AI Pricing Masterclass with HubSpot Co-Founder Dharmesh Shah

We made a very pivotal decision. AI was not an add-on product that you bought. We can't say that we are going to be the AI-native CRM company and then say, oh, and by the way, AI is something you add on. That's like Elon coming to us and is like, we do EVs and you have to buy the battery in order to get the electric features of a Tesla.

Today my guest is Dharmesh Shah, co-founder and CTO of HubSpot, the $18 billion company with 270,000 clients. He has been way ahead of the AI wave with initiatives to drive what the new AI-native CRM looks like.

We're going to have a hybrid pricing model that is partly seat-based and partly consumption-based. I expect that over the fullness of time, credits will become the predominant model. Will AI kill SaaS? So I think what's going to happen is when Salesforce came along and introduced the idea of SaaS, he didn't literally mean no software, but he meant software as we knew it, as it was sold, as it was deployed.

In our conversation, we get into why Dharmesh believes That credits will replace seats as the dominant pricing model. The specific criteria to use when AI bets are worth making, and how HubSpot is navigating the AI transformation from the inside.

If you are finding that you're struggling making the AI transition, you don't solve that problem with an innovation bet. You solve that problem with a team change or a leadership change or an org change or a culture change or all of above.

If you like this podcast, make sure that you click the subscribe button in Apple Podcasts. Spotify or YouTube so you can get more updates with the one and only Dharmesh Shah. So how you been, man? I appreciate you jumping on this.

I've been well. It's been an exciting and challenging year for most of us, I think so.

How is it challenging? I see nothing but excitement from your post.

Yeah, I'm an optimist at heart, but I think the challenging part is the entire industry is going through a transformation, the likes of which, you know, the closest thing I think we've experienced is the dawn of the internet. And so with that comes a lot of change, a lot of introspection, a lot of, you know, what are the right things to do for our customers? What's the right thing to do for the business long-term? So yeah, it's all fun. And it's like, I've never had, I'll say this, I use the word challenging, but I've never had, you know, HubSpot's almost 20 years old now, 19 years old. I've never had more fun in the history of the company than I'm having right now.

That's amazing. That's amazing. Do you feel a little lonely without Brian being on your side? How is it, founder to founder, this is—

Yeah. Yes. And for the longest time, because Brian had kind of talked about years before he stepped into the chair role that he was going to want to do that someday. And I would just refuse to talk to him. I was like, talk to the hand, talk to the hand.

That's never going to happen.

I don't even want to hear it.

I have no idea what you're talking about. This is not the Brian that I know.

Yeah, but I mean, he's still very involved. I see him in board meetings. We chat all the time, still exchange late-night emails like we always did. And it's, yeah, I totally understand sort of from a life stage perspective, you know, how all that came to be. But anyway, yeah, but I miss him in the, some of the kind of day-to-day, you know, blocking and tackling, but I think he's Yeah, still very involved in the company and still, yeah, I'm still a good friend and I'm about to move. You know, I've been living in the outskirts of the Boston area in a Boston suburb, moving back to Boston. I'll be 2 blocks from Brian, so he won't be able to avoid me.

You're joking. Is that, is it for HubSpot or because you, a life change or what makes you bring—

My son is starting high school and he's going to a high school that happens to be in the Back Bay in Boston. So I use that as his an excuse. Kids are why we moved, or a kid was why we moved out of being in Boston, and the kid is now why we're moving back, but I'm looking forward to being back in the city, so yeah.

That's awesome. Yeah, I love Boston. So let's start thematically. Let's start from the top. OpenAI released the Strawberry, had the Strawberry release, I think it was November, maybe December last year. It was the O1 model, right? That birth reasoning, To the scene. What has gone through your mind since? Like, since you saw the model all the way to now, what's exciting? What's worrying? What's the, let's see what happens? The full Dharmesh load of this.

Yeah, and I'm thinking back, so I've had multiple kind of moments of wow with GPT at large, back to GPT-2 that Sam gave me access to before that it was a ChatGPT app. So I just had the API and had to code my own minimalist chat app, and I had a conversation with it. And even then, this was 2 years prior to ChatGPT launching. It was amazing. And you sort of had this, and I was late night emailing Bryan. I sent him the transcript of the conversation that I had, which I still have for historical purposes.

What was the context of the email? Look what we can do now.

No, look what this thing can do. It's like, this is an unedited transcript of a conversation that I had with AI, right? Like, I just laid it out there. I'm like, this is this is going to be big. And it took a while for the world to catch up. And I think, and then when ChatGPT launched, everyone sort of figured out what I had seen a glimpse of, which is like, this is real. And then we've all had our kind of magical moments. Like, I can't believe software can actually do this, right?

Right.

So that was kind of moment number 2 is sort of witnessing. And obviously ChatGPT was, and GPT-3 was much better than 2 was. 2 was very rudimentary. And then so 3 sort of awakened the world to what we might now call this kind of single-shot prompting, no reasoning, but still very, very good at kind of building this world model and tackling a broad range of prompts. What reasoning changed, at least in my head, is that we had always thought, we collectively, the industry, or at least most of us mere mortals that were not working at the frontier labs, were like, okay, well, I sort of understand the nature of scaling laws that says we can throw more compute and more data at this and it'll get better. And reasoning, what that changed was that, okay, well, not all of this kind of learning necessarily happens or the compute investment has to happen in pre-training. It could actually happen post-training at inference time that says, okay, if we just give the model with the existing data that it has just time to sort of self-reflect and quote unquote think, it can come up with higher quality answers. And that was groundbreaking, both in its simplicity and elegance. And we can talk a little bit about how it actually works, but then it sort of opened up an entire new dimension of what can be done with these models, right? So now we're not limited simply by how much compute we throw at it in pre-training and train the model for hundreds of millions of dollars and go on for months at a time. It's like now we can say, oh, we can throw inference-time compute at it. And what to me that unlocks is that That's something we can sort of dial up and down as well, right? And we're starting to see that in some of the newer reasoning models that says, okay, I want you to overthink this particular coding problem, or I want you to dial up the reasoning, use more tokens.

Yeah, you have the think fast and you have to like go for a long time and go think a lot about this. Yeah.

