Good morning, everyone. Welcome to HubSpot's Analyst Day here at UNBOUND. A big thanks to all of you that are joining us live here in Boston. It is really great to see you all. Thank you to those of you that are joining us on the webcast as well. We really appreciate you taking the time to join us today. I am Geoff Koegler, Vice President of Investor Relations here at HubSpot, and this is my first UNBOUND. I am incredibly excited to be here with the team and all of you. Speaking of UNBOUND, I hope you had the chance yesterday to catch Yamini, Duncan, and Jon's keynotes. They did a great job walking through the latest trends across AI, marketing, sales, and service, and highlighting the amazing innovation that we are driving across our platform, including some exciting new product announcements that we made this week. If you missed those, make sure you catch the replays. Also, make sure you catch the replay of Dharmesh's keynote on how to win with AI by becoming a builder. It was a really great session. No Analyst Day would be complete without a safe harbor statement. You can find the full details on our investor relations website at ir.hubspot.com. We will also be posting the full presentation to the website shortly after the event concludes. We have got a great agenda for you all this morning. Yamini, Duncan, Jon, and Kate will join us up on stage. They are going to unpack our vision to win in the outcomes era. They will walk you through where we are headed, our strategy across product, pricing, and go-to market, and how all of this will drive long-term revenue growth and profitability for HubSpot. After that, we will have Dharmesh join the team up on stage, and we will have some time for Q&A. So get those questions ready. For those of you that are live with us here in Boston, we will have a short window to connect and grab some lunch afterwards as well. Finally, at 12:30, we will all head back downstairs to the main stage session with Sunita Williams. All right. With that, let us get started by hearing directly from some of our customers. HubSpot has become very easy to use for us. It is giving us everything we need, fast value we are seeing, and it is giving us that context that we need that people can trust the system fully. Now the sales team, not only are they not having to use templates and manually click multiple times to pull different things in, with the BDR, they were able to recoup 8 hours per week back. That is a whole day added to their schedule. With the AI-drafted emails, what we found is less than 3% needed to be edited. That means 97% of the time, AI email could have just gone out without anybody having to oversee it or look at it. It is our go-to-market tool. It has transformed from what we used to look at things as just a CRM that logs calls, but now it is actually doing a lot of work for us and helping us to drive revenue. My top rep actually sold 58% more in the first half of this year than he did the entire last year. Those are pretty staggering numbers right there. With prospecting agent, to be able to put those on a virtually dead channel to us, in the first month, we booked over 8 deals from that. When looking at the previous year, we had booked almost 8 the entire year. The context of being able to have Breeze in HubSpot, to be able to look at all the data that we have going across the entire life cycle is very impactful. It has actually freed me up to work on parts of the business that we have not paid attention to. We trust HubSpot with our data, so that has been our number one choice when it comes to layering in AI in marketing. From start to finish, to build that workflow using the data agent, I would say it was about 10- 15 minutes. What would probably take hours to consolidate into a single brief, it is doing that within seconds now. We have an allocation of credits every month, but we have worked out that the Breeze prospecting agent, the cost of customer acquisition is a lot lower than traditional marketing. So we are accelerating our use, and that has allowed the CS agents to actually make proactive outbound calls now. HubSpot is incredibly easy to use. It's got really quick value and trusted context, which I think is huge. What we do at Octagos is we give clinicians back the time to really focus on what matters, which is patient care, and HubSpot does the same thing for marketers. We didn't just get faster, we really moved up from making the content to deciding what's worth making, and I think that's been the key. Please welcome to the stage, Yamini Rangan. Hello. Good morning, and welcome to Analyst Day. It is so great to be in Boston. For those joining us here in person, it's great to see so many of you here. Look forward to the conversations that follow. If you are joining us online, thank you so much for the support and thank you so much for joining us here. Welcome. All right. Today, let's get started by talking about HubSpot's vision and where we are headed. AI is creating a new era in software, not just defined by the features that we deliver, but by the measurable customer outcomes that we can drive. That's a big change. We are transforming every part of HubSpot in order to lead in this new era, from our product to our go-to-market, to our pricing, to how we operate as a company. We're doing this from a position of strength. Our connected customer data as well as trusted platform and the broad, expansive distribution that we have and the ecosystem moat, they all give us durable advantages. We're staying very disciplined, not choosing between growth and profitability. We are committed to driving both durable growth as well as expanding margin. Today, you are going to hear from the leaders in product, go-to-market, as well as Kate, and we want to walk you through our strategy to accomplish all of this. Let's get started. Yesterday, I talked about the outcomes era, and the outcomes era is here. The last few years, we have seen one of the largest investment cycles in technology. Trillions of dollars invested in building the foundation for AI. The next wave of value is actually going to come from applications that take that intelligence and convert it into measurable customer outcomes and growth. That is the outcomes era, and HubSpot is built to win it. There has just been a lot of debate on what AI means for SaaS, who gets to capture the value in SaaSpocalypse. I am just going to address that directly. We do not believe that AI eliminates the value of platforms and applications. We believe it does raise the bar for how you convert that intelligence into real customer outcomes. To do that, you need context. Here is our thesis. Our thesis is that intelligence is necessary, but it is not sufficient. Context is what turns that intelligence into measurable customer outcomes within businesses. AI models are trained on vast amounts of publicly available information, and they are going to continue to get better. Absolutely no doubt about that. No matter how intelligent a model gets, it does not know what is happening within a business, what is changing in the pipeline today, which deal is stalled, which customer is at risk. All of that information is not just static, dumb database. It is a dynamic, smart substrate with workflow logic that maintains state, rules, permissions, governance, and audit trails that make all of that information trustworthy for businesses. That is the context that HubSpot delivers. This is not just our belief. We actually quantified the value of context. We went across our customers that are using AI, and we measured how context impacts the outcomes. This is what we found. We actually found that good context delivers great outcomes. We saw this with every one of the key metrics in marketing, sales, and service improve with good context across our customers. You can see that. MQLs generated up over 200%. Deals closed and won up. The number of customer meetings booked up. It takes a lot to actually move these metrics, but good context actually drives these types of outcomes. Here is the part that actually really surprised us. We found that AI with bad context is worse than having no AI at all. This really surprised us, and you can again see every one of the metrics here deteriorate with bad data and bad context. This is because I think as humans, we know how to control for this bad context, but AI does not. It might recommend a product that is no longer in your product catalog. It might actually use an inaccurate definition of ARR. It might route an approval to a manager who just changed roles 2 days back. That is why we believe that even as models improve pretty significantly, the context within a business becomes even more valuable. This is HubSpot's structural advantage, and this is why we are democratizing AI across go-to-market for scaling companies. For just over 20 years, HubSpot has done one thing consistently. We have excelled at democratizing complex technology to bring it to scaling companies to help them grow. And in that process, we've become domain experts in both go-to-market as well as SMBs. In phase 1, we democratized inbound. We helped tens of thousands of customers by giving them a playbook and a set of products to help them succeed in marketing. We created an all-new category called inbound marketing. In phase 2, we democratized CRM. We helped hundreds of thousands of customers by giving them a unified platform with data across marketing, sales, and service, and we made it easy to use. We went from a single hub company to a platform company with multiple hubs, and we crafted every hub organically. Now we have the opportunity to democratize AI. Our goal is to help millions of customers grow by providing them an agentic customer platform and making it easy to get the outcomes that they want with AI. This is our biggest opportunity yet as a company, and we're bringing all of our domain expertise in go-to-market and within SMB segment to make this possible. We're doing this from a position of strength. We have consistently grown our revenue. We've generated strong margins as well as free cash flow over the past many years, and will continue to build on that solid foundation. We also have a platform advantage that has allowed us to really bring the pace of innovation to our customers, and it has enabled us to grow our customer base to over 300,000 globally and drive multi-product adoption across that customer base. And finally, we have exceptional distribution. We talk to millions of growth leaders every single month across our channels, across social, across academy, across the events that we do, and we have a powerful ecosystem mode. We have nearly 4,000 solution partners that extend the reach of our sales as well as customer success teams. Our execution really helped fuel the growth and now is the foundation for the next phase of our growth. The question becomes: what's happening with customers? How are they adopting AI? The landscape with customers is changing pretty dramatically. We've now, over the last year, talked to hundreds of customers, looked at how they're adopting AI, and three big shifts actually stand out. The first is customers want to build. AI is democratizing building, and that's the fundamental shift. For many decades, businesses just had to adapt the way that they work in order to fit the software that they had. Now AI, thankfully, is reversing that, and that means if you're a marketer, salesperson, service person, you can actually build. You can build artifacts, you can automate workflows, you can create new experiences, all using natural language. No need to wait for weeks behind a developer sprint. But there is a clear way to build well, and building well actually requires connected customer data. It requires Growth Context, permissions, governance, security, and that is the platform that HubSpot is building to deliver. We want to be the platform where go-to-market teams can build. Next, what we're finding is that customers care deeply about outcomes. Again, the conversation is changing really rapidly. A few months ago, it was all about token maxing. Now our customers don't talk to us about token maxing. They want and care deeply about the outcomes that they want to drive within their go-to-market teams. And we love that because the outcomes within go-to-market are just incredibly clear. We help our customers build demand, win deals, and delight their end customers. That is the focus for us. Third, I would say that customers want proof of value as well as predictability in cost as they adopt AI. A huge shift, again, that we are seeing. In SaaS, the typical process was you evaluate features, you then implement the software, you drive change management and user training, and once the users begin to drive, they get the growth and the value out of software. AI is changing that dramatically. You can just turn up a trial and look at what the value you get and the outcomes that you get within your environment, and that means that there is need for proof of value before the commitment to software. These three shifts combined is really changing what customers expect out of HubSpot right now. This is exactly what our product strategy as well as our go-to-market strategy are designed to deliver. As we look ahead, we have a very clear strategy for AI transformation. We are not incrementally improving our business. We are transforming it across product, go-to-market, pricing, and how we operate as a company. This is what it means. It means that from a product perspective, we are moving from delivering features to delivering measurable customer outcomes. It means in go-to-market, we are not just selling software. We are proving value, and we are earning the right to expand and compound adoption from there. In terms of pricing, it is not about charging for access. It is clearly aligning to how customers get value and outcomes from HubSpot. Then as a company, we are scaling with people and AI. We are driving this full transformation from how we build, to how we sell, to how we monetize, to how we operate as a company. Very exciting transformation. Duncan and Jon are going to do a deep dive in terms of our product strategy as well as go-to-market strategy. So I want to take a few minutes to provide an overview across all of these four areas of transformation, starting with our product. Our product strategy is to lead with outcomes, and that means every product decision now starts with a super simple question. Can we deliver trusted, measurable outcomes, and can we do that fast? We have three clear focus areas that you are going to hear about today that will accomplish this strategy. First is that we are building an intelligence layer with trusted context. We are bringing together all of the Growth Context for customers in what you will see as a custom Context Home. This Context Home is what will provide our customers the visibility, the control, and the trust that is needed in all of the information that powers every one of the AI and agentic use cases. Really powerful. We believe this context is what will make every agent, every recommendation we make, and every outcome that we deliver compound over time. Then we want to drive measurable outcomes within Marketing, Sales, and Service using HubSpot agents and custom agents. Two ways. HubSpot agents are domain-specific agents that we