve always invested hard in product. And then on your AI question, I said, like, it’s me. You I think our we anticipate growing the team to maybe 50 or so.

Yeah.

tweet

I think it’s reasonable to assume in this massive sea of uncertainty that we’re in and the compute being so expensive, the margins being so thin, that a higher than normal amount of these startups will die for one of the following reasons. They are the thin layer, and they get wiped by an OpenAI upgrade. Right? Similar to, like, remember, like, in the early days of the iPhone, it was like, you know, it was a torch app, and there was like a copy and paste app. Totally. All these things just got slapped once iOS started chipping.

Right? Like, was ago. I guess we We still see that now. Yeah. Yeah. It’s still. Sherlocking, they call it. There’ll be more than that than there would be, say, in, like, ye old B2B SaaS or whatever. Right? So I think that’s gonna be one cause of increased fatality for startups. I think a second one will just be margin compression. The B2B SaaS model, LTV CAC dynamics, and all that sort of stuff, it does have some assumption in there of, like, 80 to 90% margins. Maybe 70 to 90% margin is kind of what you expect if you’re running, like, an Asana or something like that.

So if all of a sudden all your core workflows are now powered by reasonably expensive API calls that you you know, it’s not yours, then all the money in the front door is going straight out the back door. I think was like had this phrase and he said in 2018, 2019, giving a dollar to a startup was just a really lossy way of giving it true to Amazon, Facebook, or Google Yeah. For for, like, hosting or acquisition. It’s arguable in a lot of cases, ’22, ’23, ’24, giving a dollar to a startup is just a lossy way of giving it to OpenAI or Anthropic or whoever.

You know, that’s who they’re basically taking it and shining it up and handing it on to. That that’s gonna be a second cause of, like, increased startup mortality. But beyond that, I think the normal dynamics will play. And then the other piece is just in that world, like, let’s say, an above average number of these new investments do die. I also think an above average number of existing incumbents are going to die too. Their world is gonna get torn apart by AI. There’s a lot of tools out there that won’t make sense in a world of AI.

Sorry. Like what? You have to like take a slightly more future facing look for this. But let me just describe to you a tool that I I know exists. I just don’t wanna name it because I don’t wanna tell my friend that their startup’s dead. Let’s say it’s an ad optimization startup. Right? So you log in and it looks at all your AdWords, and it tells you which ones are, like, performing and which ones aren’t, and you can generate new ads and all that sort of stuff.

And it does that across Google and LinkedIn and Facebook or whatever. Right? And people log in every day and they generate different versions of the winners and they turn off the losers and they look for like the LTV CAC optimum performers and they splice those splice those out into like variants and run tests, and they produce a dashboard to show that they’re doing their job and they’re maximizing the spend of the advertising. What bit of that do you think AI can’t do? If you give it one ad and you say, give me 10 versions of this, it’ll do it.

If you ask, give me 10 versions of this with a visual, with a visual that’s left aligned and a tagline that’s on the right, Can you run it? Ping the API server to put it live? Can you perhaps tell me, is it performing or not? Can you sort the list by LTV CAC? Can you calculate LTV CAC? Can you produce versions of the winners and turn off the losers? Can you produce the dash? What’s the reason to log in to that product?

That workflow is not defensible because what the humans are doing is genuinely like automatable. Like, there’s no actual human intuition or deduction or whatever being applied in that case. Let’s say digital asset management, right, like, the example I gave at the outset of the podcast, I was talking about, like, hey. Like, really, like, brand advertising campaigns, whatever. Like, the the assumption in that workflow is that the human opinion does matter and that that people actually do care about what their logo looks like or whatever. Right?

If they don’t care and they’re happy to have it picked algorithmically, yeah, that that whole thing goes. But, like, there are definitely gonna be, like, full workflows end to end that no longer matter. Every step is just an if this, then that style application of thinking. As in it’s not anyone’s opinion which is the most LTV CAC positive ad. It’s effectively likely to be of low value to have human differentiation in picking the winner or even perhaps producing the variance. Maybe the greatest thing I’d say as to why there’ll be a lot of startup mortality is that, like, if you and me were to start this ad optimization startup today, would it look anything like the previous tool that’s built?

Put it in the way, if we were allowed to rob any of their code, would we? And I think in a lot of cases where, like, you’ll see the the full disruption, the answer is no. You’d build the product totally different today. But then there are some cases, for example, let’s say we said we’re gonna build Mailchimp, but here’s our thing. We’re gonna use AI for the composer, so you can just tab to complete sentences or something like that. Right? What I would say to you is, okay, is there an opportunity there?

Well, here’s my here’s the problem. To be a Mailchimp, you able to send millions of emails. That’s a massive challenge for, like, sender reliability, reputation. You need to be able to do attribution on which links get clicked so you can produce reporting and all that sort of stuff. You need to do, let’s say, partnerships. You need to, like, you know, connect with, like, Squarespace and Wix and all that sort of shit. Like, you need to have an embeddable subscriber widget, all that shit. Right? We have to go and build all that and our little bit of AI on the composer, like, that lets you complete the sentences.

Mailchimp just needs to go and build the complete sentences bit. So what we’d have to believe is that we’re gonna build all of their stacks that we can send 500,000,000 mails a day and all that sort of stuff in the space of time that they can’t build the AI. That’s the sort of calculus I think about when I think about where where the advantages are. I think if you were to rebuild it today and you’d build none of what’s there and everything would look totally different, then, yeah, it’s go time.

Right? If you actually wish you could borrow all their servers and their sender reputation and all that sort of shit, then I think you’re trying to add a thin layer on top of what is ultimately a surviving incumbent. We used say this a lot about about, let’s say, email startups. Right? We’re gonna do email, but we’ve got one new idea on top of email. And I’m like, that’s great. So you’ve got five years of roadmap to get to the Gmail quality or even superhuman quality, yeah, like email, and then you can add your little feature on top.

Long

Yeah. Yeah. But I agree with you. I think the nature of the role you have full time and I have whenever on the weekend I decide to write an angel check is I I can’t really think about like, will this be a big bang next year? What I would do is I would say like, if I see somebody going after an area where I think there’s like a real massive chance for like investment, I’d probably say I should sell my stock in the incumbent over the next two years, and I should plant the seeds in this upstart today with a view towards payday in seven years.