And so I think what— and from like a business perspective, I think the thing that it unlocks is that it can now do reasoning. That means it can do a better job at what I think of as like functional decomposition that says, can I take a larger order task? and kind of have the model itself say, oh, here are the things I need to accomplish in order to get that task done. It can have effectively a short-term memory to keep track of what it has and hasn't done and kind of cycle back and can follow multiple paths in parallel to say, here are 4 ways to solve that. This one seems to make more sense. And it's all doing this sort of, quote unquote, in its head. And that unlocks a bunch of business use cases that previously weren't possible with just a pure model reasoning model. So, I'm very excited about what that does.

So, when you think about the things that you're excited and things that you're worried about and the transformation that SaaS is going through right now. And like you're living in your own skin at HubSpot, but like it's just SaaS at large. What is this ability to reason now and inference time compute brings to bear in terms of jobs, things that we can bring to the market, threats and possibilities?

Yeah, a few thoughts. One on the jobs front, I tend to lean towards the kind of optimistic side of it in terms of technology. I know AI is different from prior waves that have come, but there are some some similarities. So, I think there's going to be some disruption because AI is going to be able to do more and more. We're moving from relatively kind of low-level tasks to we're going to move towards higher and higher tasks. And we're seeing this already. It's moving very, very quickly.

Yeah.

So, I think that's definitely going to have an impact on the job market. The optimist in me says that AI will kind of giveth more than it taketh away in the kind of grander scheme of things, which is, and we've seen this with automation and technologies in the past that says, okay, well, there were roles now that are just unnecessary because those tasks can now be done by technologists, even in the pre-AI world. And we sort of accept that. I think it just sort of elevates humanity at large to say, okay, now we can take kind of our specifically oriented brains and apply them to things that are more creative, less of a grind. And I think increases human potential. So that's the optimist in me.

Yeah.

And as a software entrepreneur, what it does from a business perspective, which is the reason we should all be excited, is that we can now do with AI, using software, things that we never imagined software would be able to actually do, right? We've always built tools and we've said, oh, we can increase the productivity of X role by Y percent or something like that. And that's great. And we've built lots and lots of market cap for lots and lots of companies and driven lots of productivity for our collective customers. But now it's like, It's a whole different game, right? It's like, okay, well, it's not just about increasing productivity by a marginal percent, or it's like we can sort of reimagine what we can do with software, which is both exciting from an opportunity perspective, but then also, and there are questions that float around around, oh, does this mean SaaS is dead, right? Because AI is going to just basically be able to do everything. People are gonna vibe code their own applications. Why do we need software companies?

Yeah.

that kind of line of reasoning, which I have to—

So I actually love your post on this where you said like one, and I like this one thing where you said the reason we're not doing this at HubSpot, even though we can, anybody can, is focus.

Yep.

Is that by, focus means that there are some things that you decide you're going to do and do very well, and some things that you decide you're not gonna do, you're not gonna do very well. So you're not gonna code your own HR application. Or payroll. Turns out that that could be a little hard and like a lot of lives, you know, depend on getting paid.

Yeah. And even so, let's assume like even if we don't go after the big ones or even the hard, like payroll's a super hard category, regulation, things like that. Even the easier ones like expense tracking, project management, pick your thing of choice.

Yeah, yeah, yeah, yeah. Project management, right, right. Collaboration, any of those things. So yeah. Yeah.

And I think I put this in the LinkedIn post. I was proud of this line, so modestly I'll put it back out there, which is, Companies should be focusing all available calories on things that create customer value.

Yeah.

Period, right? So if you start with that principle, it's like, okay, well, yes, we have the engineers, we have the wherewithal, LLMs are definitely good enough to build relatively robust applications now, but just because we can doesn't mean we should, whether it's a low-level app or it's like, okay, well, if there is an existing SaaS application that mostly meets our needs for whatever thing we're trying to get done, we should 100% of the time be using that application. Now, if it doesn't exist, I mean, there's definitely reasons like, oh, we need this particular thing automated or this thing, this tool that it's too small of a problem. There are no real legitimate software companies out there that solve it, fine. But if a software application that you can buy off the shelf that gets you even like 80% of the way there, I think it would be foolish to try and replace that.

I think every startup out there is just cheering you on. It's like, yeah, go Dharmesh. Just say it louder for the people in the back.

Yeah, so on a related note, we should talk because it's just talking about the business aspects and the impact of AI. There's been kind of discussion around, will AI, you know, kill SaaS as another kind of related theme? And I haven't written a post about this, so this should be next in my lineup. Is that—

I can see it in your head. So let's get a preview of that post.

Definitely in my head. And so here's, if we go back and you've been around enough to see this, when Salesforce came along and introduced the idea of SaaS, And built a massive company, super successful, and all credit to the team and to Benioff, because it was truly visionary. And it was interesting. So at that time, you might recall, they had this, they put on buttons and everything like the no software thing, right? They had the word software crossed out with a big white button, red slash go through it.

Yeah.

And both Benioff and others knew that that's, he didn't literally mean no software, because obviously Salesforce was a software company. what he meant, and that the shorthand was software as we knew it, as it was sold, as it was deployed, no longer makes sense. This is the new world. SaaS is the new world, right? And he was exactly right about that.

Yeah.

So, I think what's going to happen now— so, by the way, what that changed was, oh, instead of running things on-prem, you're going to run it on the cloud. As a result of that, we're going to have multi-tenancy. As a result of that, the business models can actually be different. We can charge differently for software than we may have historically done with a you know, large enterprise license and a 20% maintenance trailer, right? There was all that that existed. And so there was a set of kind of dominoes that fell together as a result of that kind of movement to the cloud. And so we all know that, and it's been a great run for software companies. But the reality is with the introduction of SaaS, what did not happen was the end of software, right? What happened was the blossoming of software. It was even better times for people to be building software companies. It just needed to be built on this new model.

Can you make the argument Dharmesh, that SaaS brought software closer to the customer, or did it really do nothing in that respect?

Oh, it definitely brought it closer. I'll give you one tactical example, you know, 'cause I've had, you know, my first startup was, you know, pre-cloud, pre-internet. And the way we had to sort it, and it was in the enterprise software category. So we would build products, we would send them on CDs and floppy drives even, you know, back then. And the hardest thing as a software entrepreneur, as a developer, is that You could ship a product out there, and you had literally no idea what percentage of your customers had installed the latest release that you had deployed.

That's true. That's true.

What the usage was, is like, does Susie use it, you know, 5 times every day or never?

