Atlassian Corporation ($TEAM)

Good afternoon, everyone, and welcome to Atlassian's Investor Forum that we're holding here at Team '26 in Anaheim. It's great to see so many familiar faces in the room, and thank you for coming out. And for those of us joining via webcast, thank you for tuning in. But because so many of you were coming out, making the effort to learn more about our platform, learn more about our solutions and hear from our customers, we wanted to use this chance to help you better understand our strategic vision, our opportunities that we're attacking and the momentum we're seeing across our business and our 3 strategic priorities. And so today, we want to talk to you about how we're serving enterprise customers, delivering that unique AI value and what our System of Work unlocks. Today, Mike, Brian and James will share more about our longer-term vision and strategy. We're coming off a strong quarter, but today, we're talking to you about why we believe Atlassian is uniquely positioned in the AI era and why the market is still misunderstanding or underestimating our runway and growth opportunities. And so you'll hear from Mike about Atlassian's evolution into a platform company, our System of Work that we've built that serves all teams and powers hundreds of millions of workflows each and every month. He'll walk through our Teamwork Graph and the powerful context that delivers, our AI strategy and the massive runway ahead across our 350,000 customer base. Then Brian will speak to you about our enterprise go-to-market evolution and how he's taken our platform and the increasing value it delivers and how it translates into bigger, stickier, faster growing customers and relationships. And then you'll hear directly from two customers, Cisco and [ Canal+ ], on what that looks like in practice. And then finally, you'll get the chance to meet and hear from James, our new CFO, who will share his early perspectives. Please note that we won't be providing long-term financial targets today. And then we'll close with some Q&A. Before we dive in, the standard legal housekeeping around forward-looking statements, which you can see on this beautiful slide behind me. Please take a moment to review. All right. With that, I'll hand it over to Mike.

Thank you, Martin. Good morning, everyone. Good afternoon, sorry. It's morning, where I'm from. That's my excuse. Good afternoon. Thank you for being here. I hope you all had a good morning this morning. Did you all attend the keynote? Just so I know what I have to replicate. Okay. Well, this will be way more fun, don't worry. And thank you. It's good to see so many familiar faces. I should say that. As Martin mentioned, we're coming off a very strong quarter. We -- total revenue, $1.8 billion, 32% year-on-year, cloud at [ $1.1 billion ], 29% year-on-year, RPO passed $4 billion at 37% year-on-year. I'm sure you all know those numbers as well as I do, but a really strong quarter, another chapter in our story. It is an example, I would say, at the highest level of our 3 strategic pillars, the Enterprise, the AI and the System of Work all coming together. And I hope over the 3 or 4 presentations today, you see how those continue to thread together and lead to that strength and growth as we have shown continually. But as Martin said, it's not about that quarter. I want to -- we want to spend time talking a bit more about the longer-term story and what that means because -- underestimated, under -- misunderstood, all these things that you said, I think it is totally true. There is a big disconnect, I believe, in how we are seeing and what we are actually delivering. And so I thought we'd go through some of those doubts and give you a lot of facts and figures and other things to go through those. So you've probably familiarly heard a bunch of the, hey, I is going to make devs redundant. We only serve devs, seat growth is going to die, seats are going to disappear, AI agents are going to make everything less critical. People can now make their own SaaS by themselves, make their own operating system. Don't probably make everything, AI-native start-ups are going to just kill everybody. We don't have the right monetization model. I'm sure you've heard some version of all of these things. Let's go through some of them. Fundamental problem, I would say that's very much looking backwards, right? That might be looking at it last 10 years ago and assuming we're a static company. We've never been a static company. We've probably always been underestimated and misunderstood. Awesome. Bring it on. The platform we built, transmissions we're delivering, I hope we're showing you every quarter that we are delivering the numbers that we say we're going to deliver durable growth that you can see in those numbers. I do think it misunderstands our runway going forward, and we'll try to give you some indication on that. Let me tell you what Atlassian is about, grounded in real numbers. So I know a lot of you, despite being extremely small, struggle to model Atlassian. You see a bunch of apps, and you usually talk a lot about Jira and Confluence, Although I still don't know how many Confluence questions we've had in any earnings call. I think I said this a couple of years in a row, like it's still amazing to me. So we'll get to that