build. At UNBOUND, we launched four new agents: Campaign agent, Content agent, Nurture agent, and Revenue agent. You are going to see this in action today. Over time, we will continue to expand the number of use cases that we support, as well as the depth of outcomes that we deliver through HubSpot agents. At the same time, we are making it super easy for our customers to build on top of HubSpot, build workflows, as well as their own custom agents. The combination of HubSpot agents and custom agents is what will drive clear, measurable outcomes within go-to-market. Finally, really excited about the all-new Breeze Assistant. This is the interaction layer. Our vision is that customers shouldn't have to navigate to screens, click on menus, and learn software. They can have a powerful conversation with all of the value within HubSpot. We have completely rebuilt Breeze Assistant. It is no longer just a chat spot. You can ask questions, you can build workflows, you can build custom agents, you can build artifacts. This is why we are super excited about Breeze Assistant adoption. That adoption has grown pretty significantly this year. The beginning of the year was 37%, and then at the end of August, it was over 55%. Over 55% of our Pro plus customers are using Breeze Assistant. That is great. At the same time, customer satisfaction of Breeze Assistant has actually improved from 54% at the beginning of the year to nearly 68%. We are very excited about the progress that we are making in this powerful interaction layer. If you step back and you sum it up, we are the trusted context layer as well as agentic layer. This is how we are transforming our product from delivering features to really delivering clear outcomes for our customers in this agentic era. Product transformation is great. It is absolutely the start. To capture the full opportunity with AI, we are transforming how we go to market and how we meet customers where exactly they are. Customers are buying AI very differently than they bought SaaS. There is a fundamental shift that is happening here as well. They do not start by comparing feature lists. They start by asking a simpler question, "Will this work for my business?" They want to see it work in their environment with their data. When they see that, they want to scale with predictable costs that are transparent to them. That changes how we think about going to market. We will land with focused use cases. We will drive outcomes very quickly and then continue to earn the right to expand from there. To do that, we are actually strategically leveraging our distribution advantage. We have millions of users who use our product every single day. We have hundreds of thousands of customers, and we have scaled go-to-market motion. We are using that reach to seed the right AI use cases with the best trial experiences, the best product experiences, and very targeted pricing offers. That first use case, the agentic use case, is strategically important to us because what we find is that when customers get measurable outcomes in that first use case, the second, third, fourth use case becomes just easy for them to adopt. If you take a step back, SMBs are very early innings of adopting AI, and that is our biggest opportunity. Our strategy is to seed the market, is to prove value quickly, and then compound adoption from there. To do that, we are also leveraging the strength of our partner ecosystem. Our partner ecosystem is our forward deployed moat. Delivering outcomes with an AI is not turnkey. It cannot get results just by simply turning an agent on. You have to look at the data foundation and the context foundation that you are building, redesign the workflows, reimagine how a business operates. That is exactly where our partner ecosystem comes in. This is one of HubSpot's biggest structural advantages. We have more than 4,000 solution partners, and they work with our customers every day to implement HubSpot, but also transform the go-to-market. They actually drive 25% of our sourced ARR. They help us close nearly 50% of our ARR and actively manage about 40% of our entire customer base. We are investing in this partner ecosystem by providing the platform and the tools so that they can custom build on top of HubSpot, by providing the right incentives to drive that motion, and also making sure that they are taking us into every industry vertical that they have the expertise in. We believe that companies that will win in AI will have a very strong ecosystem, and that is why we are continuing to build on the moat that we have here. Okay. With that, let us actually talk about pricing. Our pricing has also evolved alongside our platform. In phase one, we built inbound marketing and customers bought one hub, Marketing Hub. We had subscription pricing back then with an axis for contact tiers. Life was just simple. People knew exactly what to buy. In phase two, as we expanded into a broader CRM platform with multiple hubs, some had seats-based pricing, and the others had subscription-based pricing. Now, we are entering phase three as we democratize AI and deliver an agentic customer platform. Customers are not just simply buying software anymore, they are actually buying outcomes from the software, which means our pricing needs to continue to evolve to meet where our customers are in AI. Our vision for this next stage of growth is simple. A hybrid model of seats and credits. Seats provides access to platform data and AI capabilities. Credits, they scale with the value and the outcomes that we deliver. We want to make HubSpot easy to buy and easy to grow with. We have been delivering against this vision. Many of you know that we started by simplifying our seats model back in 2024. Specifically, we did three things. We lowered the seat price, we removed the seat minimums that we had for Pro as well as Enterprise, and we introduced a new seat called the Core Seat to capture the value of the platform. We saw clear results. We were able to add more customers. Those customers are healthier because they buy exactly what they want and continue to upgrade from there. We were able to capture the value of the Core Seat through the platform. Now, more than 25% of our Pro plus customers have continued to add Core Seats. That is up 10% year-over-year. At the same time, Pro plus customers have continued to buy more Sales Hub and Service Hub seats. In fact, the seat count has grown by 16% year-over-year, and seat ARR has grown 23%. That's what we're seeing in terms of the seat adoption. Now, as we look ahead, we want to be even more clear about our focus. We want to remove barriers to AI adoption. That is the focus in terms of how we are approaching pricing. In order to do that, we got to drive predictability in terms of the cost, as well as flexibility in terms of driving the right adoption. This is a big change for application buyers, right? In terms of predictability, we are tying our pricing as close as possible to the value that we deliver, as well as the outcomes that we deliver. At the same time, we're providing visibility through dashboards and the ability to control spend. Our customers can set up thresholds right now. In terms of the flexibility, we're making it super easy to swap between seats and credits, super easy to buy in chunks for credits. We are continuing to test, iterate, and learn within this model. For example, we have a pricing test and experiment that was running in Nordics that makes it super easy to buy seats and credits, and we'll continue to expand that pilot within just the EMEA market in Q4. As we test and learn, we'll continue to share the progress with you. Over time, we expect seats and credits to become the primary source of our growth drivers. Speaking of which, let's actually talk about our growth model. Our core growth levers that have powered HubSpot over the last many years remain solid and consistent. We're continuing to drive up-market expansion, and the number of large deals continues to grow. We will deepen our platform consolidation as well as multi-hub adoption, and we will land and drive volume in downmarket. We plan to do that by focusing on single use cases that are powerful, drive value quickly, and then continue to expand to multiple use cases from there. As we look ahead, we have very clear emerging levers for AI monetization. Seats monetizes platform and AI value when we assist the work that people in go-to-market teams are doing. Credits will monetize the AI value when we deliver work. As I talked about, we think about it as HubSpot agents consuming credits, custom agents consuming credits, as well as the APIs that we deliver consuming credits. So if we put this all together, customers grow with HubSpot, and as the value that we deliver with AI continues to grow, we expect our monetization to continue to grow. This is our durable model for growth. Okay. Let's finish by talking about how we are transforming as a company. AI is not just creating a new growth opportunity, it's really changing how we operate as an organization. We've embedded AI into how we grow, how we build, and how we operate and transform ourselves into being almost AI native with the speed of that. Our product organization is just completely all in on AI, and they are one of the most advanced in terms of using our own agentic harness, something you'll hear Duncan talk about in a few minutes. Nearly all of our developers are using agentic coding, and as they do that, the lines per developer has increased more than 123%, and you'll see the interactions within our internal harness also continue to improve. In terms of go-to-market, we are an AI first go-to-market organization. From a marketing perspective, AI now handles almost all of our inbound chats. From a sales perspective, prospecting agents help us set up the meetings as well as continue to help our teams drive more win rate. In support, AI has enabled us to scale customer support without scaling the headcount. As an organization, we have completely transformed. Everything happens in six-week sprints with continuous experimentation. All of this means that we are committed to driving durable growth as well as profitability, not by choosing between one or the other, but by really transforming how we operate as a company. To bring this all together, we know that the outcomes era will really be won by companies that can take that intelligence from AI and convert it into clear, measurable outcomes for our customers. We have clear structural advantages to win. A trusted platform with the right context that can drive customer outcomes, distribution that can accelerate adoption, an ecosystem moat that helps our customers transform with AI, and an AI-first operating model that helps us balance growth as well as profitability. The combination of these structural advantages as well as the deliberate transformation that we are doing is what will drive durable growth as well as long-term shareholder value. Thank you so much for your support. I look forward to your questions during Q&A. One of the big things about using Breeze across our entire portal, whether it is Sales Hub, Service Hub, Marketing Hub, operations, is that it knows our brand voice. It has that context of how we talk and where we want to use different terms. Last week, I was building a marketing campaign from scratch, and I told Breeze kind of what I was envisioning. I was having an off day, not really feeling creative. It took the context from previous marketing events and other things that it knew about our portal and literally nailed the brand language, the social media assets, the ads, the emails. It just kept on building based off that previous context that we had. One of the most awesome things about Breeze is for new hires. That new hire is able to navigate to a contact or company record and day one knows the context on that exact record. In the past, they were looking through shared email inboxes, but now they are just presented with all the real-time context from Breeze directly in HubSpot. Reps are always just flabbergasted by it. When their time is freed up, sales can just sell, allowing our teams to be proactive versus reactive because they know that they can just check HubSpot and all the data is right there already for them. Please welcome to the stage Duncan Lennox. Hi, everyone. It's great to be here with you this morning. Yamini just made one thing very clear, AI needs context to deliver outcomes, and that's where I want to pick it up. My job today is to show you what context-driven AI looks like inside our product and how we're building HubSpot to democratize AI for scaling businesses. It starts with our vision for the outcomes era. Build the best agentic go-to-market platform that delivers an easy, fast, and trusted experience that customers love. To do this, over the last year, we've rebuilt much of our platform from the ground up. We've reinvented the CRM for the outcomes era. It's still the beating heart for go-to-market teams, but it needs to drive outcomes through people and agents, not just be a system of record for tools and workflows. We added a brain to make it all work seamlessly together. Let me show you what our new platform looks like, starting with how we've rebuilt the CRM. The CRM is really three things. It's data, trust, and intelligence. As always, it has to start with unified data at its core, bringing structured and unstructured data together. This, of course, is what customers have always relied on HubSpot for. Next, we built in security, governance, observability, and quality, which is foundational to trust, which of course is more important now than ever, especially as agents and AI are taking action on someone's behalf. Finally, we've given the Smart CRM a new brain with two key components, our agent harness that we call Aviator and the Growth Context graph. This allows us to give customers what they've always wanted, a CRM that updates itself. On top of this foundation is the action layer. This is where go-to-market outcomes across marketing, sales, and service are delivered, whether that is through the dedicated apps that HubSpot has always been known for, like Marketing Hub and Sales Hub, or the agents, both those that we build ourselves at HubSpot and custom agents that customers and partners are building on our platform. They all run on the Aviator harness in the Smart CRM. Then there is Breeze Assistant as the interaction layer. This is not the Breeze that you may know. We have completely reimagined it as a go-to-market expert that does the work for you. It figures out the best way to achieve your outcome, choosing the right tools and agents to get the job done. This fundamentally changes how you work in HubSpot, bringing all of the power of agentic AI to users to drive the highest quality, best value, fastest outcomes, all without them needing to know anything about AI. When we talk about democratizing AI, this is exactly what we mean. Finally, we want customers to be able to get the best go-to-market outcomes from HubSpot, regardless of how they choose to engage with our platform. We will get into all of the details about how we optimize for agents running on HubSpot and agents running HubSpot as part of our headless strategy. Okay, let us start with a look under the hood of the Smart CRM. I want to make an important