Like, we had no clue. Yeah, the software didn't call back home to report itself, right?

Exactly. It had no way to phone home, right? And then we got, you know, some instrumentation back over the fullness of time, but that was effectively the kind of the standard model. And once we went to the cloud, it's like you're shining this, you know, bright light, onto your customers and what they're doing and what they're struggling with. And we now have such kind of rich visibility into what our customers are doing. So just by that alone, I would never trade SaaS for the older model, put everything else aside. Let's say we didn't have multi-tenancy, didn't have, you know, but just that alone was worth the shift and it was transformative as builders.

I think that point of like understanding how people are using the software and being able to see it, with your own eyes. I think that is probably the most important contribution of SaaS to bringing software closer to the people. So what happens with agents now? What happens with AI?

Yeah, I think we're gonna see effectively a similar thing, right? And we may not be running around with labels that say, you know, SaaS with a red line through it, but effectively I think what's happening is that we're going to be working at an even higher level of abstraction as builders, right? So before we've had, And the languages have gotten better over the decades that I've been at this for 30-plus years now. And you see, we went from assembly to C to C++ to 4GLs and all, and we've had this kind of movement where we can work at higher levels of abstraction. AI takes that up a massive, massive notch, right? So as builders, we can say, and I mean builders now more broadly, it doesn't just have to be programmers, it can be builders of any sort. can sort of take the ideas that are manifested in their head and with a relatively high degree of fidelity, translate that into like working product for some definition of working, right? It's like, okay, that wasn't possible before. So now we can prototype quickly, we can iterate more quickly, we can produce products and even launch products more quickly than we had before, right? There's a step change already. Even if let's say tomorrow something happened in tech, this is the complete sci-fi scenario, And it's like, the models are not going to get any better. What you have is what you have. Everything's locked down now for whatever reason we've all collectively decided. We have enough technology in our hands as builders and software companies and people of the world to last us another decade, if not longer, with just what we have, just to absorb the impact of the model.

I agree with that. But you were talking about, so the scenario you're talking about of like the higher level of abstraction where English is the new interface for building.

Yeah.

Sort of, you know, it speaks a lot of vibe coding and you just posted on LinkedIn that there will never, there has never been and it will never be a vibe coded piece of software in the HubSpot core code. Is that a point in time statement or is that a forever statement? Like, feels a little declarative and final.

Yeah, no, it's, so here's the, thing I actually meant. I'm gonna have to go back through and edit it because it clearly came off unintentionally.

They gave you shit for that.

I'm not suggesting— so we have VibeCode, or, you know, code built through AI coding agents in HubSpot today exists. I'll say, just to make it more accurate, at least 99% of HubSpot engineers right now use, you know, VibeCoding, Agentic Coding on a daily basis, right? So this is not something that we're It's like, oh no, this could never work, even including our most senior engineers. So everyone's using it and it's part of the HubSpot product. The point I was trying to drive at is that this kind of idea of, because, and I prefer agentic coding, but we'll use vibe coding, it's more popular, is that because vibe coding can produce code with increasing levels of accuracy, increasing fidelity over longer runs, semi-autonomously, And that's, all of those things are true. The thing that I would posit though, is that code is a small part of the actual solution of actually putting something in customers' hands.

Yeah.

Especially if you're, now, there's multiple dimensions on which you can measure the kind of viability of vibe-coded things. For instance, if I were building an application just for myself, vibe coding is absolutely perfect. Now, as you move to larger, larger user base or whatever, I think it becomes less perfect. Not that it can't ever do it or whatever, but that's kind of one criteria. If it's a temporary app that says, I need this one time a quarter or something like that, I'm not gonna run it. It's not like running my entire business or whatever, a better candidate for vibe coding than other things. So, I think we're further away from is to say a single person is going to vibe code an entire software business and do all things necessary in order to make that software business successful. Because that's going to involve the coding, yes. It's gonna involve design, yes. It's gonna involve UX testing. It's gonna involve customer— discussions and customer feedback. And a lot of this individual pieces we may actually, you know, have AI help with, but I just, I'm not a believer that you can sort of vibe code your way into an iPhone app. And I just don't think it's sustainable. Yes, we have existing examples where, and you see this on the social media channels all the time, where so-and-so built a million-dollar app on the App Store, you know, vibe coding something that they'd never written a line of code before. Sure. We will see those examples. But as a baseline, I don't think that's going to be the predominant model. What I think will happen is I think we will have the equivalent smaller, like software teams, people that says, I'm doing this for a living for this particular category. And it's going to be in categories that historically were too small in some cases to actually pursue because the economics didn't make sense. But it's still going to be small teams of people that are devoted to building an application that they will use vibe coding tools for.

Yeah.

But it's not just a kind of one and done. I can move to the beach and have martinis or whatever and just the money keeps rolling in over Stripe, right?

Yeah.

I don't think that's what's gonna happen.

So let me get a little, let me turn the knob a little to the controversy side.

Sure.

So HubSpot has been patently a little late on the adoption of agents in their core offerings. A little, not too much.

Okay, okay.

How do you introduce this new technology that is fast to build, fast to release, in a company that's 20 years old? Is it part of the core fabric? Is it a separate team? Like, how do you accelerate? Do you think about this a lot? Like, how do you accelerate the release and the usability and the adoption of agentic type software for a company that is already successful with a ton of users that has a culture that's already sort of like cemented it? What's the right approach?