Even in seven years time, the revenue streams might not have crossed. Like, this thing could be like this, but still far below the other. You’re betting on the future. Like, it was ultimately what we are betting on when when I invested in AI for a startup today. I’m betting on what the world and adoption and market appetite for AI is gonna be like in, like, at least four years. If I try and give the Fin pitch to a lot of, say, old school banks or insurance company, they’re like, woah, horsey.

Slow down. I liked the bit where you said you could handle a simple reset password, but now you’re telling me it’s gonna, like, talk to my team. You have that dynamic of, like, you know, it’s that old that, like, crossing the cosmos style adoption where some people are, like, early adopters, some people are, like, open minded, some people are, like, cynics, and some people are laggards or whatever. We’re definitely only selling to, like, the first 20% of the market at the moment.

Yeah. Some digital transformation project that got spun up for a $100 or whatever.

Yeah.

We care a lot about the ROI of Fin, and, thankfully, it’s kinda self evident. Like, as an example, yesterday, I think OpenAI had a forty five minute outage, and I think everyone got a real taste of what the world was like before you had Fin. Because you got these support teams being like, yo. We’re all what’s this new spike in volume? And we’re like, oh, Fin had to, like, hand over to human. But areas where it’s very speculative and open ended, we thought it would be cool if we could run a little bit of AI to generate ideas for a workshop or whatever.

Right? They’re definitely gonna get cut to zero, and somebody’s gonna check a box saying, and we’re very future focused. But I think anyone who’s got a product that is genuinely, like, delivering real value, I think those budgets are gonna stick around and actually massively increase.

I think there’s loads of obvious applications, like our world most directly. Like, if Fin is indeed doing, like, 50%, then, like, will businesses buy as many seats? What we’re seeing a lot is, like, businesses are choosing to keep the seats and keep the people and, like, just turn the dial on quality, so they just get really good at support. In general

It’s a good question. I mean, another way of saying it will be like, if in a year’s time there’s another market downturn, will they still hold firm on that? But, like, in general, his point about, like, consumption pricing outpacing or, like, ultimately taking the place of seed pricing, I think, is going to be true. I think a lot of work is gonna get handed over to LLMs over the next five years. We’re gonna start trying to, like, price against the work that’s being done, not price against the seats or the employees, but just say, much is it worth for you to have all of your digital assets created dynamically?

Or how much is it worth for you to have, like, your customers get sub second replies to common questions? Questions? Like, that’s the actual right way to think about pricing in the future. The wrong way to think about it is, like, is, you know, you we’re selling work. We’re not selling seats, and we’re not selling, like, you know, incremental tasks completed. We’re just like, what does it cost to get done? When somebody actually does a good job of an AI coach or personal trainer or financial adviser, the person who does it really well will be able to justify pricing it similar to what the human costs because people have already been paying humans for this for a long, long time.

But people will be comfortable transferring that fee to a bot because it’s $24.07, it’s multilingual, and they can bring the price down as they see fit as well.

There is no tech really in history that has not gotten cheaper. There is no piece of hardware or software that hasn’t ultimately been ripped off by a cheaper candidate to offer a lower prices to more of the market. No matter what way we go, we will see a kind of a flashing out of the spectrum. So right now, I think we’ve got a high end NVIDIA pairing, high end LLMs, and we should assume on the fullness of time that they’re gonna, like, continue to produce, and they’re gonna find more market points that they can sell potentially simpler, but more basic chipset.

And same with the LLMs. There’ll be the fast LLM, and there’ll be the accurate LLM, and people will make choices between them, and they’ll be the the cheaper one and the the other one and all that sort of stuff. We’ll see the solution space fully explored. And in that exploration, I I don’t think NVIDIA in five years time will be sitting here proud as the only people who can do a GPU. And I don’t think that OpenAI will be able to, like, be the only people who can do LLMs properly.

You know, there’ll be winners and losers. They’ll be, like, the best, but the best isn’t necessarily the biggest market share. This is all gonna be a spectrum, and, like, people are basically gonna be finding where is the most dollars to be made. Right?

buyer of OpenAI at 90,000,000,000? I know that they have a funky thing where you can only make a certain amount of return back on your investment as far as I’m I’m aware. Like, I think they have some sort of policy, like, after a 100 x, you have to sell back or something like that.

As a net new investor, would you buy at 90? Yeah. Just straight up on clean terms, would I buy at 90? And I think they’re doing a billion in revenue or something like that is what I’ve heard. I probably would buy a bit to be along for the ride, but I don’t think I’d be, like, putting millions into it. It it would be a small fun check for me. I’d on the whole, I’d make the bet that they’ll pass 90 in value. I think that’s doable.

But you gotta see multiples on that. You need to see a five x on that. So you actually is it

Yeah. Yeah. Do I so so you need to see multiples. I don’t. I’m just throwing in fun money. Right? Like, if I was an investor trying to, like, you know, make my name, would I go all in, like, on putting, like, I don’t know, 10,000,000,000 in something high for 90 or whatever. I don’t think I would on the reason whilst I think they’ll pass 90, your question was, could they go to 500 in the near term or, like, within the life size stage of the fund?

The areas I’d be wary of is Amazon. Amazon played this game well. So, like, I I could see Amazon just, like, flat up, like, buying Anthropic and being like, let’s just make this part of the EC two cluster, and that’s just a very easy route to market. And I think if OpenAI run out of new vectors of differentiation and the commoditization starts to kick in even for basic stuff, I think it’ll just become easier, like, why would you choose Amazon? It’s already in the cloud. It’s already virtually private.

It’s you know, you you can, like, leapfrog a lot of other adoption concerns. So I think they’re one risk. I think Microsoft are probably less of a risk because of the pretty close partnership they have. And then, like, I don’t really know what Google has up their sleeve. I don’t even know if they have sleeves at this point, but I suspect that they won’t be totally napping on this. It it is Have you not

I played with Bard a bit. I wasn’t unimpressed with it. It felt like another version of a ChatGPT. I’m more underwhelmed by, like, their leadership in this area. Like, it feels like Google, who kinda owned all this at the start, bear in mind, with the original paper on transformers, feels like they have, like, rescued defeat from the jaws of victory. And what we saw on Tuesday of this week, from my mind, was like Sam and, to some degree, even Satya out there talking from a leadership position, from an authoritative position about where we are, where we’re going, what’s happening, and what the future holds, etcetera.

