HubSpot, Inc. (HUBS) Earnings Call Transcript & Summary

Perfect. Don't worry, Chuck. We'll leave it there. Maybe moving on to Data Hub. So Data Hub at INBOUND, it seemed like there was a greater emphasis on that, which resonates and makes a lot of sense because the common refrain that we hear is that in order to unlock a lot of the value with AI and the application layer, you need to get your data state in order. So is that the exact opportunity that HubSpot is targeting? And maybe you could just talk about some potential emerging growth opportunities out of Data Hub.

Yes. I think you're exactly right, Taylor. In order for you to unlock any value from AI, you need data in one place. And you don't just need data, you need high-quality data in one place. If you take a typical 500-person company, a lot of times, data is disparate in multiple places. They need to bring it together. And data is incomplete or just duplicate data. I mean I used to live in this RevOps world for many years. And the problem that you have is that even if you have customer data, it is inaccurate because it was not updated in real time, or it is duplicate because you will have the same Taylor McGinnis with multiple emails. Which email am I going to send to you based on what I know about you, that is a huge problem. And if you don't have that sorted out, then you're not going to be able to put agents on top or a use case with AI on top to get the real value from it. And so Data Hub does exactly that. It does 3 things. The first thing it does is it helps bring data from across the go-to-market stack and across the enterprise stack into HubSpot. We have something called data sync that allows you to pull all of that information in one place. The second thing it does is improve the data quality. We now have the ability to use LLMs to say, give me the exact name, email as well as the role of this person and fill that information out. That is something that you can now use LLMs to do and make sure that, that data that comes back is high quality as well as validated data. That's the second thing it does. And then -- we -- the third thing from a Data Hub perspective is we created a workspace for data analysts and RevOps analysts. It's called the Data Studio, where you can now make joins and make changes to that data and drive automations from there. And that is the foundation to then be able to use a data agent or a customer agent on top of it. And so it is almost foundational to getting value out of AI. And we see that as a multi-hub opportunity. If someone wants to actually drive the Loop Marketing, they need clear data for segmentation, for personalization. So they need to start with Data Hub and get that high-quality data, then they'll be able to use Marketing Hub to drive better personalization. So it's a multi-hub opportunity to drive Loop Marketing for our customers as well as improvements on the sales side.

Yes. So moving up the stack into the AI agent layer, HubSpot introduced a number of AI agents. On the last earnings call, there were some pretty impressive statistics just in terms of the quarter-over-quarter customer growth that you're seeing with those offerings. So maybe you could just talk about the pace of adoption today, how that's trended relative to your expectations? And when do you think we're going to reach -- I know this is a hard question to answer, but when do you think we're going to reach that tipping point where we get more widespread production use of AI agents?

Yes. I would say we have widespread production use. I would say there's a lot more of adoption to go to. But if I step back, we have customer agent, prospecting agent and data agents. Those are the 3 featured agents that we have launched, and they are now in general availability. Customer agent resolves tickets, whether it's support tickets or sometimes sales and marketing questions, it resolves that. We have over 6,200 customers on the customer agent and with an average resolution rate over 60%, and that was in a fairly short amount of time since we got to general availability. Of course, the opportunity for us is to get it to multiple times that in terms of customer usage, and I'll talk about what will help us get there. Prospecting agent is another one. It does something very, very important within sales, which is research accounts and helps you figure out which accounts that you need to prioritize and who within that account you need to contact. And that's one of the foundational, kind of, like, jobs within a BDR team or sales team that used to be exceptionally manual. There's almost -- like, nowhere -- when I started my career in sales, I used to, like, literally spend hours trying to get to which accounts do I need to do? Prospecting agent does that. And we are seeing really good traction in terms of that particular use case. We have over 6,400 customers that are using that particular use case. I think you asked a question of how does it get, like, widespread. I do think that it starts with some executive within the customer account that wants to drive AI and wants to get value as well as growth out of AI. A lot of times, people have asked me this question of like, "Do you see more usage in upmarket customers or down-market customers, or what does it take?" I actually think it takes a leader within the organization to say, let's just not -- let's jump in and start adopting AI. That's the #1 thing that we look at. Then you need high-quality data. We talked about the data that is needed for AI. Then you need high-quality data, and then you basically need a road map for driving adoption. Where do you start? What are the set of use cases that you start and drive. It's not dissimilar from every other technology cycle that we have seen. If you step back and think about the other technology cycles that we've gone through, you always have customers in 4 broad buckets. You have bleeding-edge customers that are running towards a new technology. If Gemini Model 3 comes out, then next day, they are actually beginning to use that. And that's like the first group. The second are early adopters. They wait to see that there's a little bit of momentum in certain use cases. That's the second group. And then there is a majority of customers that are waiting for early adopters to get really good benefits so that they can start. And then there are laggards that we're still talking to people who are -- for the first time, are adopting SaaS, right? And so we're seeing exactly that. When we talk to customers, we categorize them into 1 of these 4 buckets. And based on that, we help them come up with a road map of how they should be adopting AI. And so I think I continue to believe it's like a huge tailwind, and it's all about delivering great customer value that is repeated, visible for them so that the adoption begins to kind of like increase in pace.