Yeah, and to answer your kind of first question or one of your first questions, I think about this a lot. It's almost all I think about, right? It's like as a founder, And this is not just AI. As a founder, the thing that keeps me up most nights is, are we doing everything possible to achieve our ambition or to maximize the probability that we achieve our ambition, right? It's like, okay, well, and although I'm an optimist, the paranoid side of me is always thinking, like, what decisions are happening in the business today that are going to plant the seeds of our own destruction? We just haven't seen it yet, right? It's like going to happen a year, 2 years, 5 years from now. So those are the questions I ask myself. late at night. And so the way we're tackling this, and this is not necessarily new, so we have the entire core team, and we did, like others have done and have talked about publicly, in Q1 of '23, we threw out the roadmap we had at the time and said, okay, well, this is not a drill, this is not small. And so we had believers through the exec team. So it was not a lack of conviction on AI at all. And we moved, I thought, pretty quickly, but opinions can vary as to how quickly we launch things and got things out there. So, but the way we approach it is that there are the things that impact the core platform and the core product, and we should do those things and are doing those things and have launched a bunch of AI features. We have agents out there that are built and supported by the core team, like our customer agent, which is a nice, I'll call it mature for just on a relative scale as a category. But then we have projects within HubSpot that we internally call like these HubSpot Next projects. And these operate effectively as a startup within, and it's not an innovation lab per se, it's like literally a startup. So we make an investment decision around, are we gonna put a seed bet down on this particular startup? It's gonna have a founding team, it's gonna have, and it is not bound literally by, I'll say, Any, because it's close enough rounding up to any of the constraints that HubSpot has. You don't have to use the brand, you don't have to use our billing system, you can create your own Stripe account, you can say, you can build on your own stack, you can use whatever tools you choose. You can say, I wanna go after this market. So we have a limited number of guardrails, one of which is you have to have, just like a startup would need to have if it were out raising capital, a thesis about the world. That's like, what do you think is going to happen in order for you to accomplish this? That's kind of one of the guardrails we have. The second one is that— The would be huge if true. So we don't want these bets to say, oh, well, we're gonna have a— this is gonna be successful and we have an 80% probability of adding $50 million ARR to HubSpot's bottom line. We don't care. Like, that's not the charter for those particular projects. Those projects are, if it works, Yeah. it can be a multi-billion dollar business. It's got a chance of being a bigger business than HubSpot itself is, like those kinds of bets. So we have a slew of those. That we fund, and we are not sure.

So wait, wait, so I buy the concept of a startup. Like, how's the org structure change? Who do they report to? They still have a boss, right? Like, they're still, like, where's the startup end and where's the startup begin?

Yeah, it reports directly. So we have a HubSpot org structure has gone through a change this year. We have a new CPTO, Chief Product and Technology Officer, that's head of product engineering, UX and these HubSpot Next projects. So it all effectively rolls up to him and I'm biased, but he is frigging awesome. Of course he is. Anyway, of course I'm gonna say that, but it's true. And yeah, so that's how it fits kind of structurally, but effectively the startup. So we have—

So that person will be like their board, if you would.

So we have a 3-person board for these projects. It's me, our CEO, Yamini, and the CPTO are the 3 people that are on the board that make the to make these decisions and similar to how you might have. And so, when we kind of put it together, and this is the first time I actually talked about this, but it's fun to kind of talk through in real time is that—

Yeah.

And I've been around startups and venture and things that made a bunch of my own investments. So, I sort of know the mechanics of how that works. And so, one of the heuristics that we use is that when possible, a HubSpot Next project startup should work like a venture-backed startup would. And the only way to, only reason to deviate from that is if we have a very good reason, like, oh, in this case it's different, so we should do it this way.

Right.

And so for instance, it's like, okay, well, you know, we have a board. It's like, yep, we have a board that kind of oversees. So let's think of us as the GPs in the HubSpot Next kind of fund, so to speak. You know, we sort of make a decision around if there's gonna be like a thematic area and ours is, yeah, we still want it. We don't like to start pizza chains or, you know, going into hardware or something like that. It's gonna be related HubSpot. Doesn't have to be our existing user persona, right? It's like it has to be go-to-market kind of related, but doesn't have to be exactly our customer base. In fact, we'd sort of prefer in some ways for some of the bets for it not to be. And then we have a, what we'll call a, just like we would have in the VC business, there's a partner that's effectively sponsoring the deal, right? That's the lead partner on the thing or whatever, that's going to be more involved.

Yeah. And so So what's the approval process that you, Yamini, and the CPTO have to agree on unanimously to get this, the grant, or just one has to agree? Like, how's that work?

It depends. We don't have it down to a science yet, and I think it varies across venture firms. Our thinking—

Maybe we need an offsite somewhere really expensive to really—

Yeah, exactly. To really encapsulate the thinking. The way we've approached it right now is if one of us feels strongly enough about it, Especially at the seed stage. And this is, you know, most of them are kind of seed A-level bets right now, is if it's a seed stage thing, well, we don't need unanimous consent. We don't need forms in triplicate or anything like that. We just write the check and say, if there's someone willing to go do it, we go do it. As long as we have a founder that's kind of passionate about it and the idea even obliquely makes sense, we don't kind of second-guess it too much.

Yeah.

Once we get to kind of A and post-A rounds, then I think, and we haven't explicitly stated, you know, it's not in the bylaws or anything like that. We have no bylaws for this, but it's, my sense is that like a 2 out of 3 vote would win. It's like if 2 of the 3 of us like disagree on something, then it's like, and the 3rd one, maybe if they're passionate enough, can kind of convince the other 2 or at least one of them to kind of switch their vote. But yeah. We haven't had to deal with that yet. It's a very collegial team, which has always been the HubSpot way. Yeah, we know what we're trying to do. We're, yeah, so it's—

What other forms did you consider before you came up with the HubSpot Next venture?

I'll say this only partly in jest is pretty much all of their forms. This idea of trying to do innovation in a larger company is one that is not new to HubSpot. It's a It's a problem. It has been for decades, and books have been written about it, and lots of talks have been given. I've listened to a lot of them. So we've tried variations of it. We've tried the incubation model where we say, hey, we're going to have this company. We're going to carve out one of the core product teams that's closest to the idea that we're pursuing, but we're still going to give them some constraints like, okay, you're still kind of flying under the HubSpot banner. You're going to have these things. It has to move within a year or two. numbers that we actually currently track that like makes sense. It's gonna move the needle on numbers that we already measure. So it's, which, you know, these are, you know, what some might call like Horizon 2 kind of bets for those that follow the Horizon framework. So we've tried that variation. It's funny though, it's like, I don't know that there's any real perfect answer.

Yep.

This one seems to be, you know, having been at this for a while now and been very close to these kinds of projects at HubSpot, Seems to be working the best. And it may be just a sign of the times that maybe this is the time to be doing those kinds of things because we're going through a transformation, right? And everything's, you know, the deck's being reshuffled in a way anyway. Yeah, so.

So, if a CEO of a $100 million company were to come to you and say, hey, I gotta move faster. All my agentic projects are stuck somewhere in the product roadmap, in Jira tickets. They tell me they're coming. They're just not coming. Like, What would be your advice to all those SaaS CEOs out there who are trying to like get back to growth?