distinction here. The CRM is not simply a system of record. It is a system of context, the intelligence layer for your go-to market. That is the bet that we have made on CRM for the AI era, and we rebuilt it with two defining elements. First, our new agent harness called Aviator, and secondly, rich growth context built in. I want to dive into both of these. Let us talk about Aviator, our agentic harness. A lot of people are throwing around this term at the moment, but few people are talking about what it actually takes for it to truly work. We have thought deeply about this, which is why we rebuilt the CRM with the agent harness instead of bolting it on to an old system that was not built for AI outcomes. I want to talk about what it is and how ours is different. At its simplest, an agent harness is designed to let you run an execution loop, calling an LLM to do work repeatedly until you get the desired result. In practice, that is a pretty inefficient mechanism and delivers slower and more expensive results that often miss the mark. Our harness is deeply integrated into the heart of the CRM with quality, trust and safety, and governance. How we built Aviator is key to our ability to deliver the best go-to-market outcomes. First, we made it model agnostic. We think of this as taking the Switzerland of AI approach. Not every model is good at everything, and it changes on an almost weekly basis. We are highly opinionated on which should be used for which task. Next, we use our domain expertise in go-to-market to build evals that run continuously to pick the right model. That means we know what great looks like. Our customers do not need to know or care about any of this. They can just be assured HubSpot will deliver the best outcomes for them, even as the state of models and AI continues to evolve rapidly. The second part of our unique strategy comes down to context. In the outcomes era, good data is necessary, essential, in fact, but it is not sufficient. Building on our unified data model of structured and unstructured data, we developed the Growth Context graph. It computes insights and intelligence from across a customer's unified data. This intelligence then allows us to use precise, high quality context in real time for a given task or outcome. Growth Context has three dimensions that work together. The first is business context, what you do, everything that makes your company yours, your brand and voice, your product and positioning. Next is team context, who you are, how you and your team work and win together, your roles and goals, your workflows and processes, your methodologies and habits. Finally, of course, customer context, who you serve, the full history and nuance of every customer relationship, your ideal customer profile and personas, your communications, contract, and revenue details. A point solution might have one part of this, but that is why they hit a ceiling. In HubSpot, all three live together, and now that is in the Smart CRM. The Aviator harness uses Growth Context to provide the right context for a specific task at the exact moment that it is needed. This is so important because the wrong context adds noise and cost, which is how businesses find themselves wasting money on tokens with minimal ROI. Our unique combination allows us to deliver the highest quality at the lowest cost. If at any time the best outcome favors another model, we route accordingly. Okay, so we pick the right model, we inject the right context, all without customers having to do anything. It is all built into the trust layer of the Smart CRM, which has security, including things like role-based access, governance for things like permissions and approval workflows, observability for full visibility into what every agent did and why, and quality, including safety and evals. This is what makes us the go-to-market harness for scaling businesses. Now, I want to show you how customers experience all of this in HubSpot. Yesterday on stage, I unveiled our all new AI native CRM. It is the CRM that you know and love, now rebuilt for the outcomes era. This is because our agent harness now orchestrates and manages action, which changes everything. This new CRM works on our customer's behalf and updates itself. This will save customers hours of manual updating, the headache of forgetting to add details, but most importantly, it will give them accurate insights at the moment they need them. Those insights are stronger and more proactive than ever because we built Growth Context directly into the Smart CRM. We did not stop there. Businesses need to be able to see and control their Growth Context to understand what it is powering and how it is delivering outcomes. So we built Context Home as their one-stop shop. This is a new capability that lets customers see everything HubSpot knows about their business and manage it directly. It also gives customers a score on how complete their context is, along with recommendations for where to add additional context that will let us deliver even better outcomes for them. All they need to do is upload a file or point us at a source, and HubSpot will do the rest automatically. All of this context is available to every agent and every AI feature across HubSpot the moment it is relevant, from our own HubSpot agents to custom agents customers build on our platform with Agent Builder. I want to show you how this comes to life for customers in something that they use every day, email. This is what a generic email written by AI that has no context might sound like. We have all seen emails like this. It is AI slop. This is what happens when AI has the complete, precise context it needs. Let me point out the pieces of Growth Context that went into this email. First, from business context, we knew the product that was trialed, how it is positioned against the competitor, and even the Gartner comparison to show that. From team context, we knew how the team typically follows up after a trial, the rep's unique voice and tone, and even that UNBOUND was the right moment to push for a meeting. From customer context, we knew this customer's specific usability and price concerns, the team member that they should loop in for next steps, and that Percy pricing was particularly relevant to their company size. Every sentence in that email earned its place and helped our customers drive better outcomes. The user didn't have to futz around with figuring out what docs or other context to upload or what prompt to write or which agent to use. Let me recap what we have covered. We have rebuilt the Smart CRM as the foundation that makes everything else work. Our agent harness delivers high performance at a lower cost. Growth Context delivers accuracy and control. This is the intelligence layer, the new brain of HubSpot. Now let's see how it powers action. When it comes to driving outcomes, agents are the most effective way to scale how businesses build demand, win deals, and delight customers. In the past 6 months alone, we have built several new agents and tools in HubSpot aimed at driving these exact outcomes. I will get to those in a minute. First, I want to talk about another foundational update we have made to HubSpot, how our customers actually use those agents. Yesterday, I introduced the all-new Breeze Assistant. Many of you know this tool. It has been on HubSpot since we first started building with AI. But what we unveiled is a complete overhaul. It might look similar, but just like the all-new CRM, it has been rebuilt for the outcomes era. You now get work done just by talking to Breeze. Instead of navigating menus or learning workflows or even having to understand the capabilities of our agents, customers just describe what they are trying to accomplish, and Breeze takes action across the platform on their behalf, pulling in the right tools for the job. Customers tell us that they don't have time to become AI experts, and Breeze Assistant means they don't have to. Now, I know many of you watched the spotlight from yesterday, and I definitely encourage you to do that if you haven't because there's some really exciting stuff in there. I want to highlight some of the new tools and agents we launched and how our customers use Breeze Assistant to drive outcomes. For marketers, the outcome is building demand. This has never been harder. Organic reach is declining. Buyers are asking AI instead of filling out forms, all while the volume of content needed to cut through the noise has gone up and ROI has gone down. Marketers need a dramatically new way to build demand, so that's what we created. Our all-new Marketing Studio is their single home where they can just have a conversation with Breeze and then go from insight to a live campaign in minutes. To start, they just ask Breeze for a full campaign plan. It pulls in our new Campaign agent to build it, everything from messaging angle to an asset plan and audience segmentation. Next, we've made it easy for marketers to create their campaign content just by asking Breeze. It calls on the new Content agent, which takes the campaign plan and drafts the assets in the correct brand voice and targeted to the ICP. When marketers then have their campaign up and running and need to nurture the leads, they just tell Breeze. It pulls in our new Nurture agent to write automated emails for every contact using CRM data and engagement history. With a new home in Marketing Studio and Breeze Assistant as their guide, we're helping marketers build demand all from a single conversation. Now let's see a sales journey where the outcome is, of course, winning deals. We know reps are spending more time than ever on admin and less time actually selling. Deals stall, follow-ups get missed, and context doesn't carry between conversations. We're giving sales teams valuable time back to win deals, and they start, just like our marketers, by talking to Breeze Assistant. Let's say they want to break into a big prospect. They just tell Breeze, and it calls on Prospecting agent. Immediately, it pulls the buying committee and drafts personalized emails for each person. We already saw the difference Growth Context makes in those emails. Next, when the reps meet with prospects, they want to capture the details from the meeting. They can use our Meeting Notetaker to dictate updates so nothing gets lost. Deal progression feeds those updates back to the CRM automatically, so deals keep moving and reps keep selling. No other platform makes it this easy to prospect, research, coordinate, update the CRM, and land the deal. To summarize, what all of this means for customers is a platform that works for them, a go-to-market harness that selects the right model and context, a CRM that updates itself and delivers insights, and a single tool in Breeze Assistant that launches agents just by chatting. Together, these democratize AI for scaling businesses because they take the burden away from teams so they can focus on their customers. Now, in a world where customers are increasingly using agents, we're very aware of a different customer challenge, and that's agent sprawl. Leaders are losing sight of what's running, what it's doing, and whether it's working. This leads to difficulty managing the agents, higher TCO, and a poor customer experience, all for agents that don't actually work together. We believe there's a better way, and customers will want a single view into it all. We built Agent Hub to help solve that. It's one place to see, manage, and build agents across the entire go-to-market, and they all operate from that shared context, which a collection of point agents can't do. Agent Hub is where customers gain confidence that their HubSpot agents and custom agents are actually delivering the outcomes their business needs. We're also seeing a new kind of go-to-market team member emerge, the builder. These are people who want to quickly build agents that extend what they can do on HubSpot, whether it's as simple as qualifying every lead before a rep touches it, or as sophisticated as monitoring your entire pipeline and flagging at-risk deals before anyone has to ask. That's exactly what we're making possible with our new Agent Builder. This lets customers and partners build agents that run on Aviator and use Growth Context just like our own HubSpot agents. You can build all of it just by chatting with Breeze Assistant. Our last layer is all about extending the way customers and partners interact with our platform. I want to spend some time here because I think it's one of the most important things HubSpot has built. There's two parts to our ecosystem strategy. Agents can run HubSpot using our data and intelligence as a building block. Think of custom agents businesses can build on our platform. Agents can run HubSpot, operating the platform end to end through APIs without a human ever logging in. This is what most people refer to as a headless CRM. We're unique because we think about our API in layers. First is the data APIs. This is raw access to everything in the CRM. Over 1,000 public endpoints covering contacts, deals, sequences, content. We're committed to giving API parity with everything you can do in the UI. We want developers building on us, not hitting walls. The second are our context APIs. Instead of raw data, you get insight that we've already computed. Think of it as asking HubSpot a question and getting an answer instead of a spreadsheet. This is actually the same layer our own HubSpot agents run on. That's how good it is. The third and most important are work APIs. This calls either a HubSpot agent or a custom agent to drive an outcome end to end. Let me give you an example of how this works. Imagine you're a rep and you want to know which deals in your pipeline are at risk. With the data API, you'd pull all your deal data, bring it into an external third-party tool, and f igure it out for yourself. That's a lot of data and a lot of work. You're also going to burn a lot of tokens. With the context API, you say, "Give me the 10 most at-risk deals in my pipeline," and you get that insight back directly. With the work API, you can say, "Figure out which deals are at risk and give me an action plan." You will get the 10 deals and a unique action plan, next best action for each one. We believe deeply in our agentic ecosystem strategy, and we have been executing on it for the last 18 months. Over a year ago, HubSpot became the first CRM to support MCP, so agents can interact natively with our software. We did this first with Anthropic. Then in June 2025, we were the first CRM to launch a deep research connector with ChatGPT. In July of that same year, we became the first CRM connector with Claude. In May 2026, we shipped the Agent CLI, a completely new layer that lets agents automate bulk scheduled work on HubSpot with no human in the loop. By mid-2026, we were the first CRM to have connectors with all four leading frontier models, adding Gemini and Copilot. Today, we are the number one CRM connector on both Claude and ChatGPT, with weekly active users on ChatGPT up 250% since June. Yesterday on stage, I announced that HubSpot is the first CRM to integrate with ChatGPT Ads. We launched the AI Growth Bundle that gives new customers 65% off HubSpot Starter and HubSpot Credits, a free ChatGPT business seat when they buy one, and up to $750 spend match on