Yes. And another interesting move that HubSpot made in the AI space, it was the first B2B SaaS company to create a number of connectors LLM providers. So I'd love to talk about those partnerships, right? So on one hand, it seems like there's a lot of value that can be unlocked by partnering with the LLM providers. But then you have, as an example, OpenAI, where they released a number of videos showing what they're doing internally with AI agents, especially as it relates to front office use cases. So I think that, that sparked, amongst investors, a question of, well, could these LLMs start to get into the AI agent space, be potential competitors in the future to the SaaS incumbent. So first question for you on this topic is, can you provide a bit more color on these partnerships? Are they mutually beneficial on both sides? And because on its DevDay, OpenAI actually showcased HubSpot using AgentKit, maybe you could just talk about what that is and what that means.

I mean there's a lot going on there.

A lot going on there.

Look, I mean, I was just really surprised with the reaction for the demos because almost every large company internally uses their own products. It's not been -- it's not a new thing for you to use your product internally and to -- like, Google has done this for years. They have an internal CRM. And Meta has done this for years. They have an internal CRM. And so I think, like, it doesn't mean that you're immediately going to productize it and build an ecosystem, build an army of sales reps and start working with customers and drive deployments. I think there's a lot between those 2. But having said that, maybe taking a step back, we do believe that LLMs and HubSpot are complementary. There is tremendous complementary value that we add on top of each other, which is why we were one of the first CRMs to build connectors, like you said, with Claude, with OpenAI, with Gemini. This is a new surface right? If you think about LLMs, there are a new surface and a new operating system where people are asking questions. And we want to be the one that is bringing insights -- business insights into that surface. If you go to an LLM and you say, 'Write me a sales email," it's going to write you a very generic sales email. But if you can bring the HubSpot context and say, "Based on what's in my pipeline, based on the last 20 conversations, write an email," that is a much better email that converts better. So I think our strategy has been let's take the value that LLMs provide, which is insights, and add the context of the business application in order to deliver exceptional value to our customers. And that's why we built these connectors. And what we are seeing is that a lot of the high-level questions of, "Give me the trends between the last couple of weeks," happens within an LLM, and the actions that people take come back to HubSpot. "Let's build a campaign, let's create a set of sequences to reach out to 1 million-plus customers in our contacts." Those types of actions happen within HubSpot, and we're pretty comfortable with insights generating from LLMs and actions being taken within HubSpot. The other thing is that as applications, we all need to be present within LLMs. We talked about AEO. When you think about AEO as a new source of leads, you have to be recognized within the LLM. And so partly, what we are also seeing is that when customers or prospects use an LLM to start with asking a question to generate insights, it also -- it's a pretty easy citation process to get them back into HubSpot, and there is tremendous value. We want to be the best AI go-to-market customer platform on top of LLMs that add value to our segment of customers, and that's literally the strategy we are executing.