Yeah, as kind of, I'll say this as founder CTO, my advice would be you don't solve that problem with an innovation bet or HubSpot Next or a startup within a startup. You start up, you solve that problem and it might be painful is that you might be looking at a team change or a leadership change or an org change or a culture change or all of the above, right? It's like, okay, if the team is not bought into that, if there's passive resistance, if there's And this is, you sort of have to, and this is, we're not, we don't like to do kind of mandates and edicts at HubSpot. That's not the HubSpot way and historically has not been. But there are times where like we do make a top-down decision that says, okay, this is what we're doing. This is not an experiment. And it looks like, I think what happens, and this is actually a good story since we're having this conversation. So the last time we had this level of, I'll call it, kind of an intervention is probably overly strong a word, but like that is when we got into CRM. So, when Brian and I made this decision—

Well, that must be a tough decision. Wow.

That was hard, right? Because we had been so successful as a marketing software company. We were growing and this is a year before we went public. And we knew we were like, that was on the path. It's like, really? You're going to effectively pivot the company the year before going public?

Come again, tell me this one more time.

It was like, exactly. Like, play the tape again one more time. I must have missed something, right? In the board meetings and like every meeting. And there were a couple of lessons from this, right? And really good lessons, deeply held lessons. And it turned out well. And we have lessons on things that didn't turn out well too that we can talk about. But this one, what ended up happening is if I do it all over again, I would do it. And here's why is that I am a big believer in experimentation, doing innovation bets and startup within a startup, all of those things. But there are some things where you have what I think of as self-fulfilling prophecies.

Mm-hmm.

So, when we made that decision and we very clearly communicated to the company that says, hey, we're doing the CRM thing. It's not an experiment. It's not, oh, we're gonna go try this and see if it works. Otherwise, we've got the, you know, market. It's like, we are going to take— we're gonna do this. It doesn't matter if it takes us a year or 3 years or 5 years or 10 years. We don't care. We are going to be a CRM company, we're going to build this, right? And we didn't know how long it was going to take, and it ended up taking many, many years, right? And because it's a hard problem. But what ends up happening, and this is kind of the lesson learned, is that once you show that clarity of conviction, all the stars start to align, right? What ends up happening is like, okay, well, this is not an experiment. Your best engineers and product people and the people that have been at the company that are the needle movers got to work on that thing. It's like, that's the thing that the founders have said is the thing, right? That's our next big bet versus like on some of the experimental projects, it's not that you can't bring in great people, but some of them are like a little bit kind of, it's like, okay, well, like, is this a good career move for me? Is this, you know—

Oh, absolutely.

Will this be around in 6 or 9 months? Like, is it really needle moving? So lots of things fall into place once you make that decision, make it clear. So back to your question, if you are finding that you're struggling making the AI transition, this has to be a conviction bet. You have to, in the upper layers of the company, in the inner circles, you sort of have to decide, like, do we all believe that this is the case? Which if you don't believe now, then you've got bigger problems. It's like, okay, I don't know how much more evidence you need at this point. And then you have to say, okay, well, Do we have the right people? Do we have the right culture? And if not, you know, what doesn't matter how drastic, what do we have to do? Because that would be to me an existential threat that says, okay, if we do not make this transition, we may cease to exist as a company.

So when you made that decision, what was the passive resistance? Like what was the people that did not go along and that you have to part ways with? Like how did that show up?

The way it always shows up, right? Which is, there's a fine line between passive resistance and lack of active passion, right? Is that, because what you're looking for, right? Is you're like, oh my God, you know, I ran Claude code yesterday or something like that. It's unbelievable what this thing can do, right? And even if not completely true, you want people that kind of see the light, so to speak, right? So, the next level down, that's the ideal state, right? And you're gonna have some number of those in most companies at scale. You're gonna have that. And ideally, you know, Some of them are in the kind of upper levels of the company and can push. We were fortunate that all 3, which I think of Yamini, Brian, and I as the 3 that we saw them very, very early and there was no resistance. And so then, but the path of resistance shows up as what I think of, and this is not just AI related, it shows up in orgs of all sizes, is what I think of as checkbox management. It's like, okay, well, I know I'm supposed to do X, whatever X happens to be. I know I'm supposed to explore AI. I know I'm supposed to try this. I know I'm supposed to do that. And I will do the— I won't call it minimum viable effort. So, they're not intentionally trying to make it minimum viable, but in their soul of souls, they don't think it's going to work. And so, they're not going all in on it, right? They're just making a bet that's, okay, we'll put a few people on it. Yeah, I know Dharmesh said this and Yamini said that or whatever, but we're going to— here's— And I'm not saying this is exactly what happened, but this is how it could manifest. And so that's the thing, you sort of have to, and part of it is just resolvable through that clarity of conviction. That said, this is not a drill. And part of it sometimes comes through org changes, leadership changes can be up and down in the organization. Some of it comes, some form of passive resistance, and this is specific to AI, shows up in the form of people that just don't in principle agree with AI. And that is actually good for humanity. That happens, by the way. And so, I'm not saying it's going to be 50% of the organization.

I'm a little shocked, but maybe I'm in that bubble. So, tell me more.

Yeah. So, there's going to be more than zero. I'll bet you money in any organization, I'll say over 100 people, there will be at least 1 or 2 people that said, and these are the same people, well, and it's because they've seen other technologies come along like crypto, right? And you have the same kind of names like, oh, I understand the technical, but this is just not good for humanity. It's like, AI, oh, this is just not good for humanity. It doesn't really work. Then you have AI slop and you have all these things. It's like, and they'll sort of pull the thread and you still see this in media now, right? So based on which narrative you decide to believe in, I could make an equally compelling case. We could spend the next hour on all the things that are wrong with AI and why we should be worried and why, and the things to kind of worry about. And that's— But they sort of internalize that as a result of which, and so what we did very clearly, and this was not pass as a value judgment. We said, we are going to be the AI-first customer platform. We are like, we have the conviction similar to how we had the CRM bet, right? It's like, anything we will move, we'll move mountains, whatever it takes, we're going to do those things. If you are not comfortable with that, there are amazing companies out there, and you're an amazing person. Our advice to you would be to literally to go somewhere else. You know, life is short.

That's pretty stark. That's pretty stark. I mean, you're, that's That's not very HubSpot-y.