ChatGPT Ads. By connecting to the deep context in HubSpot, ads become more powerful and customers get attribution that is only possible by integrating their CRM. This is what real momentum looks like. These are not just announcements. These are real products with real users and real growth. I want to wrap it up. We have heard a lot today, but I want to talk about what actually makes HubSpot different. AI that runs on a foundation rebuilt with the right harness, the right context, and the right governance, it has a compounding effect. Every action makes the next one better for a customer. That is what Aviator and Growth Context give us, and no one else has that foundation. Second, we made it easy for customers. Most agentic platforms give you power and then ask you to go figure it out for yourself. Breeze Assistant means customers do not have to. They just describe what they want. Agent Hub means they are not losing track of what is running, and the CRM updates itself, so go-to-market information is always relevant. We are not just making AI available, we are making it usable, and that is what it takes to democratize AI at scale. Third, we do not believe in wall gardens. We believe in customer choice and meeting them where they are. We have used these principles to build a three-tier API strategy. Finally, trust, which is at the core of everything we do. Every company building AI right now is learning that customers are nervous. Businesses want to know what is running, what it has access to, what it did. Trust is about transparency and control, and that is what we have built into every layer of the agentic customer platform. From Aviator and Growth Context in the CRM to Breeze Assistant and our own HubSpot agents, so customers can trust us as their go-to-market harness. Here is where all of these things we have talked about add up, the outcomes. Customers using HubSpot with rich Growth Context across the platform are seeing real results from AI today. They're seeing demand built with nearly 4 times more MQLs. They're winning deals at over 3 times the rate, and they're delighting customers with twice as many tickets closed. We set out to build the best-in-class agentic customer platform that drives go-to-market outcomes that customers love. We believe that what we've shown you today and introduced to our customers this week not only meets the moment, but gives businesses the platform they need to grow well into the future. Thank you so much. The biggest and most effective aspect of the HubSpot ecosystem is the automation and the possibilities and the flexibility. At the moment, when I just checked the numbers, we got 39 meetings and 34 deals created through Prospecting Agent. I think our response rate is somewhere around 8%. The overall pipeline impact, we're talking about something like 40 deals that have been created in the pipeline this year. Mainly, this has been spearheaded by people using the Prospecting Agent. It was pretty easy to see right away that this actually is doing a job of a brilliant SDR. Inevitably, that's going to spend a lot of credits. I think we have spent roughly 400,000 HubSpot Credits as we speak this month. I would say that we're quite well on track in terms of spending almost all of them this month. From my perspective, they are money well spent. Please welcome to the stage, Jon Dick. Thank you, Duncan. So great to meet you all. How are you? Nice to see you. Thanks for coming to UNBOUND. I'm Jon. I'm the Chief Customer Officer at HubSpot. I lead sales and customer success. I'm a longtime HubSpotter. I've been at HubSpot for about 10 years now, and I've worked across pretty much every part of our go-to-market. Most recently, I led our customer success team. Leading that team, I had the opportunity to get to know our customers deeply, their goals, their challenges, and their hopes when they choose HubSpot. I think that our customers are going to absolutely love the HubSpot that we're building for the outcomes era. Not only do I think that the product that Duncan just talked to you about is going to absolutely blow them away, but the transformation of the business model will align HubSpot's growth with our customers' growth in a way that was just never possible before. It's good for customers, and it is good for us. My mission as the Chief Customer Officer is to build a go-to-market that meets the moment for the outcomes era. It means democratizing AI, delivering exceptional results for customers, and ensuring durable growth in the process. That's what I'm going to talk to you about today, how our customers are changing, how our go-to-market is transforming, and how we're going to drive growth both down-market and up-market. Let me just start with the customer. In the SaaS era, the customer journey was very well established. You've seen it before. It looked something like this, and it was a journey that was built on establishing upfront promise. What can this solution possibly deliver for my business? This has all changed with AI. In the outcomes era, the customer journey is built on upfront proof. What will this solution deliver for my business? I think that new journey looks more like this. Let me just talk you through it at a very high level. The customer journey used to begin when buyers started doing research on their own. They found a whole bunch of solutions for their problem. In the new journey, buyers are outsourcing this. They're outsourcing it to LLMs and to influencers, and the whole thing looks very different. They're being given a short list of solutions to consider for their problem, and they're just showing up much more ready to buy. Buying also looks very different now. Buyers are no longer interested in spending a tremendous amount of time upfront evaluating if a solution works for their business. Instead, they want to jump right in. They want to start building for their business. They don't want a demo from a vendor. They want the vendor to help them actually build their first use case on the product in their environment. That first use case then becomes the business case. Instead of developing a theoretical ROI and committing based on promised value, buyers are able to look at actual proof and have confidence in the outcomes that the solution will provide. With that confidence, they grow, and their spend grows too, not only by spending more in that first use case as it scales, but also by adding other use cases, because in this new customer journey, one use case leads to the next. This new customer journey has really inspired us at HubSpot, and it's caused us to really obsess about two big things in our go-to-market. The first is how we drive distribution so that more companies consider HubSpot on day 1 and continue to consider us on day 100. Second is how we engage customers to help them actually build that first use case and see value immediately. Let me start with distribution. Distribution has been an advantage for HubSpot since our founding, and it continues to be a real cornerstone of our go-to-market strategy. The big shift, of course, is that buyers increasingly rely on more trusted sources to make purchase decisions, especially when it comes to AI products. They bypass the traditional channels, and instead, they look to the LLMs, the influencers, communities to learn what they should be buying. We've been ahead of this shift at HubSpot, and we've been diversifying into these channels for several years now. Our YouTube and newsletter reach, for instance, beats major business media publications. And you can see that with the live stream of our spotlight from yesterday. We're already at 1.65 million views on YouTube, well ahead of other keynotes that have been posted. Meanwhile, our global aided awareness for HubSpot up-market across core geos and segments is at all-time highs, and we continue to be number one in AI share of voice for the CRM category across all of our major markets. This is our playbook, and we are continuing to invest in this. But distribution, of course, isn't only about new customers because the other place that buyers look are to the companies that they already work with and they already trust. With over 300,000 customers, of course, we have a substantial distribution advantage and a trust advantage within the SMB space. And as you know, we've been successfully leveraging this advantage when it comes to upgrade and multi-hub expansion for some time, and that continues to go well. But we're also now focused on leveraging our existing customer base for the emerging AI adoption opportunity. We are actively working to get as many of our customers building their first AI use cases on HubSpot, and we're making really good progress with that. 19% of pro plus customers are using HubSpot agents as of August, which is up 2X since the beginning of the year. We're working with them actively, and so this brings me to how we are doing that. We are transforming customer engagement at HubSpot, and we're putting our focus on helping build that initial use case, helping customers achieve value, and then growing from there. What this looks like in practice is fewer handoffs and more help. To do that, the lines between sales and customer success really start to blend in a whole new way. That's why we combined the sales and the customer success team under my leadership earlier this year, why we've aligned everybody in sales and customer success on customer outcomes, and why we are rebuilding our motions with customers. But we, of course, are not the only ones who help our customers at HubSpot. Our engagement model also includes our partners, who are really an extension of our sales and customer success teams. They act as HubSpot's forward deployed team. Our partners are a major advantage for HubSpot, and we are leveraging them to help our customers adopt AI. Partners have the industry experience, the technical expertise, and the change management resources that our customers need in this moment. Accordingly, we are doing a lot more with our partners. We are doing more co-selling. We are doing more co-delivery. We are rewarding them for their efforts by providing incentives for customer AI adoption, and it's paying off. When a partner is involved in an up-market deal, we see nearly an 8-point higher win rate, 75% higher ASP, and a 14-point lift in net revenue retention. So as you can see, when partners are involved in selling and expanding, it really works for driving value for our customers. We are going to continue to do that. Now let's actually talk more about our customers, and I want to start with a group that's been the foundation of the HubSpot story for our entire 20 years, which is small businesses. What we're seeing right now down market is that customers have more options than ever, but the same challenges. They're more price sensitive, they're time constrained, they just need easy ways to grow. That is what we have always delivered at HubSpot. I think by what you saw from Duncan, what we will continue to deliver, because we've built them an even easier way to grow in the age of AI. No more data entry, no more managing omni-channel complexity, no more learning the software. I've been using HubSpot Meeting Notetaker on the HubSpot app all week in meeting with customers, and it's been absolutely great. They can build and scale on HubSpot without friction now. This gives us an opportunity to reintroduce HubSpot to the down-market customers in a whole new way. The first thing we're doing is we're positioning ourselves differently. We are obsessing much more about the outcomes that we drive for customers than the technology that drives it. That's what our customers care about. They care about building demand. They care about winning deals. They care about automating all the stuff they hate. Second, we are ramping up our promotion in Q4, and we are pursuing new partnerships, and helping more customers see this whole new and improved AI native HubSpot. Third, we're being more flexible with our customers. We are letting them shift their spend between seats and credits, just so it's easier to build, prove, and grow. So better product, better positioned with the pricing flexibility that customers expect. Now, let me tell you two stories. I'm going to tell you two stories of down-market customers who built and quickly saw value with HubSpot. The first is Phin. They're a security awareness and phishing simulation platform. They're an existing HubSpot customer who had a common problem. Their reps were spending more time doing prospect research than actually prospecting. The reps were spending their time combing through thousands of emails and deals, searching for insights on who to re-engage, what messaging worked, how to personalize their outreach, and they knew that they needed a better solution. They did what any modern buyer would do in the new customer journey, and they started building. But they didn't just build on one solution. They built with a frontier lab, they built with a point solution, and they built with HubSpot, three different experiments to see which was going to work for them. HubSpot won, and we won because of our trusted context. Prospecting agent went out. It researched all the old deals by pulling notes, the contact details, the call history, pulled it all from HubSpot. It drafted relevant re-engagement messaging for the reps. Now, you saw from Duncan the difference between an email written by generic AI versus one written by HubSpot with Growth Context. It should come as no surprise that Phin's email response rates more than doubled on HubSpot, and they won multiple deals that had previously been written off. These results were great for them. It led them to 8x their credit commitment with us, and they are now on track to continue to increase their spend this year. I just love this story because we won on proof, not on promise. Another example of a down-market customer that I wanted to share is VisionWave Holdings. They are a microcap defense technology company, and their COO had a problem. He had a major defense trade show coming up, and he had no way to collect leads at scale. He needed outcomes in literally days, not months, so he did not have time for the old traditional customer journey and to go through that whole process. He just came to HubSpot, and he came to HubSpot ready to get started. He got onto our free product, and he began building. In one week, using only Breeze Assistant, he built a custom lead registration system for the trade show. He configured his sales processes, and he ingested the company's existing leads and deals. The COO described the experience as the fastest and best CRM implementation in his 30-year career, and he upgraded to our enterprise product touchlessly because of it. As for the trade show, according to him, VisionWave got an overwhelming pipeline of deals from governments and tier 1 contractors around the globe. I love this story because it shows the power of Breeze Assistant in action for our customers and how powerful it can be, especially down market. Let me tie it all together. With our new AI native CRM, our value proposition down market is easier to use software, faster time to value, and better outcomes. Let me switch gears now and talk about the fastest growing part of our customer base, our up-market customers. What we are seeing up market is