And I don’t think we’ve had that yet from Google. To take a leading position, I think you need to, like, do that thing where you walk out on stage and say, like, here’s our mission. Here’s where we’re going. Here’s where the whole world is. Like, hey. We have a big wide eyed vision. Elon Musk saying he wants to terraform Mars. That leading a company in a certain direction saying, here’s the end goal. I feel like Bard unfortunately felt like we have to release this because ChatGPT was getting a lot of traction.

It didn’t feel like we’ve actually cracked search again. We’ve reinvented ourselves all over again. You know? They need to have that sort of a Jay z, like, allow me to reintroduce myself moment, right, where they come back and they say, like, Google two is here. That’s we what need to have here for.

Now now you’re pulling onto the real potential problem, which is are they willing to risk it all to win it all. Right? Like, are they willing to disrupt themselves, or are they happy to take the, like, long slow decline into obsolescence or irrelevance or whatever. Right? Versus saying, no. We’re gonna, like, take a massive hit to our own revenue, and we’re gonna go, like, LLM first, and the future of Google is actually gonna be a q and a style thing even though we haven’t yet worked out how we’re gonna monetize that.

It’s a massive existential risk, and people have thought about I mean, it’s easy to say here, but if you are CEO of Google and you have shareholders of Wall Street I kinda hate what I’m about to say. I’m gonna say it because I think it’s what I it’s what I would feel compelled to do. I’d be scrambling to find ways that companies can sponsor injections into the LLM. So if I said, like, who is the best footballer in the world? And the answer is clearly like Lionel Messi, but you could say something like, according to transfer market, the answer is Lionel Messi.

Right? Like and that’s really my sponsored injection. Right? And then, like, you could augment the Bard style answer with such injections so that it’s presenting you facts that it it is kinda disowning. Right? Like, it’s kinda saying, hey, like, this isn’t the LLM deducing this. This is just what we think. This is what we’ve been paid to say whenever we talk about this type of thing. That’s kind of the attack factor I’d go on, and I’d try and float that with all my big ad buyers and sort of say, hey, look, Let’s be honest.

The world is gonna go to LLMs. Even if this doesn’t work, we have to give it a lash. Then I’d try and, like, get some sort of traction going for that type of ad model. Then I’d explain to the investors, we have to move to this because the alternative is the company basically starts to be on a ticking clock. Right? And

I just think the idea of, like, let’s not talk about it is not a great strategy either. Right? You know? You have to be willing to stare irrelevance in the face and choose a different path.

It depends on the winner of what. I think Apple will make massive, massive strides forward with AI. I’m kinda disappointed how long it’s taken them, but I think

Yeah. Yeah. For sure. You have to assume Apple’s a really well run company, and you have to assume that there’s a head of AI in there, and you have to assume that you’re training LLMs, and they’re looking for LLMs that can possibly run on their hardware natively and not even have to talk to the cloud. And Apple are very privacy focused, so they’re gonna get all that shit correct. And you have to assume it’s all gonna work with your AirPods, your watch, and your phone, and all sorts of stuff.

That that’s like I would be shocked if that wasn’t the case. So then what will they win is the question. I think what they’ll win is this idea of Siri might finally become useful. Siri is currently not useful because it doesn’t really have enough smarts, but I think when Siri can be as conversational as ChatGPT and can take actions on the device, it’ll change the entire interaction model across desktop and iOS as well in huge ways. So I think Apple will win there, but that doesn’t outrule OpenAI’s success or Amazon’s success selling Anthropic or something like that or any of the other scenarios.

I just who owns consumer AI? It could well be Apple. And then in that world, what’s interesting is we might see a new type of commoditization of, like, services. So you might just say, Siri, call me a taxi, and you might not give a shit which taxi shows up. And Siri, order me a burger, and Siri, play this song. Siri’s default is gonna be to play that song on whatever, like on on Apple. If they become this really nice intermediary layer where you’re you don’t even pick the app you launch, that is a form of user selection we don’t even want in our lives.

You just say the thing you want to happen and leave it to Apple to work out what should happen, then Apple will take this new primal position. It’s like owning the homepage of Google or something like that. We get to redirect all user intent to anywhere we want.

through our service first. I think that’s true, but I think Apple will say absolutely no way. That’s not their shtick. Right? Like, Apple are always customer first. Like, another argument there, I think this is something more googly. If okay Google or whatever their thing is called, their serial equivalent is called, I could totally imagine Google doing a thing where they give away effectively free Android phones powered by the fact that they now control the intent layer, which is when you say, okay, Google, call me a taxi.

Google goes and they get you whatever taxi which comes from whichever of the highest paid provider. That’s so interesting. So you give away the whole butt for free as your CAC, your customer acquisition goals. Yeah. And you make it all back in lifetime services. That’s I mean, Google give away a search engine, and they make it back and pay its sponsorships. Right? It’s it’s it’s the same model. It’s just a different rendering of it. What

have to look at their actual business, which is, like, Instagram and WhatsApp, and it’s Facebook itself and it’s Oculus. The question I’d be scrambling to answer if I was them is just like, what does AI mean for all of these services? And what does the LLM advances mean for all these services? So I could imagine they might have a bot that sits inside WhatsApp. It could become a new way to, like, find me the nearest florist. And, like, maybe WhatsApp will put you directly in contact with them.

You can speak directly to them or speak directly to their bot. I can imagine them kinda interjecting from that point of view. Oculus has really obvious applications of AI, LLMs, infinite worlds and insane games. I take a combination of LLMs and Dolly and put on a headset and live in that world. It’s fascinating to think of what’s possible. And then facebook.com, that feels me feels like it’s like up there with LinkedIn and one of those, just like a bastions of the old Internet. I don’t know what people use it for anymore, but I think we all log into it every day, but we don’t really know why.