It's not, I agree. But in a way, sort of it is, right? So HubSpot-y-ness is partly around empathy. And it's like, okay, well, if we were them, would we want to? And all of it was true, right? It was not like we were making something up. I think we were being very transparent that says, this is not like an exercise. We're not just saying it so we can get on earnings calls and tell the investors that, oh, we too are an AI company kind of thing. It's like, we feel this in our soul that this is the transition that HubSpot is going to need to make. And part of what has kind of driven the modest success that HubSpot has had is we are on a noble mission. Like we are still have conviction around SMB. We did not get pulled and we can talk about that as well. That's another important lesson learned. But it's like, if you're no longer aligned with the mission, which is now, you know, incorporates AI, go to somewhere where you are.

What were the hard choices around You know, not to judge, but becoming an AI customer platform doesn't feel, you know, burning the ships too much. If you, if you, what were the actual hard choices at the point of the engineer, the product manager, the person selling, the person for successing that had to be made for you to move from a customer platform to a customer AI customer platform?

Yeah. Well, first, you know, so when we made the original call in Q1 of '23, We said, this is a transition we're making. First, it's like not to burn the boats, but burn the product roadmap, literally. It's like, okay, well, like all bets off the table. And as you might imagine, we've had that planned out, right? Like it's—

Right. There's a lot of people and careers that were in that roadmap. No, I get it.

These are all things that we, in most instances, had very animated debates about it. Like what should be on the roadmap? What are the priorities? What are we doing? Where's the product headed? Where's the platform headed? What's happening? Like, it's not like we have no plans. We had plans that we were all very passionate about that we had already kind of debated through, then to come in and say, yeah, so we're not doing that. We're going to start, you know, it's like the— what's— there's a cost term in accounting called zero-cost budgeting or something like that.

Yeah, when you erase the budget, you start from scratch.

Yeah, you start from scratch, right? And as would happen, you know, some things are obviously going to kind of come back on it, but we start from the zero things like, okay, what are the things we need to do. And the hard calls, and this is still a call, right? It's still happening as we speak, because as much as we all love AI, the challenge is that, and it's on an exponential curve, and all the things that we hear about the technology is very, very true. As it turns out, humans are not on said exponential curve in terms of human change. So when you go to a 50-person law firm or one of our other 270,000 customers, and we say, hey, there's this new thing called AI, and you should be using it in this, this, this, and this way. And we have features and products that you don't even have to pay for. Like, you can just try them out. They're included in the price you're already paying. Change management, it's always hard. You know, just because this technology is revolutionary does not necessarily mean that, you know, Susan, the VP of marketing, or Joe, the VP of sales, is going to say, oh yeah, of course, I'm going to throw out my plan and my quota and my budget and everything and rethink it in—

Right.

In terms of AI, right? Like it's, there's a slower adoption cycle.

That brings it to a very important point, which is, so who, who in, in the organization was in charge of getting, making sure that those new AI tools, applications, experiences were adopted and successful? Like whose job changed to ensure that indeed HubSpot became the AI customer-centric platform?

Yeah, the, the biggest change, it's, is the obvious one is the, is the product team, right? Because that they're the ones kind of building building the product. And most things at HubSpot are kind of product-led and platform-led. But then it really comes down to, and it's, once again, we don't like mandates and edicts. It comes down to the, or up to the CEO, I will say, is that, and like, because she's going to have to, you know, with assistance from me and Brian and kind of instill down into them, it's like, this is very, very real. And then have a very kind of hard look at what's happening. And we, alongside of this, right? And culture was brought up earlier. One of the bigger changes we've made to HubSpot culture, which I've always thought of as a product and led the charge on for a long time, is that the HubSpot culture actually needed to change as well. And this was unrelated to AI. This was just that the big transition sort of surfaced this is that, so HubSpot has always had this reputation of being very empathetic, kind, which we are, And that has been a large part of our success. One thing that sort of happened, and I think we were not alone, I've talked to lots of founders of companies at scale where it was the same, and some of them have publicly talked about it, is during the boom years, the post-COVID years, when everything was headed up into the right at double or triple-digit rates, it was easy to sort of become, I'll use the word complacent, it's the one that jumps to mind right now. And then what ends up happening is that And we were growing fast, hiring lots of people really fast. What happens is people join the company and they sort of pick and choose the elements of the culture that sort of reads like, oh, I hear you guys are like empathetic and kind and this and this and this and this. It's like, yeah, but did you miss the parts that we are intense? Like we are like, we are accountable.

You have to be accountable.

We didn't get here by, you know, so it's like, and so what ended up happening is that we sort of, had this sort of drift in the HubSpot culture. So it's— and I'm a product guy, and we talk about HubSpot culture as the second product that we build. It's like, oh, our product had drifted in terms of it no longer serving the needs of the actual customer, which is the team and the company at large. So we've made a shift to say, okay, well, we're going to— not that we're not going to continue to be empathetic and nice and all those things, but like the number one slide in the HubSpot 128-slide deck is solve for the customer. And you cannot solve for the customer by— it's like, they are literally like, if you can draw a path for me that says, I'm doing this and this, this way, because I think it's actually in service of the customer, you will have me as an advocate. But you can't do it. It's like, oh, well, that's because it'll make the employees happier. It'll do this. That's a secondary, tertiary benefit. It's like, oh, if we have happier employees, they'll serve the— but if that, like the first-order direct benefit to customers and solving for the customer should trump things every time. So yeah.

So you use the same rubric and the same true north to reestablish this.

Yes, but put the spotlight on the elements of the culture that, to make it less ambiguous as to the fact that this is what we're solving for. And we had that same, you said, a broader kind of company-wide scale that said, we are changing our culture, partly driven by AI, partly just this awakening that was happening in the kind of post-COVID boom years, it's like, we don't think this is right. I think the culture has drifted and here's how it's going to— and once again, this may not be the right company for you. Like we didn't say leave, but it's like, it may not be the company you can— but yeah, so.

That's amazing. So this podcast is about getting paid building AI. So let's talk about getting paid building AI. So how has your thinking evolved In monetizing HubSpot now that the AI agent is here. So, you know, marketing automation began with the, you charge for accounts for like the number of leads that you had in the database. Very clear model. Everybody was doing the same model. CRM is more on a proceed basis, et cetera. So now you have these agents that are driving, you know, work and outcomes and tasks and full resolutions.

Yeah.