more deliberation than ever before. Honestly, I get it. It makes sense. It is a once in a generation platform shift, and companies want to make sure that they get it right. What this is showing up as for us is more scrutiny from buyers, larger buying committees, and a requirement to just de-risk the implementation. The good news is that despite this, our product and our go to market have never been stronger for an up-market buyer. As Duncan showed, we have a legitimate platform advantage. Our trusted context drives better outcomes at a lower cost, and our headless strategy future-proofs HubSpot for our customers. Customers, when they come to HubSpot, can be confident that we will scale with them despite a changing AI landscape. Meanwhile, as you saw, the core marketing, sales, service, and ops use cases, they have all gotten more valuable for our customers. On the go-to-market side, we have more credibility up market, more logos, more proof, more peer companies willing to be references for our new customers. All of this together means that we just have a real opportunity up market. Here are the new ways that we are meeting the moment. The first place we are focused is the initial sale. We know these up-market buyers are more hesitant and have higher expectations, so we have updated our sales playbook to match. We are doing more account-based marketing. We are involving C-suite decision-makers earlier, and we have built out a dedicated, specialized sales team for service and CPQ use cases to help those buyers. To de-risk the sale, we are investing in more resources to ensure that our customers see tangible value from their first use case as quickly as possible. What this looks like in practice is funding more partner services, building outcome-oriented implementation plans, and offering HubSpot go-to-market engineers to support their AI setup. To help drive expansion into AI, we are partnering with our customers much more deeply than we ever have before. We are doing discovery focused on how their core go-to-market processes work. We're mapping their problems to agentic solutions, and we're granting them credits to actually trial our agents to solve those problems. We are embedding our HubSpot partners into the delivery and ongoing success of all of that as well. So a better playbook, more protection upfront, more partnership long-term, all guided by greater personal engagement up-market. We're already seeing the results. We are competing in and winning much bigger deals over $120K of ARR, up 38% year-over-year, and our partner attachment to up-market customers has grown as well. Now, let me walk you through two stories of up-market customers that I really love as well. These are stories of customers who are driving incredible outcomes with HubSpot. So the first is Nations Roof. Nations Roof is one of the largest commercial roofing contractors in the U.S. They've got over 40 locations across the country. So they originally came to HubSpot for marketing. They needed more leads, and they needed to automate more of their marketing activities. While the marketing team got going on HubSpot and experienced what it was like to engage customers with HubSpot and experienced the ease of use of HubSpot, their sales team was on a different planet. They were on an incumbent platform that they referred to as a Franken system. They told us it was over-customized, it was over-expensive, and it was so painful to use that reps simply avoided it. I hear this story multiple times a week, by the way, from different companies. Within a year, the CRO decided that it was time to make a change. So he went to the market looking for something different. I remember meeting him. It was very distinctly the way that he described what he was looking for and his hopes and dreams in a new CRM platform. He said, "I just want a CRM where the reps don't have to do data entry. That's what I want. That's what I need." He got excited as I started talking to him about what we were building with our Meeting Notetaker and our self-updating CRM and our mobile note-taker. He also got excited about the idea that a sales and marketing team would actually work out of the same system and could work together. He got the most excited about the lower total cost of ownership that he would have if he consolidated all of it onto HubSpot. So that's what they did. What started as a marketing use case quickly turned into four enterprise-level hubs and 50,000 additional credits, bringing their annual commitment with us from $10,200 up to over $300,000. The next example I want to talk about is a company called Aventus. Aventus provides software that helps financial institutions detect market manipulation and meet regulatory requirements. The firm sells into one of the most compliance-intensive segments of fintech. The cycles are long, the buyers are technical, the addressable market is small and very high value. This means that prospecting for them requires extensive research, depth, and personalization. That doesn't mean it needs to be done by a human. They had a vision for building a prospecting system that would augment their sellers and not replace those sellers, one that could do the research, track the signals, and work with their reps. To their credit, they jumped right in. They pulled some engineers and were like, "Let's just build this prospecting system." They went at it, but when they learned that HubSpot had built a prospecting agent in the same system that their sales reps already worked in, they decided to give it a try. They made a small credit commitment, and they ran a pilot on HubSpot. They loved the early results. What they loved was that each rep got an agent that was trained on their voice and their selling style, and that agent did all the work. It pulled the context on where a lead came from. It pulled the engagement history. It pulled the campaign goals to draft an outreach. All the reps had to do was review it and approve. Based on this, they made the decision that it wasn't worth the cost for them in dollars or engineering time to replicate what HubSpot had already built. They went all in on their first use case, and they started to scale. Since then, their teams have seen a 27% increase in influenced pipeline. They've got back 87 selling days across the team. Because of it, they've also began to expand into new agent use cases. They are now enriching, tagging, and routing good fit accounts with Data Agent, and they're building a big ABM program for their account management team for a top 50 ABM approach. In the last four months, they've increased their credit commitment 3.3X with HubSpot, and they are still expanding. In both cases, our upmarket value proposition of an easy, fast, trusted platform with lower TCO delivered exceptional value to our customers. As I hope you can see, we have an immense opportunity up-market and down-market, and we are executing against both of these segments very thoughtfully. I want to close where I started. It is a noisy time in the paradigm shift, and there are lots of options, but HubSpot is positioned to win for the outcomes era. Against the incumbents, our AI native capabilities and our trusted platform deliver on buyers' expectations with a CRM that is easier, faster, and provides better outcomes at a lower total cost of ownership. Against custom-coded tools and everybody who wants to vibe code their own CRM, our message to customers is that starting on this path is easy, but scaling it is exceptionally hard, and HubSpot does the hard stuff really well. We provide the governance, the compliance, the maintenance, the integrations, the model neutrality, the ongoing innovation, everything you need to confidently grow, and we do it without the complexity and cost of doing it yourself, and we make it available in whatever way you want to access it. Against AI point solutions, we think history will repeat itself here. In the software era, nobody wanted a Franken system. In the AI era, we don't think anybody will want agent sprawl. Our unified platform has both breadth and depth. It delivers the trusted context that every agent needs, and it scales exceptionally well. This era is going to reward the platform that delivers outcomes the fastest, and that platform is HubSpot. That's why we're leveraging our position to win more customers and drive more growth. Kate Bueker is now going to take you through our business performance. But first, I'd like to share another example of a customer getting outsized value from HubSpot. Thank you all very much. Customer Agent is our best ally now. We are managing a database of 2 million contacts with only 3 human agents. In Synergee, we are using 4.5 million credits. The reality is that we expect a higher volume of incoming conversations that will move onto Customer Agents. That, combined with the fact that the price of conversation resolved by AI was cut in half, lead us to aim for a number like that. According to our calculations, scaling customer service for the opening of 200 gyms without Customer Agent will have required approximately 12- 15 human agents. The return on investment is really amazing. Please welcome to the stage Kate Bueker. Thank you, Jon. Hello, everyone. It is great to see you all back here in Boston. You have heard from Yamini, Duncan, and Jon Dick about where HubSpot is headed as we transform with AI. I am excited to share how this is going to drive durable revenue growth and margin expansion in the years ahead. In terms of our agenda, I am going to start by sharing how the solid foundation that we built as we democratized inbound and CRM puts us in a position of strength to democratize AI. Next, I will walk through how that strategy drives new growth levers as we align our business to delivering outcomes for customers. Finally, I will share how we will drive margin expansion and earnings growth as we transform how we build, grow, and operate. Over the last 20 years, HubSpot has built the leading go-to-market platform for scaling companies and a solid foundation for growth in the age of AI. This foundation includes our financial scale, platform advantage, and exceptional distribution. It is what sets us up to win as we democratize AI. Let us start with our financial scale. HubSpot has a consistent track record of delivering revenue growth, margin expansion, and strong free cash flow growth. Since 2021, our revenue has grown at a compound annual growth rate of 23%. At the same time, we have delivered consistent margin expansion, expanding our non-GAAP operating margin by 12 points. Our cash position is strong at $1.4 billion as of the end of Q2. In 2026, we expect to deliver $750 million in free cash flow, representing a 30% compound annual growth rate since 2021. Next is our platform advantage. We have grown our customer base at a compound annual growth rate of 20% over the last four and a half years to over 300,000 today. We have also driven significant multi-hub adoption, with 70% of our install base ARR leveraging three or more hubs. Across our 300,000 customers, the scale of our CRM and context is massive, with 5.5 billion company records and 18 billion contact records. We have nearly 2,000 technology partners that have integrated into our ecosystem. Finally, the platform is unlocking new ways for customers to engage with HubSpot's platform. I know Duncan shared we were the first CRM to have connectors with all four of the leading frontier models, and we are the number one connector on both Claude and ChatGPT. Last, our exceptional distribution. We are one of the 10 largest business media networks worldwide, with over 350 million organic views over the last 12 months. We have a sales and customer support team that is 4,500 strong, and we have 4,000 solution partners that help extend our reach by helping customers build their go-to markets on HubSpot. Our distribution extends across the globe, with 50% of our ARR coming from markets outside the U.S. Our strong foundation sets us up to win in the AI era. Next, let us talk about how we will drive growth. Our growth strategy remains strong. Our core growth drivers are up-market expansion, multi-hub adoption, and down-market volume. Our strategy also enables a set of emerging growth drivers across seats and credits that align our business model to delivering outcomes for our customers. Let's dive in, starting with our core growth drivers, which should all look familiar to you. Our up-market performance remains strong, reflecting continued demand from larger customers for a unified AI-powered platform. We saw this again in Q2 with large deals growing 38% year-over-year. As a result, deals with over $120K in ARR represent 18% of our install base, up more than three times since the beginning of 2021. Our multi-hub momentum continues to grow as new and existing customers consolidate their go-to-market technology stacks on HubSpot. The share of our new Pro Plus ARR that is multi-hub has grown steadily, with more than 80% coming from customers with two or more hubs and more than 55% coming from customers using three or more hubs. As a result, only 9% of our install base ARR is on a single hub today, and more than 70% use three or more. Down-market volume remains a key element of our strategy. You heard this from Jon Dick. As of Q2, we had 150,000 Starter customers on HubSpot, up from 50,000 in 2021 and representing about half of our total customers. We are transforming our down-market product to deliver an AI-native CRM that helps small customers achieve the outcomes they care about at a very low cost. We aim to capture customers early in their journey and grow with them, and we remain committed to this segment. Our current growth drivers are working. As we execute our strategy to democratize AI, we are seeing traction with a set of emerging growth drivers. These AI growth drivers include seats and credits. Seats are about making human users of HubSpot even more effective. Credits are how we monetize the outcomes we are delivering for our customers. Seat growth has been a meaningful driver for HubSpot. With paid seats up 80% since our 2024 pricing simplification. We had more than 2.5 million total paid seats at the end of Q2, including more than one million paid Core Seats. We're committed to making every go-to-market employee as productive as possible. As a result, we've continued to add value to our seats and notably our Core Seat with features like Breeze Assistant, HubSpot Meeting Notetaker, and now a self-updating CRM. Breeze Assistant adoption continues to scale, growing from 37% of Pro Plus customers at the beginning of the year to 57% in August. Breeze Assistant agentic actions are up 4.4x. We will continue to monetize the core value of our AI and platform capabilities, and we're going to continue to increase the value of seats and drive deeper penetration within our customers. Now let's talk about agents and credits. Duncan outlined our agentic strategy, which includes three opportunities: HubSpot agents, custom agents, and APIs. Each of these will consume credits and drive monetization over time. Across all of these opportunities, we will measure our progress with a consistent framework: reach, depth, quality, and growth. First, reach. How many customers are adopting our agents? Second is depth. How often are customers taking agentic action? Third is quality. What is the quality of the business outcomes we are delivering? Finally, growth. How is all of this translating into credit and seat consumption? Let's dive in. The first opportunity we see is with HubSpot agents. Our strategy here is to deliver best-in-class agents across marketing, sales, and service use cases. In addition to customer prospecting and data agent, we launched campaign, nurture, content, and revenue agents this week at UNBOUND. Let's unpack the early momentum we are seeing, starting with reach. 