Perfect. And then before I take one from the audience. You introduced this new pricing model, and you're one of the few SaaS companies that's actually talking about seeing a pickup in activity as it relates to seat expansion. So could you just elaborate in terms of what inning are we in terms of this transition? How you expect that to evolve going forward? And not to throw too many questions at you, but there's been the introduction of this credit space model as well too. So how are you thinking about for the customers that have adopted agents? Like what does that mix look like going forward? And how do you think about the balance of those 2?

Yes. So maybe I'll step back and talk about what is our pricing strategy and how do we think about the opportunity of core seats as well as credits. And if you step back, our pricing philosophy has always been very consistent, which is add value before we monetize. And we've been very disciplined about that approach because when we do that, we know how that delivers. Once we deliver the value, then we don't see the churn associated with it in the back end. So the repeat value and visible value to customers has been the North Star for us. Having said that, our pricing is hybrid. The way we monetize is through persona seats like Sales Hub seat or Service Hub seat that is for a specific role. We also monetize through core seat, which we'll talk about in a minute, and credits for certain aspects of agentic actions. Those are the 3 mechanisms through which we monetize our product. The Sales Hub, Service Hub seats, you know it, it's been consistent. We have embedded a lot of AI value. And as our customers use AI features within that, we begin to see seat upgrades. We're seeing that within Service Hub seats. We're seeing that within Sales Hub seats. And as they engage with our platform and use embedded AI features, they continue to upgrade. Core seats was something that we launched in 2024. And when we launched it, we actually added platform value. Now think about a user that wants to edit a CRM record. They want to create a new custom object or custom property. They needed the core seat. And when we did that, we saw pretty good traction, and we've built a $100 million business in a fairly short period of time with the core seats. What we did at INBOUND this year is we added AI and data value into the core seat. Typically, you'll see a lot of SaaS companies have a separate copilot, and they'll charge x dollars per month per copilot seat. What we did is we took that, Breeze Assistant, which is like our copilot. We added it to the core seat. In addition, we took a lot of the data that we had from Clearbit, which is the company enrichment data and contact enrichment data, and we put that as part of the core seat. So now the core seat is what you will need for every go-to-market employee that is not using a Sales Hub or a Service Hub seat. And we think that as we add more AI and data value, we can drive a lot of adoption, and we are in the early innings of core seats growing within our customer base. And then the last part, which you already mentioned, is credit. So agents use credits as well as Data Hub syncs use credits, and we're in the early stages of rolling that out. We talked about some of the early traction that we are seeing with customer agent and prospecting agent. And customers will get included credits, and once they are done using those credits, they can upgrade to $100 packs or $1,000 packs of additional credits. And so early days, but again, the combination of these 3 is how we monetize AI across the platform.

Perfect. And to wrap it up, just one from the audience. So the question is Loop seems more of a playbook than a product. It seems like this could evolve into an actual product and be a potential revenue opportunity for you going forward. So doesn't HubSpot have the right to win here given the breadth of the customer base for whom you're already solving demand gen problems for?

Yes. Loop is a playbook. Just like INBOUND was a playbook. INBOUND was not a product, and Loop is not -- it's very similar to that. And we think about giving -- when so much is changing within marketing, we want to provide a step-by-step approach customers can take in order to grow their top of funnel. One of the things that -- primary things they come to us for is, "We want to have digital presence, and we want to grow our top of funnel," and that is, "The playbook is Loop." Now what we have done is that for every step of the playbook, there's a combination of features within Marketing Hub, Content Hub and Data Hub that powers that playbook. So we talked about diversification of sources. Well, we have within Marketing Hub and Content Hub, the ability to drive content across multiple sources, and that is an example of how we deliver it. So the way you should think about it is Loop is the playbook. We have Marketing Content and Data Hub that enables the steps within the Loop playbook. And we are now training our internal teams across customer success and sales as well as our partner ecosystem to help our customers drive that methodology and adopt it. And we do think it is a pretty big opportunity as customers look to us to help them navigate where AI is taking them.

Perfect. Well, we'll wrap it there. Thank you, everyone, for joining, and let's give Yamini a round of applause.

Thank you so much. Thanks a lot.