Yeah. I work here.

they’ve produced some really cool stuff as it relates to visuals, and I think if they play their cards right, they can actually expand beyond creatives, if you know what mean. Like, Adobe previously or heretofore has been, like, the creatives tool. Like, you have to be good at design or whatever to use Adobe, And and that’s even kinda true for their Figma acquisition with the exception of FigJam. Like, if you’re not good at, like, drawing or whatever, the tool is not that rewarding to use. I think if they play this correctly, loads of people want to draw things but can’t.

Right? Like, I said draw me a picture of a penguin, right, like most people will make a ball to it. Right? I guess people who want to draw and then a much smaller subset people who can draw, and then a much much smaller subset people who are very good at drawing. It’s an example of the total addressable market and the serviceable market and like how much of them can we actually charge money for.

With the idea of like some of Adobe’s features where you just select a bit of text, you say make the sky darker, and you select another piece and say take this person out of the photo, or you say give me a rendering of like a leprechaun walking over a rainbow, and it does all this like instantly to a pretty high degree of fidelity and like just nice polish, like certainly far beyond what an average illustrator could do, and it can do it instantly. So I think we should see a world where there are more people being creative.

More people able to like weaponize the ideas they have in their head in in a visual form than previously could. Previously, you’d just be like, oh, I have this vision in my head, and then whenever I, you know, go to pick up my mouse and open up like Photoshop, it falls apart. Now you don’t need that. Now like, you know, thinking and rendering become like this new sort of skill set that you can have. Adobe could see a significant expansion of people who care about Adobe tech if they play well, because Adobe is gonna be like, if you have an idea and you’re struggling to visualize it, just use Adobe.

Like, and you’re like, here’s what I’m trying to draw. And here’s I wanted a diagram and I want this and I want that. Oh, you got that nearly right. Let me select this piece. Alright. Edit that so it’s not like this, it’s like that, and have her face in the other direction. Okay. It can do all that shit really, like, at least the tech demos I’ve seen, stuff that, like, Scott Belsky shares and all that. It’s, like, phenomenal looking stuff. So I think, like, they could have a really significant expansion of people who could use Adobe software far beyond just creative geniuses.

Where does

I think right now, a lot of value is going straight into the infra. Like, as in we’re handing it all at the back door to OpenAI in in in this case. Right? In your first question about commoditization, if that happens and the value starts to reduce there, right, because if there’s more competitors offering the same thing, that’s the beginning of commoditization or at least you go from monopoly to an oligopoly to ultimately perfect competition where you, you know, our price goes down on each step. That that’s how I expect that’ll play it.

Unless OpenAI can continue to find mass market differentiation, which is always a hard thing to find. Right? Like like, it’s in the differentiation that all the customers care about, not just specifics. Right? So let’s assume over time, I believe that, like, just, you know, the amount of people investing in the space means that I think the infrastructure layer will gradually get a lot more AWS like. Right? Which is like razor thin margins. Well, sorry. Not razor thin margins. That’s the wrong way to frame it. Price mixed with strength of product as a competitive weapon, if you know what I mean.

Right? Like, Amazon make good money off Intercom, you know, but they’re also, like, we’re all in on them, and we’re very, you you know, we’re very committed to them. So that’s what the competitive battleground. There’ll probably be three or four big providers is my guess. Probably GCP, AWS, maybe OpenAI directly, and I’m sure Azure. That’s how how I think the that layer will play out, and they’ll end up in direct price competition with each other. Where does the value accrue beyond that? I think just differentiation.

Value just generally follows differentiation around the stack. Whoever has the stuff you can’t get anywhere else can charge the margin state, no one else can charge. That sometimes means quality of software products, sometimes it means quality of network or social network, sometimes it means, like, we’ve got the most distributors in our marketplace, But, like, whoever has the differentiation that no one else can get is the person who can actually, like, basically charge the highest price.

Correct. I generally think if you’re playing in somebody else’s playground and you’re bringing one little thing, even if you do have a reason why your thing will be definitely better persistently and sustainably, it needs to overcome a lot just to be worth the cost of an extra tool in the stack.

With AI, a lot was made at the start of like, oh, everyone’s in trouble. And I remember when we first released our first wave of AI features, which is just AI in the inbox, it was things like summarization, expansion, change tone, like, take this blunt reply and make it polite. Those sort of features, one of the big dogs, you know, use it as a showcase of, like, this is why there won’t be new startups built in this. All the value will go to the incumbents. And I think the hypothesis there is that if the value is available on top from OpenAI, it’s on top.

In our earlier example, it’s on top for Mailchimp, and it’s on top for the new startup too. Right? So the thin integration, where you’re shelling out for your differentiation to a third party, it does not create a wave of disruption. The disruption comes as we were saying earlier, like, it’s it’s when you would build the entire thing differently in a world post AI. It’s hard to give an answer overall. If if you put a go into my head, I’d say incumbents will probably do better out of AI than, like, startups thinking they can displace incumbents.

But that’s not the same thing as saying that, like, startups have no business in AI. I think there’ll be loads of new stuff in new areas that no one’s ever seen before because of AI.

I think. Yeah.

correct. I think, like, the areas where AI will deliver most value in startups will be things that just weren’t possible before are now possible. Companies like Synthesia will, like, will have, like they’ll produce massive new things that just literally weren’t being done before, and they’re, like, new categories and new jobs, new workflows, new capabilities.

I think what we’re seeing there is a growth in how do you say AI powered video customization to do lots of like things that people bear in mind, the cost of video editing and all that sort of shit is so high. You have a video that explains how your product works, and then you change the name of one of the features. Now, in the current world, that’s a reshoot. In the new world, you edit the JSON file. We’re we’re moving into a different fucking world here.

I think what we’ll see is the area of, like, AI powered, say, video creation in that case, didn’t exist three, four years ago, really, and it’s gonna be a big area is my guess.

If I was investing in, like, a classic commoner, gardener, b to b SaaS enterprise company in this day and age, I would need to be quite convinced that AI has no role, or if it does have a role, the founders are all over and I’m gonna lead. Otherwise, I just don’t

No. But, like, maybe I’ll come out of a different way. Like, if you’re building a web analytics product today, if someone’s trying to pitch me, like, we’re gonna redo GA. Everyone hates GA. Go Google Analytics. It’s gotten too complicated. So we’re gonna do a better version of Google Analytics. My first few questions, because, like, this is an area that isn’t obviously affected by AI and that it’s reporting on data points that, like, that I gathered on websites about, did the user visit the checkout page or not.