19% of Pro Plus customers using HubSpot used HubSpot agents as of August, and this is up 2 times from the beginning of the year. Over the same time, we have seen the number of monthly agentic actions increase by 3.5 times, reflecting the quality of the outcomes we're delivering for customers. We've seen monthly credit consumption more than double, even with the change that we made in April to lower agent prices and shift to outcome-based pricing, which created a near-term headwind. Equally encouraging, we are seeing credit usage increasingly diversified across our agents. Customers are not just adopting a single agent, they are beginning to use agents across the entire customer journey. Let me give you an example. Sesame HR is a 400-person software company with more than 18,000 customers worldwide. As the company grew, its support team struggled to keep up. At best, they could cover 70% of their tickets, leaving the rest unanswered. Instead of hiring more reps to handle the volume, Sesame ran a 1-month pilot with Customer Agent and they saw results. Within a few months, they were able to cover 100% of the tickets, with 60% of the tickets resolved through Customer Agent without any human escalation. Customers got a better experience and the support team had more time to focus on higher value work. The value is translating directly into credit consumption. Sesame quickly used its initial credits, and they purchased additional credits that represented about 100 times the initial amount. This resulted in $80,000 in agentic ARR. The company has been very encouraged by the success with Customer Agent, and they are testing additional HubSpot agents. Sesame is piloting Buyer Intent to help its sales team drive conversion rates and better outcomes, just like its support team. The second opportunity we see is with custom agents. As Duncan shared, the strategy here is to make it incredibly easy to build agents and go-to-market automation directly on HubSpot. Breeze Assistant and our solution partners will help drive adoption. Early, we have seen strong initial interest and engagement. Over 3,000 custom agents have already been deployed since we launched in July. The most common adoption and deals are things like internal workflows around deal management and revenue operations, places where customers can quickly save time and drive consistency in their process. We see a significant monetization opportunity with custom agents as our partners and customers build more agents and consume credits. Let me share an example. Cyber Advisors is a nationwide cybersecurity leader with over 400 employees. The company saw an opportunity to make its go-to-market more efficient while driving incremental growth, and they turned to HubSpot agents to help. To start, they launched pilots with Customer Agent, Content Agent, and Prospecting Agent, and they saw results. Prospecting Agent now sends a personalized email to every new lead before the team member picks up the phone. So far, more than $700,000 in closed revenue has been attributed to Prospecting Agent. Cyber Advisors are not just using HubSpot agents. They are building custom agents on HubSpot that suit the unique needs of their business. For example, the customer built a scoping agent that parses call transcripts in real time and creates and sends a ticket to the solutions architect to prepare a quote. This is a process that used to take days, and deals would often go cold as a result. Now, it happens in minutes. Cyber Advisors also built an upsell agent that automatically analyzes customer accounts based on things like industry and company size, and then it recommends services that they are most likely to need to the sales reps, allowing the sales reps to have much more targeted conversations. Together, HubSpot agents and custom agents are transforming Cyber Advisors' go-to-market, and they are mapping out plans to build additional agents on HubSpot's platform. The third opportunity we see is with APIs. This is all about extending the ways that customers and partners work with HubSpot's platform. As Duncan shared, this includes accessing our platform through data, context, and work APIs. We are confident in our ability to monetize the value we provide to customers no matter what work surface partners and customers choose to operate. I will end where I began. As we deliver AI value with HubSpot agents, custom agents, and APIs, we expect credit consumption to scale as we monetize the work and the outcomes we are delivering for our customers. As we see traction with these emerging growth drivers, we want to make it as easy as possible to buy and scale with HubSpot. This means lowering the barrier to entry to get started and giving customers more choice and control over their AI spend. We took a big step in this direction in 2024. We simplified pricing for sales and service by lowering seat prices, removing seat minimums, and introducing the Core Seat to give all go-to-market employees access to the platform. These decisions eliminated large upfront commitments and made it easier for customers to adopt and grow with HubSpot. What we expected was more customers, lower initial ASPs, and better net revenue retention, and that is what we saw. The change drove approximately 2 points of benefit in net revenue retention in 2025, and we are seeing a similar benefit in 2026. The decision to simplify seats was the right one, and it gives us confidence that further simplifying our pricing to drive AI adoption is the right thing to do for our customers and also for our business. Let us talk about where we are going. We want to make it easier to consume functionality across all of our hubs by giving customers predictability in their spend and flexibility in how they purchase HubSpot. We have made AI pricing more predictable by aligning agent pricing to outcomes, by making credit usage visible across our platform, and by letting customers set limits so they are not surprised by runaway AI costs. Customers have more flexibility than ever in how they choose to purchase HubSpot. Customers can pay for credits by buying capacity packs or simply by paying as they go. As Yamini shared, we have been piloting a simpler seats and credits-based model with new customers in the Nordics and Benelux. Our early learnings give us confidence that we are on the right track, and we are expanding our pilot to new customers in EMEA starting next month. We will continue to test, learn, and evolve pricing in a thoughtful way. Over time, we expect that seats and credits will be the primary drivers of our revenue. Let's recap. Our core growth drivers of up-market expansion, multi-hub adoption, and down-market volume remain large opportunities. Our strategy drives new growth levers of credits and seats as we align our business model to delivering outcomes for customers. We will continue to remove barriers to AI adoption and simplify pricing. Let's shift gears and talk about the meaningful margin expansion opportunity ahead. Yamini set the table well for me. As we transform how we build, grow, and operate, we expect to deliver durable margin expansion and drive outsized earnings growth. Let's dive more deeply into how we think about delivering leverage, starting with COGS and gross margin. We plan to invest aggressively to drive agent adoption and usage across our platform. As a result, we expect that COGS growth will outpace revenue growth in 2027. That said, Aviator gives us the flexibility to leverage models from different vendors and route tasks based on which model will deliver the best quality at the best price. This gives us a structural advantage. As model capabilities grow and prices change, we will continue to optimize. Turning to OpEx, we're transforming how we operate internally, and we see meaningful opportunities to deliver leverage across the organization. In R&D, we'll continue to invest to drive innovation, but our best-in-class internal agentic platform lets us do this more efficiently and deliver leverage. In sales and marketing, we use AI today to handle the majority of our support tickets, to create and localize marketing content, and to drive growth in AEO visibility. We're going to build on this momentum and continue transforming our teams with AI. We're focused on standardizing usage and driving consistent adoption of AI-enabled processes. In G&A, we'll see modest leverage by continuing to lean in to automation, systems, and data. We are confident in our ability to balance gross margin pressure with OpEx leverage and drive continued margin expansion in the years ahead. Looking ahead, our 2026 guidance for non-GAAP operating margin is 21%, representing a 2-point increase year-over-year. In 2027, we expect non-GAAP operating margin of 23%-24%, representing 2- 3 points of incremental margin expansion, and we are raising our long-term target for operating profitability. We expect to deliver non-GAAP operating margins of 30%, GAAP operating margins of 20%, and stock-based compensation as a percentage of revenue of 10% by 2030. Before wrapping up, I want to reinforce our longstanding principles of balancing growth and profitability and share more about how we intend to manage the business moving forward. First, we will solve for long-term revenue growth. We have a deep conviction in our strategy, and we're making the strategic decisions to set ourselves up for durable long-term revenue growth. Second, deliver operating margin. We expect to continue expanding GAAP and non-GAAP operating margins. Third, we will deploy capital strategically using free cash flow for organic and inorganic opportunities that accelerate our AI strategy and for opportunistic share repurchases. As of Q2 2026, we repurchased more than $1 billion in shares over the prior 12 months. Fourth, manage dilution. We'll manage dilution through a disciplined stock-based compensation strategy. If we take all of these together, we believe that the business can compound earnings growth at a very healthy rate into the future. Let me wrap with three key takeaways from today. First, the solid foundation we built as we democratized inbound CRM and built a durable platform company puts us in a position of strength to democratize AI. Second, our strategy creates emerging growth drivers as we align our business model to delivering outcomes for customers. Finally, we are unlocking durable margin expansion and earnings growth as we transform how we build, grow, and operate. Thank you for your time today and for your continued support of HubSpot. With that, we'll move into our Q&A session. Yeah. All right. Great, so we are going to move into the Q&A session now. I am sure there is going to be lots of questions, so if you have one, please raise your hand. We have a couple of mic runners that are going to be running around, so you can just stay seated. We would just ask that you please introduce yourself before you start your question so we know who you are, and try to stick to one question so we can get to as many folks as we possibly can get to. I will give it just a moment here for hands to go up, and we will get going here. Yeah, let us start with Jackson. Thanks. Whoa. Hey, guys. Jackson Ader at KeyBanc Capital Markets. Thank you so much for doing this. I guess we will start with Yamini or Dharmesh, really. The big question right now that I think I have and investors have are around the headless strategy. Yeah. Can you just explain it simply as why is headless good for an application software company? I feel like I have spent my entire life thinking that the higher up the stack you are, the closer to the user, the more value accrues to that company rather than being down the stack. Why is that different in 2026? Okay. I will start, and then Duncan as well as Dharmesh can add if they have. Take a huge step back. I think software was always meant for users, and there is a way in which users got value, and that continues. But what is different with AI is that agents can operate on top of software and drive value as well. There is a big paradigm shift that is happening. The way we think about headless, and I think Duncan did a great job of laying out what our headless strategy is, there are three ways in which now people can interact with HubSpot. The first is by having Breeze Assistant as the key way in which you get the value out of HubSpot. Think about that as users not having to learn how to navigate, not having to learn how to move screens, and really having a conversation to get the most value from an application like HubSpot. That is a new interaction mechanism, and that is what is growing, and we are seeing that adoption within HubSpot. The second way is through connectors. Everybody is talking about connectors. We were one of the first ones to actually build connectors with all four of the frontier model companies, and we have been having that within our ecosystem for more than 15 months. That is a new interaction way. Just like way back when the internet came, there was a browser, and applications became available on the browser. Now there is a new interface in which customers can actually go and ask questions about the software applications that they are using, and that is what connectors enable. Our connectors, we have seen nearly 350,000 users on connectors. When they leverage it, what they are doing is they are going to an LLM of their choice that they are working on and getting insights, getting reports, maybe even creating artifacts to really get more value. We have seen engagement go up. The cohort of customers that are using those connectors, we have seen 12-point improvement in retention. More users are using HubSpot. They are engaged more deeply with HubSpot, and because of that, their retention is going higher. That is the second mode of interacting with software that is completely new with the age of AI. The third is actually an agent talking to software and getting more value. The first two are users interacting with software to get value. The third is really an agent that runs. We are now making it easy to get to the data layer, but more importantly, from our perspective, to the context and the work layer, which is what Duncan laid out in terms of the strategy. Why is that valuable? Because the number of ways in which you can build with AI, you can automate more things with AI than you ever did before with just the software applications before. It is expanded, which means agents get to build, and that means our data, our context, and the work that we deliver also continues to expand. That is how we think about the strategy. Look, lots changing, and we feel really confident in terms of where we are going with our strategy. I do not know either of you if you want to add. The only thing I'd add on top of it is if you think about that three-layered API strategy, when you're interacting