What's your thinking there? Like, you know, how much of the old, Pricing model monetization is going to stay, how much is going to change? What's the latest?

Yeah, the latest, our thinking is that the seats model in some form still makes sense for some portion of our products, for some portion of our customers, because it's like, okay, if you're selling a tool that humans will use, that's a close enough proxy. There's a reason it worked. It's because it works. Okay, great. Yeah.

But wait, is it because of that or is it because there is less friction in the sales cycle because people are used to buying per seat?

Both. Both. I think it makes sense. And as a result of which, and then we're used to selling it in that way. I think it's both. But then, so let's look at the kind of the AI side of it. So I'll tell you the answer first and then we can provide— Yeah, yeah.

We'll backtrack into the explanation.

Yes. The answer is we're going to have a hybrid pricing model that is Partly seat-based and partly consumption-based. And the consumption-based side of it is going to be this, or is this credit-based model that says, okay, certain parts of the HubSpot product consume credits at certain rates, and you buy a certain number of credits, and there's a plan and the mechanics are in place for that. The thing we have not done, and I can tell you the debate that we had, or discussion we had internally, is we did not We elected not to do the outcome-based pricing, including for customer agent. And our motivation for that, not that it doesn't make sense for a customer agent, and there are people and companies that are really good and have had success with it. Our thinking was we wanted, we think different pricing axes add dimensional complexity to a business, right? So when we were pre-CRM days, when we didn't have seats at all, it was all kind of the number of contacts in your database. That was a nice simple model. That's how you bought. It was fine. And even before that, we had a single price, which was like the stupidest thing in pricing history. We can talk about that.

But you can wrap your head around it.

That's how it works. And yeah, so we want to reduce the number of like, what's the minimum number of dimensions we can have, the variable axis as we think of it, right? And so one, we think it's going to continue to be seats, partly because industry's used to it. We know how to sell it. It makes sense if you're still selling software to humans. And if there were ever a day that there are no humans to sell software to, or that's no longer a good proxy for value. Like, we're not dogmatic about it.

Yeah.

But the reason we kind of decided not to do outcome in addition to the kind of credit-based, pure kind of consumption-based model was that that was an additional axis that we thought the marginal value of that did not outweigh the complexity that it would introduce. And for a couple of reasons, the other thing, and this is And part of the discussion we had as well is that, you know, we sort of gravitate towards the examples we have right now, and they're great examples on the kind of customer agent side. But I'm like, okay, but this doesn't really— so what makes it successful, made it successful in that particular sector of the market is that customer service has been around for a long time. You know the cost per ticket roughly, you have an approximation, most organizations know what it costs them. You have an objective measure for successful outcome, which is, is the CSAT high enough? Was the agent able to resolve the issue, right? So there's a set of traits, I'll call it, that make customer service agents almost near-perfect fit for an outcome-based model, right? So objective measure of success, economic value associated with that successful outcome, and a way to sort of measure whether you're doing it.

But Dharmesh, how much of that is hindsight? I mean, it's obvious now. It wasn't obvious 2 years ago.

No, but it was like, but we didn't have to make a decision 2 years ago, right? We were still like in the midst of like how, what, does this need to look like? Yeah, so—

So the market is there, it's clear, settled.

Yeah, so the other thing, I mean—

You still decide not to do it because it increases dimensional complexity to your—

Yeah, yeah, of course. So let's look at an example, right? So let's look at other parts of go-to-market. So let's say we had like a content creation or marketing-oriented agent, which we do. And we said, oh, this thing, I'm making this example up because it doesn't exist, but let's say it did, is like a logo creator agent. It's like, okay, so we, and because we serve SMBs, lots of startups, maybe we have the world's best logo creation agent someday. It's like, well, so we could charge and say, oh, well, it costs $100 a logo. Okay, is a logo worth $100 or worth $10 on Fiverr? Or is it worth $5,000 if you go to a big brand? What does a successful outcome for a logo look like? And when would you pay? It's like, well, so the things and what's the economic value per logo? We don't know any of those things, right? So that's almost at the other end of the spectrum. which there's no objective measure. The economic value is variable, especially across industries and across companies. And so, you know, our thinking was, that's not the only example of the logo because I just made that one up. There's going to be examples of things where outcome-based pricing is not going to solve it. So, we can't go to full outcome, right?

Right.

So then, the question is, do we have 3 dimensions or do we have 2 dimensions? And we made the decision, let's go with 2 dimensions because anything that we could do in outcome, as we're learning with customer agent, can sort of be formulated in the form of a credit-based model that says, okay, it's going to be a lower price, but the customer is sort of absorbing the, I get to decide what the outcome is or whatever, but I'm just gonna pay based on X credit consumption, which is—

Yeah, yeah, no, and the credit thing actually makes total sense, especially for companies that are moving from seat-based to seat-based plus some kind of usage because the agent is actually doing all this work. Let me ask you a question around credits. Why introduce an intermediate currency that is not dollars? What does that do to the model? Why couldn't it just be dollars instead of credits?

Yeah, that's a really good question. And one we talked about, and the reason is, and this is not to, and I'm not an economist, I'm not even smart enough to—

But you play one on TV.

Sometimes I play one on TV, is that the value of a currency existing is because that currency is then kind of is fungible, right? So you can sort of change the rate card, so to speak. When you go to dollars, it's like, okay, well, so we talk about dollars. It's like, okay, why not just dollars? But it wouldn't be dollars because we are actually a multi-currency product, as you would expect, right? We're in 100+ countries or whatever. So it's like, so we needed a currency that says, okay, well, things cost varying amounts or whatever. So we need some currency. And so one of the decisions we made is we're going— a credit is a credit is a credit. So we have a unified currency. It's across all of HubSpot. And the only thing that varies is that some actions might cost multiple credits instead of one credit. But what we like about it, we like the simplicity of it. And we tend to, when in doubt, take the simpler path because there's always time to make things more complicated. It's always harder to try to make things more simple again and go back the other way because the way the universe works, entropy and everything. But that's now me playing physicist and economist on the same show. But so the idea was, can we keep it simple, come down to a single currency, and then we can sell that currency, so to speak, because it has some nice pricing mechanics that I think would be kind of relevant to your business because we can sell batches of credits. We can sell subscriptions to credits that says, oh, just like, and there's, you know, and these kind of credit-based models have been around for a long time. We've had them in the mobile industry with minutes and things like that. And then that, You know, obviously evolved over the years, but you have a little bit more flexibility in terms of rate card changes. And for instance, it would sort of be weird if we wanted to sell bulk. It's like you're selling me bulk dollars. So you're selling me $500 worth of currency of AI consumption for $300. That is like, I don't think that would fit in people's head as well.