Right? But I what I’d wanna see is show me your reporting engine. Show me have considered how, like, chat UI could be relevant here, like, basically, instead of building, like, a 150,000,000 drop downs to, like, you know, query by browser type, by day of week, by this, by that. Can I just type the query I wanna see and can you generate it? Can in fact, can you go one step ahead of me and suggest new things? Can you use AI to like, power your reporting engine to suggest new reports I should be looking at that I haven’t seen before?

I’d wanna be confident that the founders have thought through that and come to a conclusion, and it might be we’re not gonna build any of that, or it might be, you know, we should really go hard on natural language as a query engine for reporting. That’s like where I’d be like, yeah, okay, I’d wanna see that. And then I’m I’m kinda making this up on the fly, but like, if they said to me, no, Des, we don’t think AI is gonna change anything to do with how people query analytics tools, at that point, I’m probably leaning out.

I I don’t know how you can be so confident in that, and if I’m honest, I actually think it will.

you look at your deployment today, has it changed much? I would say I’ve probably, like, the last five or six checks I’ve been I’ve written have all been into companies that are either pure AI or heavily AI infused.

it’s about I’m probably more b to b SaaS exposed, so maybe maybe I I’m probably more like, 60%. But, yeah, if the question was, like, how much of it mentions AI in, like, sub bullet tree or whatever, then, yeah, a 100%.

being

Yeah. Here we go. Yeah. Yeah. It comes back to this, like, does the product make more sense with AI or not? Like, I think what we saw over the last three, maybe four months has been a lot of companies. This is what that GIF I tweeted or that video I tweeted was about. You’re seeing a lot of people doing really pointless shit with AI. Kinda reminds me of the early days of JavaScript, people did all these fancy effects that were totally pointing to the scripts. They used libraries like Scriptaculous and all that like to do just fancy like, oh, it fades away, you’re like, why does it fade out, you know.

There’s definitely bits of bullshit you’re gonna see, which is people reaching for the easiest to use features and trying to like reverse engineer a demand for them inside their product. Summarize this shopping cart in natural language, English. You’re like, well, it looks like the user is buying a robe and it’s a size me. I was like, who gives a shit, you know. So like there’s a lot of that happening where in the scale up space where I think it’s somebody looking to, like, wave a flag or get the, like, marketing juice that comes with saying you’re now something AI.

What I look to to sort of interrogate the value is just, like, is there something happening now that wasn’t possible before, or is there somebody who doesn’t have to do something now that they had to do before? If the answer to either of those is yes, then I’m sort of bought in. If it’s basically like, you know, oh, you know what? It’s really, really important that people like, whatever, like, you know, summarize the Google Doc in real time on the right hand side as they’re writing it.

I’m

To leave any sort of lasting impression on somebody, I think you have to say something that is like unique, as in like, can’t be the same thing that everyone says, valuable, as in it has to actually matter to them, and simple. Right? They have to be able to get it in one read. If you can speak to your users in a way that presents some value that they haven’t heard before, that they can only get from you, and that they can understand, then generally speaking, you tend to, like, resonate with your users.

And there’s a lot you you can’t just walk in on somebody’s life and just start sharing stuff with them. So you need permission to speak to them, and that’s where you lean on thought leadership or, like, make yourself somebody worth listening to. But I think most businesses fail to say something that’s either they fail on one of the three tests. It’s either not unique, it’s not simple, or it’s not valuable to them. So you see things like inbox reinvented. You know, like, I wasn’t really looking to reinvent my inbox or it’s like forget everything you know about blah.

So didn’t know anything about blah. You know, so you have a lot of like this like marketing speak, if you know what I mean, like which is just people just lean for like their marketing veneer or polish that you get, and it just gets a bit too salesy. And in doing so, it loses its specificity and it loses its perhaps bite or its uniqueness or whatever.

like, jacked about this. Excited to announce our our, like, ongoing relationship with. You know? It’s like I think there’s a, like, owner or CEO. It’s just so he’s, like if I could ascribe him one defining quality, it’s probably authenticity, but I think it’s just speak the words you wanna say. That’s the thing that is, like, so important to have a, like, credible marketing is that you say the things that you would actually say. You don’t put on, a marketing hat where you rattle off a lot of bullshit, you know what I mean.

Most startups with marketing with B2B or B2C, they they forget to say the thing that makes them unique that’s valuable to the users. And if they’re going up against an incumbent, the worst thing they can say is, like, oh, we’re like a new type of inbox, blah blah blah blah blah. Because they say, oh, I put you into the same category as the incumbent. You need to say, here’s the things that we’re different. Like, here’s the reasons that we’re better than blah. And, like, that’s usually hammer his home.

Like, again, simple so they understand it.

Simon’s point’s not bullshit. It comes down to what type of marketing we’re talking In my head, I’m talking about product marketing here, like, I’m talking about, like, what does the homepage say when you go there. Right? If you go to Asana and, like, it opens up of, one day, Dustin was struggling to manage a project at Facebook, and he really wanted to get an overall view. You’re like, dude, I’m closing that top. What the hell is the thing you do? Now, when it comes to consuming the overall brand, if you like, you know, like, what is this whole company about?

What is its raison d’etre? Which I think is more important in certain industries and verticals. You’re like, hang on. If I’m gonna, like, really align myself with these people, I need to understand what makes them tick. That’s when you’re going to, like, slash about or slash vision or whatever to sort of read a bit more about what the hell is going on here, like, in, are you guys just growth hackers who stumbled into a project management company, or do you actually care about the nature of human productivity?

Simon’s point’s not wrong, but I think most of the time when you’re trying to sell software, the best thing you can do is probably say what the software is about and not talk about the origin story of or whatever.

If I could change anything in terms of things that I would trust myself to do a better job on, I would probably try to more tightly align the audience we were building with the product that we were selling.