with HubSpot at the context and work layer, you're getting all the benefit of us doing our fundamental job, which is our job is to go out there and deliver the highest quality, best value, fastest go-to-market outcomes. As part of all of that, you get everything running on the Aviator Agent harness, leveraging Growth Context. At the end of the day, that's always our number one job. The second thing is we want to be able to meet customers where they are. If they're interacting with us via an LLM like Claude and ChatGPT, we want them to be able to do that really easily. Okay, great. Let's maybe go down here to, Elizabeth, go ahead. Awesome. Thank you so much. Elizabeth Porter from Morgan Stanley. Thanks for today. You highlighted the flexibility of being able to choose between seats and credit consumption. Just as that credit consumption starts to expand, how much of the opportunity should we think about it truly being incremental to the wallet versus just a shift from what would normally be priced into a seat or a hub? Just talk to some of the confidence that you're seeing that credit expansion really drives average revenue per customer higher than simply shifting the composition of spend. Yeah. Thanks a lot, Elizabeth. I'll start and feel free to chime in as well. Look, the easiest way in terms of the framework that we think about is that if AI and our platform adds value to someone within a go-to-market team, it belongs to seats monetization. You'll see us add more and more value through that, which is Breeze Assistant, where people charge for Copilots. We just include that within our Core Seat, as well as Notetaker, as well as a lot of the functionality. If we help go-to-market team members get much more effective, that belongs to the seat. If we deliver outcomes or value that scale beyond people that are using our software, then that belongs to credits. That's the simple way in which we frame these two. Now, over a period of time, we do think that the agentic use cases are going to continue to expand. Today, we laid out our framework for how we think about it. HubSpot agents, our domain-specific agents, will continue to expand both the number of these agents as well as the depth of outcomes that they deliver. As we do that, we will see it in our framework of reach depth as well as the quality. Custom agents, pretty early, but we are seeing really good publish rates in terms of how many customers are deploying custom agents as well as custom workflows. They consume credits. Then we laid out the structure for API. The question you are getting to is: How can you be sure that these are all additive? Well, we are very early in the stages, but the examples that both Jon as well as Kate gave today show the cases where it is additive. It is additive because on the Customer Agent side, we are adding to how our customers are resolving tickets. It is much bigger than the opportunity that we had before. Prospecting agent, we are continuing to see how we can drive even more outcomes by delivering qualified leads. That is additive. At least so far, we have continued to see. I think in the fullness of time, even if the seats are capped based on the headcount, we will continue to deliver more agentic use cases. As those agentic use cases expand, we think there is a huge opportunity for us. It is early stages, but lots and lots of opportunities. Yeah, I guess just to maybe restate and reinforce what Yamini had to say, with small and medium-sized businesses, there are many, many cases where they just are not getting to all the stuff that they can and should be doing to grow their business. You again heard that from me and from Jon, and that the new value that we can deliver through agents or even just our AI-enabled customer platform allows them to do more work, and we are delivering more value than we ever have before. Great. All right. Let us go over to Matt, maybe here in the Joe, right? Yeah. All right. Great. Thank you. Matt VanVliet from Cantor. As we see more companies approach this sort of headless architecture, it seems great to unlock a lot of the value that's sort of trapped in the data model today. How is more of the work maybe gets done on an interface that isn't native to HubSpot, do you ensure that the data gravity remains around the Smart CRM, and you ensure that data continues to be written back into instead of over time, they start pulling in data sources from other areas, whether it's cloud data warehouses or even other systems, and it sort of eventually diminishes the value of HubSpot? How does that multi-hub approach and getting that into more customers' hands sort of fortify that? Sure. Look, the way we think about it is this is why Aviator and the Agentic Harness is so important to be built directly into the Smart CRM. What we're doing is focusing on that goal I was talking about of delivering those highest quality, best value, fastest go-to-market outcomes, and we're doing that by being able to always have that shared context that's available for all of our agents, our own agents, or custom agents to be able to use on top of that. Then exposing that and making the ability to drive those outcomes available via things like the context and work APIs, even if that's headless and inside of another product like a Claude or a ChatGPT. So that goal of us being able to always have that mix of structured and unstructured data, computing, pre-computing insights out of that in the context layer, and then using that to inject into our own work layer will mean that you'll always be able to get those best outcomes directly through HubSpot, regardless of how you're interacting with it. We think that's going to continue to be a huge opportunity for customers, even in headless. Yeah. One thing I'll add just on the headless, since it's a point of discussion, is that Headless exposes the data work API through that mechanism. The other thing we're going to start to see, and this is evolving already, is we will see interfaces coming across the wire as well. We'll have a hybrid conversational interface. I don't think the future is everyone's just going to be chatting with their business software, and that's the only way to interact with it because the affordances aren't there. It's not that user-friendly. We'll start to see is, okay, yes, we're going to pass some data across the wire. We're going to get some work insights across the wire, but we're also going to have UI elements that are exposed from HubSpot because we've been doing this for a while, this is whether within Breeze Assistant or within a third-party AI application, it's going to be a combination of data and UI that's embedded in that conversation. It's not just data traveling across the wire. Final thing would be within Breeze Assistant, which we're seeing increasing usage of, yes, it has access to all of our context, but it also has access to the user context that says, we know what part of HubSpot you're in, we know what you're doing, and we can actually bring the power of both the frontier model and HubSpot's domain expertise inside Breeze Assistant. I think of the AI apps more as just like a browser. This is like a container for functionality. It's the new way to embed software and Breeze Assistant is our embedded way. It has access to the same frontier models. It can do model routing, it can do and now UI embedding. Over the fullness of time, my sense is that, yes, for ad hoc queries, you're going to go headless and maybe you'll sit within Claude or ChatGPT, but there's increasing gravity just to Breeze Assistant simply by where it sits in the stack that has access to everything. It's going to be more economical and efficient too. Great. Yeah, go ahead. Let's go back. Yep. Thanks. Brian Peterson from Raymond James. As you guys hit phase III of this growth trajectory, I would love to understand how you are thinking about the net new customer opportunity versus expansion with existing customers. Kate, I know you are not guiding to that metric and you historically have not, but if we look at what you outlined, it would seem to be that the NRR should be increasing. Is that the right way to think about it over the long term? Durably, what should that NRR look like? Yeah, I think it is a really good question. I do think that there are a number of opportunities that you should have heard throughout today to really drive NRR over the next couple of years. Certainly, seats and credits are two big opportunities that will fuel net revenue retention over time. I think in the near term, we have been very clear about where net revenue retention goes, which is largely flat based on some of the dynamics we shared in Q2 earnings. But over time, if we can execute this strategy well, there are a lot of tailwinds to net revenue retention. Yeah, maybe let us come down here and grab Tyler. Thank you. Tyler Radke from Citi. Great to be here. A lot of great vision on AI, the things that are going to drive long-term revenue growth. At the same time, you have talked about the evolution of pricing, more promo discounting at down-market, changes to how you are monetizing and charging for the AI credits. Can you just talk about the puts and takes on revenue? Does any of this have an impact in how we should think about the shape of revenue growth or ARR growth near term? Then I think you referenced some early good traction in the Nordics, for instance, that you are rolling out to Europe. Maybe just talk a little bit more specifically about what you saw there and maybe is that going to extend globally too? Do you want to take the first part? Go for it. Okay. I think what you should have heard from both Yamini and I is we are experimenting broadly in pricing and packaging. This is certainly the new part of the transition to AI, and what we want to ensure is that we are removing as much friction to our customers adopting and really growing their use of our AI tools over time. One of the ways that we are doing that is through the seats and credits, simplified pricing in the Nordics and Benelux. I think what we like about that is it in many ways is a parallel to what we did in 2024 for Service and Sales, where you are lowering the barrier to get started by really just making it super simple to buy a couple of seats and some credits and really get going and then just grow from there. What we have, it's obviously very early, but what we've seen in those early cohorts is a lot of those same mechanics, which is maybe lower initial ASPs, but larger growth trajectory from there. It gives us the confidence to continue to learn and expand. We're moving into EMEA for new customers next month. Yeah, I think I completely agree with what Kate said. If we take a step back, there is a big shift that is happening within pricing. Customers, I would say, are in the spectrum of foggy to confused to worried about what the pricing changes are, because suddenly they are not just buying subscription seats. Is it token? Is it credits? Is it usage? How is it tied to and how can I control it? I think as we think about how we help our customers navigate not only a platform shift, but also pricing associated with that platform shift, our focus is looking at what are the barriers, what feedback points that we get from our customers and just make it easy. I think that is what you will see us in terms of the different ways we are driving these pricing pilots and experiments. One of the other things that we think about is that everything that is not tied to a go-to-market team member should just be based on credits. Today we might have contact yours in one way, messaging app in another way, and a different mechanism another way. We just want to simplify it, and that is part of what we are trying out. Look, we always focus on One, making sure that we are delivering clear value. When we do that, we want to see repeat usage as well as the depth of usage. When we do that has to drive the right outcomes. If we do all of those, then it drives the monetization and growth. This is why, with all of you, we have shared how we think about the strategy working, which is reach, depth, quality, and then that basically drives growth, and you should expect us to continue to drive that. Yeah, let us go over there. Thanks for having us. Billy Fitzsimmons from Piper Sandler. I have both a financial and technology question, so it is going to be applicable for a few of you on stage. I want to hone in on the sight line into the longer-term margin targets because one of the things we talked about today was that COGS will outpace revenue growth, and over time, that model routing and the optimizations could potentially be a little bit of an offset. Can we talk about where we are today for some of the first-party products in terms of usage of frontier models versus open weight models and what that could potentially look like over time? One of the things some other application software companies have talked about is the ability to build first-party models, which could be more efficient on cost. Is that something you are thinking about? Lastly, on the OpEx lines, can we just talk about how your own use of AI internally has changed here to date? Right. I think this is Duncan first and then Kate. Sure. I would say firstly, we certainly see Aviator and building our own agentic harness as a strategic differentiator for us. Primarily, number 1, as I was talking about, to make sure we are driving those highest quality, best value, fastest go-to-market outcomes for customers. Certainly, as part of that, and that Switzerland of AI approach, that gives us a lot of control and flexibility about how we manage this. We have built it in such a way that we can do this very much on a task-by-task basis. So there are hundreds of tasks across the go-to-market journey, and building out this rich set of evals that we have, we have thousands of these evals that we have built out and continuously run, and that lets us evaluate on an ongoing basis a wide spectrum of model families, model sizes, and so forth. We do that across the full spectrum. We work very closely with the Frontier labs. As you can imagine, we have looked at open-weight models and so forth as well. It is changing weekly, having that flexibility, that dynamism, I think, is a real advantage for us, at the end of the day, that lets us deliver the best results for customers. I would say on the notion of training models and so forth as well, again, we have a pretty sophisticated organization in terms of infrastructure and how we look at these things. You can imagine that we are looking at and evaluating everything across the board again on an ongoing basis. Yeah, I think there is a ton of good news in what Duncan's team has built in terms of the sophistication and flexibility in that core platform technology. In terms of where we are on the maturity side, this is a huge opportunity, right? I think we have always approached things as first deliver the best quality then optimize. As I think about where we are in that cycle, I think we are still very early. It is sort of quality, number one, then optimize cost over time. In terms of where we are as an over— I will sum that up in terms of where we are on the gross margin side. More broadly around where is the organizational maturity in adopting AI, I would say we were very early on in really enabling our