in a second. But the Jira and Confluence is not how Atlassian works. It's really not how we work, right? We are building a singular platform. As you now see, we have all 5 collections. And plenty of companies claim to have a platform. Most of them, I would argue, as a technical person, do not have a platform, not like we do, right? We have a single platform that delivers a single System of Work that continues to compound and multiply the more collections and apps that you have on top of it. It delivers a common way to do things to the flexibility for an organization. And the value is in not the data, but the graph, the ontology. We will spend a little bit of time talking about that. And that's why I'd argue we're becoming an operating system for work for a lot of the world's largest companies. Customers are increasingly committing on a platform level. Brian will go through that. We have more and more platform level committed customers. I've talked a bunch even this week that are making even greater commitments to us, which feels excellent. It's a justification of all the things that we've invested in. As they add collections, they deepen their investment in us. It also makes our AI smarter. I hope you see that. We'll have a lot of slides, and I'm sure there will be at least a couple of questions on AI as we get there. Now one of the reasons that our growth is so durable is that it's incredibly diversified. So we told you in our shareholder letter last week, the Service Collection has passed $1 billion ARR and is continuing to grow at 30%. That's ARR, not even revenue. Now that's just Service Collection. Confluence by itself has passed $1.5 billion in ARR. Jira is past $2.5 billion in ARR. So we've built multiple billion dollar-plus businesses, all of which work on a single platform, which is really important for our customers, that serve every industry and every team, from finance to HR to IT to software teams to designers. And that is an incredible set. And people will keep asking, well, how did you and Jira and Confluence keep chugging? They must be slowing down, right? And it's not the case. Today, Jira and Confluence are accelerating as we have said in the cloud. And it's because of the platform, right? We serve more and more workflows and business processes, which I think people greatly underappreciate. Right? With the Teamwork Collection, people aren't buying point products anymore. They're buying a collection of things that work together that are part of the platform. And that's a massive signal for us, as we've shared, of 1,000 customers that was quarter a bit ago, I should say. It was in previous quarter's results and more than 1 million seats in the first 6 months. So those are some really, really strong numbers, but the point is diversified business and multiple large businesses, all of which are growing very strongly. And again, the collections make it easier to expand. The changes we've made from a set of products 3 years ago to the apps and collections that we have today working on a singular cloud platform make it easier for customers to consume our software. We have less things for them to buy with far more value. And part of the reason we're seeing that runway ahead of us for durable growth and the growth we're seeing today. Now you can see all of this diversification playing out in our numbers. Right? Customers are not just buying in, they keep expanding. So let me give you an interesting chart, I think it would be interesting for you. Let's look at total cloud seats, excluding all migrations. No migrations involved. And despite the seat based, it's going to disappear. You can get your rulers if you want. I know you do. You can see visually, it's not slowing down. It's a pretty amazing curve. Again, this is excluding all migrations, cloud seats. They continue to grow every quarter, every year. Durable growth, as we talk about. Why? Because workflows continue to get more complex, more intertwined. Teams and companies are buying more into the Atlassian platform and getting exponential value as they increase their commitments. And that's really important, and we think that only keeps accelerating from here. This team started orchestrating AI agents alongside their people. You heard me talk a little bit this [ mongo ] graph. I think we have more ahead of us that we can continue to do because of our investments in that platform. [ Ryan ]? Now you might be asking, okay, they're adding those seats, but seats for whom, right? This is where the old narrative, I think, breaks down. People pigeonhole Atlassian as a [ fellow ] company. I might argue there's going to be more developers in the AI era. I think that's going to be true, but fine. You think we're the developer company, or some people do. We love that heritage, I will say. Developers are awesome. Technology and business teams working together is the core of what we do. We think developers are incredibly important to the future of every single large business on this planet. Technology is the core strategic advantage of almost every single big company that you will talk to. If not, they have some sort of regulatory barrier, et cetera. But the growth we're seeing hasn't been driven by developers for a long time. The System of Work is for the entire company. It's for finance teams, it's for HR teams, it's for marketing teams. It's for operations teams and service teams. When we talk about assets and buildings and laptops and everything else, we're talking about far more than developers. We're powering hundreds of millions of workflows every month. Now heard us talk before about how half our users are not in technology teams, and we've got a half technology, half non business. We shared that before. So how about this one instead? Break it down, and let's look at developers versus everybody else, knowledge workers. In Confluence, 7 out of 10 users for that $1.5 billion plus business that's growing very well in the cloud. 