Correct. Yeah. It was large, but it was then harder to translate that into demand that you could capture. Yeah. Qualification was tough. Yeah. Even yesterday, was at a conference here. Dude, I read the Intercom blog. I’m like, what do you work on? Like, you know, our program manager on the infrastructure team. I’m like, you’re not gonna buy customer support software. I think probably I prioritized growing an audience for the blog over growing the right audience for the blog. And the net net is that we have a large audience for the blog, people who like, design or into product or even into engineering, people who are maybe into, like, startups in general, like, even people who into marketing.

What we didn’t go deep on early enough was just the actual, like, sort of people who are best positioned by Intercom.

It’s a different approach. I agree with you. Like, as in I agree, if you have a million readers a month and, like, 1% of them are, like, your buyer or if you’re, like I can’t do the math on the fly here, but if you have, like, 10,000 readers a month and, like, and whatever, like, get some the same percentage that nets out the same numbers total, they’re both following strategies. Like, a small focused audience or a big unfocused audience that contains in it, the small focused audience, that works.

Like, I mean, you’re asking me what would I do again, I think I probably would have just tried to be a bit more focused early on. An interesting thing we measure, like when we look at say stuff we share on social, it’s easy to get like a viral tweet these days. What does it drive? So that’s thing, not a lot. Right? Like, it doesn’t drive a lot of follows, doesn’t drive a lot of anything other than like a few like quote tweets and a few retweets or whatever, like, you know, hundreds or whatever.

But what’s interesting to me is we’ve been having popular viral moments. The thing I care more about and I’m more excited about is, like, when we share something that’s pretty I don’t wanna say boring, which is pretty focused on, like, the only people who give a shit about this are support people. The people who engage with that content is growing, and that’s the thing I care about more than, say, just general followership. I’m like, effective followership is the thing to look at.

Yeah. Yeah. And like the reality is if Becks enters investing, his followership isn’t that useful to you. No. You know what I mean? It’s not at all. Yeah. See, if I tell you that Beckham wants to be on a podcast, don’t want

That’s a tough question. I think I probably would have taken integrations more seriously earlier. Understanding more about Intercom’s role within the tech stack, knowing that we’re not the only show in town, understanding that we have to play nice alongside, like, workforce management solutions, CRMs, whether it’s like Salesforce or HubSpot or whatever, QA tools, Slack, you name it. I think we could have gone harder earlier on that, and I think it probably would have been more valuable to our users.

I mean, you’re literally asking me, like, a question I I can’t answer the the decision.

I think there is very little best practices advice in startups that actually survives more than five or six years in my opinion. Because I think everything is just constantly evolving around us. Like, even like all this AI shit’s gonna, like, upend the load or, like, product strategy advice advice that has been holding solid for ten years. Right? Most of the time, advice is expires. How you go to market? If you’re in 2012 and you’re talking to, like, VCs, you know, senior experienced people about what’s the best way to go to market.

What they’ll all tell you is basically hire a sales leader and you kinda run up market and get away from those scrappy little companies. Right? That was the sort of norm in all the blogs and all that back then. Was just like, build a sales machine. Hope you go. Right? That’s how DocuSign did it or whatever. Right? That’s how forbill.com. Yeah. Yeah. Exactly. Boxers, somebody from like from that just slightly previous era. Around twenty thirteen, fourteen, fifteen, people started to realize Dropbox was actually doing well.

They actually had managed to do it without having to go like all enterprise, but the big dog in the story here is obviously Atlassian, who presented an entirely different way to go to market and build a massive business without running up market. If you play it out maybe in parallel, I don’t think it’s necessarily sequential, but Shopify, say, example, are still as relevant if I just wanna sell this bottle of Fiji Water. Still probably to go to, but also they’re a massive business. Right?

So they’ve managed to, like, not have to turn into, like, how just like the, you know, the old school big commerce company where, like, you go and you have, like, nine people show up to do solutions engineering and plug in your stuff.

approved and exactly that there’s massive businesses to be built without ever having to go, like, all the way up marker. Right? The advice is almost out of date by the time it’s best practice.

When I think about, like, I if had to start another company tomorrow Yeah. I think this stuff that I would definitely bring into the next company is not specific, like, go to market tactics or, like, bottoms up adoption versus, like, going into the top of the market. It’s not like marketing channels or tactics or any. I think what it would be about is, what is the makeup of a senior executive team, and what do you want? What are the attributes of leaders at different stages of the company?

How to interview? How to, like, give the right feedback to people? How to assess where a hole is in your organization and how to deal with how to, like, kinda, like, lose yourself defending a narrative, but instead, like, you know, take reality, face it straight or whatever. I think they’re all the sort of things, the more timeless pieces that I would learn. Like, I think it’s reasonable to assume hiring executives is going to be a thing. It’s reasonable to assume building a company culture is going to be a thing.

It’s reasonable to assume having a mission and a vision on the values by which your business performs. They’re all kind of timeless skills, and I think serial entrepreneurs know the mistakes they made and not being firm enough in the right areas or flexible enough in other areas or whatever. I So think that’s the sort of stuff that you’re investing in. They they won’t make that mistake again.

Yeah. I mean, it’s a fair question. Like, is managing managing humans the same today as it was five years ago? I I probably not in a lot of ways. Yeah. And, like, definitely, generationally, I think it’s different as well. But, like, I’m not trying to be evasive here, but if you just take one step up, I’ll just talk about it as in, like, design a culture that attracts the maximum amount of talent and motivates them to work as hard as possible for the climate you’re in, for the environment you’re in, for the country you’re in.

Yeah. It comes out, like, we already trying to weigh talent here or not. Like, I guess, like, what you want But I don’t think

Yeah. Well well, specifically, would say you’ve got people who are early in their career, people who are, like, seasoned, super experienced, super tenured, brilliant. We have interns. They’re not yet a players. But, like, to your question, like, is it a maximum quality, or is it a maximum quantity? Like, I guess what I mean is that you’re not trying to maximize headcount. And I think for us, right, that’s another thing that, like, most seasoned entrepreneurs won’t do a second time or third time, whatever. They won’t see headcount as a vanity metric at all.