employees to get access, experiment, try AI. I think each of the functions has— Broadly, there is significant usage of AI across the organization. I think the functions are at different states of maturity in terms of the shift from the individual productivity gains that we have seen from just exposing and letting people experiment to driving real organizational efficiency that will deliver leverage. Actually, Duncan's team is probably the furthest along, he can comment a bit more on that. The internal platform that the team has used has allowed a real shift in how we build that is broadly impactful across the product and engineering organization. I would say they are leading the way in terms of driving real innovation and efficiency in how we operate. I think there are parts of Jon's organization that are really far along, the support organization is one that is a huge call-out for me, where we have been deflecting tickets for, I do not know, 3, 4 years. Yeah, 3 years, 3 now. We have not added a net new support rep. I think there is more opportunity there to scale up the usage more universally around the sales and marketing organization. I feel like there is a ton of opportunity for us to do that. Then, as somebody who lives in the, we call ourselves the company pillar and the G&A functions, I think, again, we have made nice progress, but there is lots more work we can do around systems and data to get us even more leverage on the G&A side. Okay. Yeah, go ahead. Josh Reilly from Needham. Thanks for the time today. As you think about the macro headwinds in 2026, including the extended sales cycles, how do you think about that in terms of customers figuring out what to do in their own businesses and their strategy versus actual economic uncertainty? If there is greater visibility in 2027 on AI workflows, does that improve your operating environment? Yeah. Thanks a lot for the question. Just to step back, Q1 was a solid quarter for us, but as we came into the year, a couple of things that we saw. One is that customers are navigating a platform shift. We take it as everybody has been talking about AI, but when it comes to making a decision for a 1,000-person company, they want to make sure that it works in their environment, they have the data foundation ready, and they want to make sure that they are going to see the benefits of spending for AI. It is a big platform shift. As they do that, there is a little bit of hesitancy in terms of making the right choice, as well as making sure that the budget works within that. As we mentioned in the earnings call in Q2, we said that for upmarket customers, it actually shows up as larger buying committees. If a CMO was making a decision before, now it requires a CFO or a CEO approval, or sometimes even a board and PE approval. I think we will see that continue. This week, Jon, myself, we are all talking to literally tens, maybe even hundreds of customers within UNBOUND, and they are navigating that shift. I do not think this is a one quarter kind of an effect. We are driving that by engaging earlier, by really focusing on the outcomes that we can deliver to customers, by making sure that our partners can deliver with the right implementation early on. We are changing our playbook to be able to drive that, and we will see some level of pricing sensitivity. To take a step back, we understand that customers are navigating a big platform shift. We understand that there is pricing scrutiny as well as hesitancy with the new model. Our job is to make sure that our playbooks and how we execute is really mapped to that. You asked about 2027. It is pretty early to say what happens in 2027. But as the outcomes become clearer, as the customer stories that we share with other customers become clearer, and people get much more comfort, over a period of time, we will see that type of navigational hesitancy of a new platform dissipate. Our job is to make sure that we are giving that level of confidence in our strategy as well as comfort in the pricing model that we are rolling out. Anything you would add? You nailed it. All right. Thank you. Okay. Yeah, let's go back there. Thank you. Brett Huff from Stephens. Appreciate the time today. I had a question on kind of following up on the data gravity question and the multiple ways you're meeting folks where they are in their AI journey. Some are just existing in HubSpot, and that's great. Some are extracting data in some meaningful way. Can you talk a little bit about how that data's getting back into HubSpot? I'm a big believer that the system of record will still be there for a long time, but what kind of behaviors do you see when people do extract it, and how do they put it back in? Maybe it kind of links in with this idea of self-updating CRM. Are you trying to kind of head that off at the pass? Thank you. Yeah, I think firstly, as Yamini talked about earlier, what we actually see greater engagement and retention from customers that are using our connectors to access the platform. That means they're getting a lot of value out of HubSpot, and part of that is continuing to contribute data back into HubSpot because ultimately, it's Growth Context and the Growth Context graph that drives all of this work. It is in customers' interest to have that shared context, be the most relevant, be most up-to-date, so that everything across the platform is leveraging that as part of it. So when they're using our own first-party agents, they're using other capabilities that are AI-powered within the platform. They're building custom agents in Agent Builder. They are accessing, in some cases, those agents, within something like Claude or ChatGPT, or even if they are building something custom within Claude or ChatGPT to help them orchestrate something. So we see data getting written back to the Growth Context engine, as well as being used within those platforms. Hey, thank you. Nick Altmann from BTIG. Kind of a philosophical question here, but just over the medium term, where do you expect the bulk of agentic engagement to be within HubSpot? Is it more on the prepackaged agents that you guys provide, or more on the custom agent side? Maybe just talk about the implications of both. I imagine if you are a partner, maybe you have more incentives to work with customers on custom agents, because that seems like a lucrative opportunity. At the same time, maybe your customer base isn't as used to buying or adopting custom software. So maybe just talk about philosophically how you expect that to trend over the medium term, custom agents versus Yeah your prepackaged agents. I will start because we have literally been having this conversation between HubSpot agents and custom agents, where would we see the mix over the long term? Look, we have domain expertise in go-to-market, so we are at the cutting edge of what's happening in AEO, which means we can bring that to an agent. We understand what's happening with content, so we bring it into the Content agent. So I think over a period of time, we are going to bring our domain expertise within the segment that we serve for the go-to-market functions, and that is going to be the HubSpot agents. We will see that adoption continue. The interesting part about custom agents is that it is workflow automation 2.0, right? If you go back in the fullness of time and look at marketing automation and sales automation, yes, they used our features, but what did they do with our platform? They actually built custom automation and workflows on top of HubSpot, whether it was Marketing Hub or Sales Hub. Now there's an even easier way to build and extend. So custom agents are, you can build new workflows, you can build custom agents on top of it, you can drive artifacts. So you can do a lot more with AI, so it's the second part of it. You're absolutely right. One of the things that we're doing is with the partner ecosystem. They are going to love it because this is how they extend HubSpot into industry specific use cases, and we can already begin to see how partners are engaged. In fact, today we have a battle of the builders. We have a USD 100K competition for partners to actually build custom agents, share those use cases, and really drive that. So we're incentivizing our partner ecosystem to do it. I don't know exactly how this is going to land, whether it is going to be more custom agents or not. I think both are really critical in delivering the kinds of results and outcomes that we want to within marketing, sales, and service. But we've been having this debate. What is your take? Yeah, absolutely. I think, look, if you look at it overall, we're going to continue to build out the set of HubSpot agents that can deliver across core go-to-market outcomes. We think that's incredibly valuable for customers to be able to just take those out of the box and get value quickly, whether it's Prospecting agent or Customer Agent or any of these other capabilities like Campaign agent that we've announced this week at the show. Agent Builder then is giving this incredibly easy-to-use tool that customers and partners can build on, and it can be the glue that connects a lot of these other capabilities together. It could be extending a HubSpot agent, building something brand new that's very unique to their business need, and everything running on that same harness with access to the same Growth Context, with all the same governance, trust and safety, quality guardrails around it, just automatically applied as part of that. You just talk to Breeze in order to tell it what you're trying to do, and it'll build the agent for you. So it's already been very exciting to see what customers have been building with Agent Builder, and it'll be really exciting to see where it goes going forward. Ultimately, it feels to me like we're getting to this very exciting place where, for a long time, companies had to adapt to the way the software worked, and now the software is adapting to the way their businesses uniquely work. Putting that power in a very easy-to-use way into customers' hands, with the help of partners often, is very exciting. All right. We've got time for one last question here. He's had his hand up. Yeah. I don't want to miss you, so let's go here then. All right, thanks. Martin Schneider with Constellation. You guys talked just now a lot about the solution providers, what you are looking at, the VARs and resellers, and great opportunity there. But when you think about where headless and where context being king and then that auto entry, how do you see your ISV relationships needing to change and that focus where Because again, features and those little adjacencies, the agents can do that now. But how does that context, where does that come from? Because a lot of those surfaces, a lot of those places where people are going to want to work in a headless environment, they are your competitors. Those entities are owned by people who you would see as competitors in CRM and in other areas. How are you going to have to play nice a little more, or force that, and how is that working out when, just to put it bluntly, small businesses are like, "Yeah, we use Slack, and we want to bring HubSpot into Slack and vice versa, and create those surfaces and those artifacts and those cool things to both lower our token utilization of all of our AI and all that." How do you see that evolving, that whole ecosystem where a lot of these are your competitors today? I think if you come back to what I was talking about earlier, we need to meet customers where they are. We are going to make sure we are delivering the best go-to-market outcomes for them and then engage with them where they want to engage with us. We already operate inside of Slack today. You can use Breeze Assistant from inside of Slack. Obviously, Breeze Assistant is there for customers that want to directly use that. We think that is the best way for most customers who are working in the go-to-market space, but we will continue to evolve. We have talked a lot this year about our open ecosystem, and we really think that is the road forward. So much is changing so fast. Customers are trying different things, but ultimately, we will take that set of structured and unstructured data that we have had across the whole go-to-market. We will use that to continue driving the best growth context out of that and be able to inject that into the agents and tasks that we are doing for them and then meet them where they are. It is going to be really interesting to see how customers evolve and how they want to interact over time. Got you. For Slack Bot, for example, they would not want to use that because A, it is probably too expensive, and they do not have the context of the data model because that Salesforce, they do not have that, right? Because they are just using Slack. So it actually, they have, you have Exactly. In that example, they are using Slack as effectively as an interface to access HubSpot and our entire platform via that. But we are able to meet them there in Slack where they are. Yeah. I would say, if we really step back, the paradigm for how customers interact with software is changing. Even within HubSpot, in a year, even now, we would want them to just live and breathe in our Breeze Assistant. That, we are making it world-class. We just want to make that experience just exceptional, and if you look at how our customer satisfaction on just that usage has grown in the last 12 months, it has just been extraordinary to see how customers are adopting. If you are a serious go-to-market user, you are building campaigns that are multi-language, multi-region, you have 10 million contacts, you are not going to sit in some other surface. You are actually going to be in the software, interact with the power of the software. We will make that world-class, and we will continue to build on that. The second part of it is that, yes, now there are new ways, whether it is Slack Bot or another LLM, that you can conversationally have interactions with the software, and we are going to make that great as well. We have all the connectors, and we can begin to see how that improves the engagement and retention that you get out of that. I think the third case is, like other software that is using CRM, it is still valuable too. Over a period of time, I think the industry is still early in the cost optimization phase and performance optimization phase. It will be much better to leverage our context API and work API than pull the entire data and do inference outside and try to figure out which model to use outside. I think there's value for us in every one of those situations, and that's why we have confidence in the open ecosystem strategy that we're building. Okay. Becoming more collaborative in grouping can actually help people reduce their spend on other products like Slack. Right That can cover consumption, right? Right As well. Okay, great. All right. We are going to conclude Q&A. For those of you that are live with us here in Boston, we are going to move out these doors now. We have got a few minutes here. We will grab a quick bite to eat, and we can connect for a few moments. I just want to wrap by saying thank you to everybody for joining us here today. We really appreciate you taking the time to be here and for your continued support of HubSpot. Thank you. Thank you, everybody.