Dharmesh Shah: Thanks for the question, Brian. So we’re really thrilled with the kind of adoption of HubSpot AI after its launch at INBOUND. And we’re particularly pleased with the repeat usage we’re getting out of some of the marketing oriented use cases. And as Yamini mentioned in the opening comments, one thing we’ve learned is that if you can put these AI features right where people do their work, it works. And so we’re continuing to double down on that. AI is the — literally the #1 priority on the HubSpot product road map going into 2024. We’re super excited. And one thing we’re going to do is take the lessons from the adoption that we’ve seen so far, and we’re taking a ChatSpot, and we’re going to kind of leave it through the entire platform.
So we have a natural language interface and kind of add the Smart and Smart CRM within HubSpot. So that’s kind of high on the list. In terms of OpenAI in their recent conference they had just 2 days ago on Monday. Super exciting things. Here are things that are most relevant that has — and we’ve been working with them on preview releases of what just got announced. The #1 thing is they launched a new version of their model called GPT-4 Turbo, which is as the name implies, is faster. But the thing that’s much more exciting about it is that they expanded the context window. And without getting too geeky, the context window defines how much data you can pass, how much information when you prompt the LLM when you ask a question. And they went from 8,000 tokens, which was a default before to 128,000 tokens.
And to put that in context, that’s equivalent of a 300-page book. Now here’s why that’s important for HubSpot. When we’re generating, let’s say, a sales e-mail that says, “Oh, we want to do an upgrade sales to this particular customer”, now we can take all the information that’s in the HubSpot Smart CRM on this unified data record and take all that context and pass it to the LLM in real time so that when it crafts that message, it does the best job possible. What has me even more excited, so it’s not just the CRM data that we have now, imagine in the future, we have Clearbit data, where we have all this information, all these 200 attributes, all this intent data, we can bring that all together into one single kind of prompt to the LLM and it can take all that information in real time and craft these really, really pointed effective messages.
So super exciting going into 2024. And then one for Kate here is that one thing they did announce is that the cost for all this kind of invocation of their latest model is going down. Our estimates are somewhere around 50% to 70% based on our usage once this goes into GA. So lots to be excited about AI is massively exciting going into 2024. Thanks for the question.
Operator: The next question comes from the line of Rishi Jaluria with RBC.
Rishi Jaluria: Wonderful. I just wanted to dive a little bit deeper into a Clearbit really exciting acquisition. I heard some good things about the company. Can you talk a little bit about how this is going to impact your relationship with other players in the space? Including ZoomInfo, which is popular to use along HubSpot, Apollo, we actually saw HubSpot speaking at Apollo’s user event earlier today. And maybe alongside that, I know privacy is obviously is very, very important. How do you think about just dealing with some of the maybe privacy and governance concerns that might come with something that is serving up data.
Yamini Rangan: Rishi, I’ll take that multipart question. Maybe start with the vision and then the competitive landscape and then how we’re thinking about privacy. Look, we’re super excited about Clearbit and how they’ll help us accelerate our vision for the customer platform. And what we’re doing is we’re bringing together this data and powerful AI tools that can consume that data and world-class engagement hubs, and that will help our customers connect and grow better. And if we step back and think about the rationale for us, the age-old problem with CRM is that when you open up the box and set it up, it’s empty. And our customers want us to pre-fill it with data on day 1. They want us to pre-fill it with company data, who else can they sell to, intent data, who else is looking to buy their products and services and contact data, who do we connect to.
And this makes complete sense for our customers. And it solves a huge problem for them. And that’s why Clearbit makes sense. Like I just mentioned in the prepared remarks, they have 20 million company profiles with rich attributes for each of these companies and makes absolute sense, especially in the world of AI. And in terms of the competitive landscape, this is not about us entering a database. This is much more about us accelerating our customer platform vision. And vision that we have articulated is we want to have engagement hubs powered by a Smart CRM, powered by data, powered by AI that helps our customers grow. That’s what we are accelerating. And the word platform really means we’ll continue to operate in a very open ecosystem. And if our customers choose to work with Zoominfo and Apollo, we welcome that, and we’ll continue to have a very open platform as well as ecosystem.
And I think in terms of the final part of your 3-part question, which is how do we think about data privacy. Look, we know our customers value having trust in our security privacy practices. And our level of commitment to security, privacy and governance will not change as we look at integrating the data asset. When it came to e-mail deliverability, we took the right side of the road in terms of how to help our customers deliver e-mails, but do it in a very thoughtful manner, and that’s the same way in which we will look at this as we go into accelerating our vision. Overall, super excited. I have to say that the feedback from customers and partners has been really, really positive in the last week.
Operator: The next question comes from the line of Alex Zukin from Wolfe Research.
Aleksandr Zukin: Maybe just a quick 2-parter for me. As we think about this macroeconomic environment, I know we’re guiding one quarter at a time, but at this point, we’re starting to kind of get our models ready for next year. Is it fair to assume that you’re kind of planning for this type of environment to persist and continue. And anything that we should think about or keep in mind as we start tuning our models for next year, or thinking about. You mentioned FX and the headwind there. But anything about from either billings or I don’t know the — just the retention environment or the putting in Clearbit into the numbers, anything there would be super helpful. And maybe just as an adjunct, any thoughts around for next year. As we think about AI monetization pathways, is it more around tiering up into more premium hubs? Is it new product modalities? Anything helpful there?
Kathryn Bueker: Maybe I’ll start and talk through a bit of the — how do we think about the economic environment in setting guidance. Certainly, Yamini and I said this in the prepared remarks, it continues to be a very challenging external environment. We believe that, that environment will continue to be challenging through the end of this year. It’s frankly a little too early to talk about next year. We will give, as you know, formal guidance on our Q4 call. In February, what you should have heard Yamini say is really like we have very clear focus areas. We’re focused on productization, we’re focused on AI. We’re executing really well. And we feel really good about how that sets us up when the economy does turn. That’s the expectation is that things remain challenging, at least through early next year.
Yamini Rangan: And I think on the last part of your question from an AI monetization perspective, really no change from what we walked you all through at Analyst Day. We’re really focused on delivering that road map, most of the features that we announced at INBOUND are either generally available now or getting close to general availability by early Q1, and that’s when we’ll share even more specifics about AI, but overall consistent with what we shared with you at the Analyst Day.