The second thing I’d say is, Macy, you wanna attract the, like, the smartest folks you can at the right level. Like, you know, you might you don’t need, like, a thousand of them in an early stage job. You need four to start. And then as you start to scale your culture, you’ll realize, alright. We can’t only hire, like, a plus plus superstars because we’ll be waiting all year for our next engineer. So you start to work out ways to, like, scale, and you say, you know what?

How about the a plus plus person leads the team, and they can hopefully drag the b person up to an a plus standard or whatever? So you start thinking about it from that point of view. What was the biggest mistake you’ve made on culture? I would say radical candor is something that, like, I struggled with for a long time, and I probably still do. I’m better at checking myself, but the idea of, like, telling somebody what they need to hear so that they have the best chance of succeeding in their own career versus telling them the thing that will make them happy.

Society struggles with this in general that, like, you know, are you ultimately looking out for the the greater good for the person, or are you just optically looking out for it in the short term? And I think my version of that would be earlier days would have given, like, nudge and prod feedback, but, like, what I probably need to do is to say, like, hey. I don’t think you can actually make it as a leader of blah until you can blah, and here’s the skills that I just think are close to a terminal gap in your skill set.

And if you don’t close them, I just don’t see you either surviving here or honestly anywhere. And I think you need to invest and take me seriously in this. Whereas I was probably twenty fourteen, 2015, I was probably more likely to say, I know I actually can make good progress, but here’s a couple of areas where I’d like to see a little bit more. You know, She was ambient. Yeah. Well, it wasn’t even just ambient. Was just way too cushioned. Like, I remember, Lauren, our head of people, saying to me something along the lines of, if this person’s not, like, really upset with you at the end of this, you haven’t actually done the job.

If this doesn’t sting for them to hear, you should be suspicious because because it means you’ve definitely pulled the punch. That was good advice at the time because it’s just so much easier in a conversation. But but, you know, everyone’s gonna go good at par Or just not to have long term Yeah. I mean, that’s kind like, the cost of entry for startups because you just you don’t have unlimited headcount. I think bigger companies can can struggle with that. But yeah. Any big lessons on riffs?

Oh, no. I they’re horrible. Here’s the lesson. Don’t put yourself in a position where you have to do them if it’s at all possible. It’s just really fucking hard to get through. It’s a it’s a big time sync. It’s a big emotion sync. It’s a big engagement sync, and it takes a long time to sort of, like, plan, perform, and doing it all. You know, the lessons I’ve kinda learned on Riffs are more from other companies than anything Intercom has gone through. The last twelve months for startups in general, every other investor update I get has been, oh, we’ve done a little Riff.

We’ve got you you know? But things are generally going good up and Yeah. Yeah. But, like, honestly, in a lot of these cases, superhumanizes feature where you can if someone mails you, you can see on the right hand side all the other mails they sent you, and you only need click back two or three words like, we’re excited to announce that we’ve hired not one but two VPs of partnerships or whatever, and you’re like, oh, you didn’t realize that you’re planting the seeds of the rift there.

You know what I mean? Like, I think, like, the bigger thing I’d say is, like, over hiring. There’s an electricity and a spark. In the same way I say you always want your office to be a little bit too small because it gets you this sort of vibrance and this kind of sense of hustle. I think you always want a team to be just a little bit like like there should always be a pile of work and like there should always be a sense of like holy shit, we’ve got so much to do.

I think you have to put it on the Federus at that point. Like, as in, if someone gives you the money, you don’t have to take the money. If you take the money, don’t have to spend the money. If you spend the money, you don’t to spend it on hiring. If spend it on hiring, you don’t have to spend it wastefully. Like, it’s a bit of a I get I get

a really disciplined run. Yeah. Well, you know, hiring fluctuates between, like, needs and coulds. Right? Like, and and and and maybes. It’s easy to lie to yourself and say, like, the existing, you know, the incremental thing that this person’s gonna do is gonna be transformative. I would bet your second time founders, they hire slower is my guess because I just think it’s an easy way to, like, burn up opportunity and ultimately lead yourself in a position where you might have to riff or, like, in a lot of these companies, they, like, they’re riffing people where they didn’t

Yeah. Yeah. Delayed emotions. I think it made me more efficient in the time I’m at work for sure. I knew that I wouldn’t be able to work as much when I went home or whatever. And I noticed this in, like, when I’m on, I need to be on and I need to be, like, emptying the tank of my a game because at some stage, you’re gonna go home and have all these like kids and I have to deal with them. I can’t just come home exhausted and I I go I need to be active when I get home too.

So I think that’s probably the first way. I don’t know if I got more patient necessarily. I mean, I definitely need more patience with my kids. I think the folks I work with, they don’t antagonize me that often. Patience isn’t the string I pull on a lot, thankfully.

Oh, I think Google because I don’t wanna fix Google Meet. Like, yeah, it’s honestly, I just I if it’s because if I have to go back to my job after a day and I’m gonna go back to my normal life, I think Google Meet is definitely the product I really wanna fix up.

It’s not that. It’s the chat bar on the side. Like, basically, if I was the CEO of Google, what I’d do is I’d say, I need you guys to build me a Slack, a fully functional Slack equivalent that also sits in the sidebar of all these things and has to work also at Google Docs. So when I’m having a Hangout with with Harry, if we’re chatting, it should be the same thing as persistent chat channel between me and Harry. And if I start a new Google Doc, it should be automatically shared with Harry.

All of these things just feel like such donkey work that I have to go through every single time I jump on a call. You know, if I was trying to solve an actual a realistic tech problem in one day, I think that’s probably thing I’d go for. I’d love to go to Spotify and basically say, like, we’re doing a new thing called Spotify Live where we cover all soccer. I think, basically, it’s Spotify for soccer, like, is in the idea that there’s a single place you can go to watch all live streaming sports sitting right there, and I think it’s just everyone’s scared shitless.

Well, you couldn’t because of rights. Well, that’s I gonna say, but, like, wasn’t that true for, like, iTunes and music as well? Right? The rights thing just seems to be like a deal that just needs to be done. I do think, like, Spotify managed to negotiate such that they can play podcasts and music for all over the planet, from all over the planet, Getting like Plymouth Argyle to agree that their matchable leads should be like livestreamed. I think it should be doable.