7 out of 10 of our users are knowledge workers, nondevelopers. In JIRA Service Management, it's over 3/4. Even on Jira, it's roughly 2/3 of the users are nondevelopers. That's HR and finance and legal and marketing and operations staff and design. We spent 2 decades connecting these business teams with their technology teams, and we have a lot more to do. Again, a different way of showing that we are a well-diversified durable growth business. Now let's look at the customer segmentations. So when you're embedded across that many teams inside a company, customers are expanding, and how do we see that? Well, we talk about having 350,000 paying customers. We know that that's the bottom of this kind of pyramid illustration. How about at the top end? So our $1 million customer base, which I believe we said a quarter and a bit ago, had passed 600 customers spending more than $1 million a year. That has grown sixfold in the last 4 years. In the last year alone, it grew at 39% year-on-year, that category. These customers have more than a 99% retention year-on-year, and they are almost all expanding. If we look at our $3 million cohort over the last 4 years, when we talk about the enterprise investments we had on the technology side, on Brian's side, that $3 million cohort has grown tenfold in the last 4 years, 54% year-on-year in the very last year alone. And we have quite a number now of those $10 million-plus ARR customers. These are ARR numbers as well, not revenue numbers. Takeaway I want you to get is we're not a nice to have tool. We are part of how these organizations run. We are a strategic partner. That has been our goal, as we've said in every shareholders' letter for the last 4 years. That is our enterprise goal. We've been working on the enterprise for a decade now since data center launched a decade and a bit ago. And I think that's reflected in the movement of our customer base. So the top end of the pyramid is going very well. The important part is we have an awful lot of customers in that bottom end of the pyramid who are moving up there. So the platform works, customers stay, they expand, they deepen their commitment to Atlassian. Natural question is, but do you have all the businesses? What is the runway? So we've updated all our SAM numbers. The way we've recalculated it now, updated the calculations is that our SAM is about $140 billion. We have a lot of runway is a simple thing to take away from this, see it broken down by different collections and other bits and pieces there. So there's a lot ahead of us. And then the question naturally is, well, how do you go after that? How does that actually happen? So we've broken out these 4 growth levers, which are the ways that we grow typically, right, expanding inside the customers we already have. Those million-dollar customers become $3 million customers, have come $10 million customers that started as $10 customers. That is expanding inside customers we already have. When we move to new groups inside those companies, we don't count them as a new customer, they're the same customer that we've always had. Most of the customers you'll hear from today, and I'll show you some examples in a second, have been with us for 20-plus years. I routinely speak to customers who have been with us for more than 2 decades. None of them are spending less than they were 2 years ago, let alone less than they were 10 years ago, and their commitment to Atlassian keeps growing. We obviously cross-sell across the portfolio. I'll show you some stats about that in a second. But we obviously have far more apps, I think you can see across the moon rise than we have ever had in collections. Upselling into new value tiers as we've added the premium and enterprise tiers in the cloud. And of course, winning brand-new customers is still something that's incredibly important to us. When we look at that 350,000 customers across the globe, they're in literally every industry on the planet. I have spoken to customers in every single one of these industries in the last maybe quarter, at least a year, every single one of these industries. And for every single one of these industries, they're driven by technology. But their technology and business teams connect together is how they're trying to win in the future, and that's exactly where we sit with our platform. And it's a massive number, but we have barely scratched the surface with the opportunity in most of these accounts. How do we know that? Again, we talk about the Fortune 500 that are already Atlassian customers, 85% of them. But again, representing only 10% of our business today, I think we only just passed 10% very recently. We have a huge