Yeah.

Now, having said all of that, one thing we are learning and still evolving is that businesses, including our market particularly, are not used to credits yet. They just aren't, right? So people don't love that idea. It's like, okay, well, and so it's going to take, so our sales force has to be kind of educated and trained and they're going to have to go through some cycles and some iterations to kind of get to how to, but one of the big benefits that came from that credit model though, that we could have actually done with Outcome as well, but, you know, versus So when we started this, we were not sure what this being the kind of the AI rollout of products and features, we made a very pivotal decision. And that decision was that AI was not an add-on product that you bought. Like at time t equals 0, when we went kind of out with it and we talked to the board and even to the investors and explained this and our thesis behind it was that, so we think we are in a new generation of our industry being CRM and customer platforms. And the last big shift that we saw was Salesforce with the cloud CRM, which is brilliant and brilliantly executed, massively successful company. What Salesforce didn't do, they didn't say, hey, we are a CRM company going against the incumbents like Siebel, and we have this cloud option now that they don't have. They didn't say that. They said, we are a cloud CRM company and you cannot buy Salesforce except for the cloud version. There is no— Non-cloud version of Salesforce, right? Yeah, yeah, yeah. This is the path, this is the way, which we think makes total sense. So when we said, oh, we are going to be the AI-native CRM company, customer platform company, we can't say that and then say, oh, and by the way, AI is something you add on. That's like Elon coming to us, it's like, we do EVs and you have to buy the battery in order to get the electric features of a Tesla.

Yeah.

You know, just at face value, that sounds sort of kind of silly, right? And so what we said is, We're going to bake some fundamental. So our copilot, our Breeze assistant, which is our copilot, included in the product and free. The early features all free. And we're like, okay. And part of the kind of core culture and kind of business philosophy at HubSpot is we always like to create value and create economic surplus. This is the second economics reference. Create economic and consumer surplus and then go back and say, okay, well. we can start taking some of that value back. And there are lots of studies around how much of the economic surplus you can capture.

So where have you drawn the line now? Like at what point does the usage begin?

Yeah, so we have now started monetizing now, as in earlier this year, we said, okay, oh, customer agent, clear value. There's market acceptance of that.

No-brainer, 100%.

No-brainer. And so now, but now that we have the framework in place, so the way we approach it is that For our kind of flagship agents that we have out there, we say every customer gets those agents for free and gets a certain number of credits for free. So we don't sell the agent per se, and you get to try it. So it's like, and that's one of the other benefits of that credit model is we can afford to do that. It's like, okay, well, the ones that are successful with it will continue consuming credits. And we should see that kind of credit line, credit consumption line go up over time in, at some point, relatively predictable ways, right? At the scale that we kind of deal with it, we have, you know, 9,000, 10,000 customers a quarter, net new, that will be joining so we can kind of follow cohorts, newer and older. So that's where we've sort of wound up is like now we have the framework in place. It's seats plus credits. Seats we know very, very well. Credits, we make the decision, is this something that customers will ascribe value to that it makes sense for us? And we try not to do it based on COGS, although secretly I think parts of the business do. It's like, wait a second, Dharmesh, we can't not charge for this because there's Yeah, yeah, yeah. COGS involved because we're paying OpenAI, Anthropic bills and such, but yeah, so.

So yeah, the last question just before we wrap. So extrapolate the business 5 years from now, if you would, and how much is the credit contribution to the business versus the seed contribution to the business versus something else you're not thinking about contribution to the business?

My sense is that the trend line will move Up, so I don't know, and I'm not allowed to provide dates and timelines and slopes of curves and such. But I expect that over the fullness of time, credits will become the predominant model. And the reason for this is that more and more of the value that HubSpot creates will be better kind of better articulated and better captured with that model versus the seat-based model. And the way we want. And this is something I tell anyone that asks me is that my sense right now is that, and so one of the questions we get sometimes from Wall Street is around, oh, well, aren't you going to see this kind of headwind because the number of seats is gonna go down as AI takes over more and more things? It's like, okay, well, yes, but my expectation is that if we create software such that it reduces by X percent, let's say someone in a particular role, That is actually higher value creation. And that actually makes the EV of HubSpot go up, not down, right? So yes, maybe seats will go down, but the thing that they're getting, the thing we're selling, the thing that we built actually has more value than the seat we were selling before that. So in an ideal world, I'm like, okay, well, if we can sort of that clearly demonstrate kind of value to our customers and on kind of return on investment, it's a very simple, it's simpler math. The seat math says, oh, well, I have X number of Y people and I can reduce or either increase productivity by X, or it's just the table stakes. You have to have a Y piece of software for those people to function. I would much rather tie it to actual kind of value optically, kind of seen more visibly.

I 100% agree. I think all systems of record, this is the recipe that is working right now and of adding credit consumption to the seat, and count on that as a lever of growth going forward.

Yeah, I think it's like, and it's, you know, the parts of AI that are, you know, we talked about the expansion and the market opportunity for software companies because we can go after, you know, much larger pieces of value. And the other one is the rent. And so SaaS brought with it the kind of subscription model. AI is sort of the underpinnings of this kind of credit-based model, which Sort of related, but kind of not, right? We could have done that before, but I think the reason it makes more sense in an AI world is that because now we can connect the dots on here's how this actually takes on some kind of human-oriented, historically human-oriented or human-exclusive tasks. So I think that sort of makes sense. Yeah.

100%. Dharmesh, we are at time. This has been incredible. Thanks again so much for making the time. Oh, appreciate it. I could have stayed another hour. Talking to you. We didn't even talk about Agents.ai, which is a whole—

We'll save it for next time.

Exactly. Exactly.

By the way, there's no rule that says guests can't come on twice, Manny.

That's fine. That is true. That is true. I have nothing but questions for you, and I really appreciate you sharing the wisdom with our audience.

Thanks.

Thanks, man. See ya. Now, if you like this podcast, please subscribe on YouTube, Spotify, Apple Podcasts, or wherever you get your podcasts fixed.