The right way to view competition is with respect and after you’ve launched and you’re out the door, increasing comprehension. Comprehension meaning you have to understand why they’re successful. It’s very easy to like their product shit, but maybe they’re just brilliant at go to market. And if you don’t understand that, you’re never gonna be like, why is not even paying attention to me? I’ve got the better product. Respect always. They had they have the market. You want the market. You have to respect the fact that they got there.

The comprehension piece, just ever likes to quote, like, Jeff Biz Stone, sort of like, oh, I ignore my competition. I just listen to my customers. Do listen to your customers. I just don’t think ignoring competition, especially if you’re, like, in a market where there is competition. I think it’s essential to work out, like, why do people pick them over you? Is that not a useful thing to know? You can’t listen to your customers if they’re using

competitors.

Yeah. Yeah. As long as you believe that beating them is the right prize. That that makes sense. But, like, oftentimes, you can find yourself ruling over a graveyard. If we just beat Zendesk, but Zendesk is, like, in in the current state it’s in, that won’t be the full realization of what Intercom won’t be.

I don’t really know because I don’t know the inner workings, but certainly the PE cash cow playbook would look like raise prices, shrink road map, bring the register, and package it up into a very efficient cash producing machine and hand it over. I’d have no idea where they are, but, like, but that’s what whenever anyone else is gonna pee, that’s usually what seems like.

did Owen coming back change culture? It changes substantially. It’s also it’s Owen coming back. It’s also, like, just the new era that of, like, of where startups are. We got hyper focused on a very specific use case. For the longest time, Intercom had been kind of a broader tool. When Owen came back, were like, well, let’s go all in on on one thing. Let’s set a very clear, like, focus for literally everything we’re doing. You know, increase the urgency. We kind of we had this initiative called, like, p 52, which is, like, fifty two weeks to, like, to basically rebuild all parts of the company.

And it gave us a kind of ticking clock, and it gave us a very clear sense like, progress and momentum and engagement. We’re just finishing it now. We went gung ho as, like, the team of 900 folks towards rebuilding, like, marketing, branding, pricing, product. You name it. By the time this goes live over the next five, six all of that’s gonna look different. Yeah. Why did you make Bobby at equals write a blog post a week? Oh, yeah. So, basically, Bobby asked me to invest, and I said I will, but I was scarred by too many businesses who forget to tell people what they’re all about.

What I mean by that is a lot of companies launch, and they kinda have this kinda weird incel like vibe of like, why isn’t anyone paying attention to me? And an investment I’d made had just finally wrapped up and it died. It was really frustrating me that they had kind of forgotten to launch. And to me, launch isn’t like, oh, we’ve put up with like an alpha page or something like that. Launch is like when you’re laying it all out there. Here’s what we think. Here’s our vision for the future.

Here’s our product. Here’s how our product matches our vision. Here’s where our product’s going. Here’s why you should use our product. Right? Looking through my own portfolio, and I was like, lot of these companies that died never really launched. I was like fearful in that period. And I said to Bobby something along the lines of like, what’s the plan for telling the world about this? Like, I mean, simply having a better excel. You could find yourself just sitting at the bar wondering why there’s no one paying attention.

What’s the actual plan for putting yourself out there and being in great shape? He’d talked through some aspects of marketing. You know, branding positioning, like Google Sheets is over here, Microsoft Excel is over there, blah blah blah. And I was right. But, like, you probably need to tell people that hundreds of times. Yeah. And I think I wrote 90 or 95 of the first 100 blog posts on Intercom. And if I’m honest, I always felt that Bobby and to some degree Ben, his cofounder, never really gave me any street cred for that.

You you know? But so so some of this was like eat some greens. You have to commit to tell the world everything you stand for your product, what it is, what it does, how it’s used, who uses it, why they use it.

Well, totally. And your thinking will evolve, and that means that something I wrote in 2013 is no longer what I think today. Fine. I don’t think anyone thinks anything ten years prior is, like, admissible in court or whatever. You know? It was ultimately it was like, please guarantee me that you’re gonna lay it all out there and tell the world why they should use equals. That that’s ultimately what they ask us.

Yeah. I can only assume that they might have realized earlier the world was gonna return to a lot more, like, normality. And I think they did smart work with, like, say, StreamYard in in that acquisition. I don’t know if it’s a lesson, but I think when you have, like, an existential crisis along the lines of, hey. It turns out everyone’s just going back to real world conferences, and they don’t even wanna livestream it that much. There’s a lot of temptation to, like, push back against that, a hedge against it, like, or, like, you know, just be like, oh, no.

No. Surely they will. Or there might be another pandemic. You never know or whatever. Whereas I think sometimes you just need to say, hey. Look. Thing we thought was true is no longer true. We now need to reorient the entire company to get ready for the new truth. Streamyard and that that move was that generally part of that. That’s how I saw it as an outsider. I was like, they had to, like, ultimately pivot to, like, live streaming a lot quicker, and they did. They made a great acquisition.

Here’s a shit lesson that’s just true. Right? If you’re in the weirdest capital market in history and everyone’s telling you to, like, raise billions of dollars at, like, ridiculous valuations, don’t do it. Like, it’s a weird lesson. Right? Like, it’s a hard thing, like, you know.

Yeah. Yeah. I think that’s true. I think, certainly, that’s for founders who occasionally write checks. I think that’s probably the bigger lesson that will emerge from 2019 through ’23 or whatever.

I think Zendesk will have their own faith. It’s not incumbent upon us to, like, do anything to them. I think Intercom, we want to be, like, the primary support solution for, like, Internet businesses. That’s where we wanna be as to where I am in that. I hope that there’s either less intensity or I found a role that in the company that requires less intensity because I don’t see myself as a 52 year old being able to do what I do today. And even if I look at myself today versus, like, when I say 30 when we started or whatever, like, I was wasn’t in this healthy shape or whatever, but, like, do it there.

There’s definitely more energy and more, like, 9AM till 2AM type energy. Right? Like, that has definitely shrank a bit. Like, I’m now probably working closer to a normal working day. I hope whoever has my position, if it’s not me, is able to do it. Or if it’s me, I hope who doesn’t need the, like, nine to nine energy.

Thank you very much.