amount of growth in that Fortune 500 cohort alone. But the headroom inside all the other companies we serve is massive. We are in the European equivalent of the Fortune 500, the Australian equivalent of the Fortune 500, the Asian APAC equivalent of the Fortune 500. So across the global enterprise landscape, we have a huge amount of runway ahead of us. How do we know that in terms of those knowledge workers? Well, if you look at the 1 billion knowledge workers on the planet, about 900 million of them are outside of technology, about 70 million technical knowledge workers and about 30 million of them are in development. Call this order of magnitude, rough numbers, there's various surveys, they all end up about this sort of proportionality. If you think about the fastest-growing segment of our customer base, which is that nondeveloper audience, we have an awful long way to go with our tens of millions of users today to that $1 billion. So a few different ways you can think about our potential. And again, when we land a new department, we don't add a new customer. Just 2, 3 months ago, we landed a 200-person marketing organization that had never touched Jira, in 1 of the 10 largest companies on the planet. What's important is those 10 largest companies, they said they had never used Jira. In fact, one of the people said, "I didn't know this was Jira. I thought this was someone else." I won't say which competitor. And that was awesome. Right? That was amazing. And then when they use it, they're really happy. Why? Because they're connected to their technology team. They almost literally repeated our core mission back to me in a meeting, and it was a truly amazing moment. So, massive opportunity ahead of us. We do have significant cross-sell opportunity just in the Jira base. So how do I keep illustrating this in different ways? Here's a different way to look at it. If you look at the Jira base, in the cloud, 150,000-ish customers, call this a sort of a round number. And again, as I said before, Jira is growing extremely fast in the cloud by itself. So this number keeps going up. However, if you look at Service Collection as an equivalent, it's only 65,000-odd customers. There is a huge runway just if Service Collection did nothing other than sold to the Jira base,right? That's over -- should have done that before my coffee. What is that, 80,000-ish, 85,000 Jira customers where Service Collection is not today? That is a huge number when you look at enterprises around the world. And these are only significant customers, I would say. So the reason is 150,000 have included a lot of the smaller ones and stuff like that. So they're of size, let's say. The best part, those aren't cold conversations. When Brian's team go in and say, "Look, your Teamwork Collection is going really well. Let's talk about the Service Collection," and you've got the same data, the same platform, the same Teamwork Graph. It's not a cold conversation. That is an amazing conversation. Part of the reason why the Service Collection continues to take share and do really well in a lot of those -- increasingly in the enterprise and strategic segments. I talked about upgrading additions. This gives you some interesting statistics on where we were. So if you look back 5 years -- again, this is Q3 numbers, so not quite the end of FY '26 yet, but roughly a 5-year period, call it 4.5 to 5 years. You can see that our mix here has moved from the standard premium enterprise we had 5 years ago to where we are today. It's an example of our enterprise journey continuing. Now I drew this chart for you, so it was illustrative on the left to the right. The right-hand bar, obviously, in quantum is far larger than the left-hand bar. We have far more customers than we had 5 years ago. But just as a proportionality, it shows that our customers are increasingly opting for Premium and Enterprise as we continue to solve scale, security, regulatory compliance and other issues that can opt into. At the same time, the standard continues to grow well because we get that broad base of millions of companies around the world. How big is that broad-base? Well, there are at 6 million companies roughly that we would estimate are in our target market, right, that aren't Atlassian customers today. We have a few hundred thousand free instances in businesses of any significant size. So again, we're not talking about a one-person instance, someone credits a test, et cetera, excluding that. So it sort of shows you where some large amount of opportunity when you look through those different things. So we talked to the System of Work, a growth, lots of opportunities. I thought it would be helpful if I broke it down into some actual customer examples. So this is real customer data anonymized. Neither these 2 customers are coming to talk to you today. They are 2 examples of thousands of customers that I could pull out literally that show a multiyear journey. So I'm going to start with the global auto manufacturer, 100,000-plus employees, a very big company. Let's look at their journey over 15-plus years with Atlassian. It's a bit of an eye chart, so let me walk you through it for a second. If you see the line at the bottom, the first line is their time on server, for those of you who have been with us for a while. The second line is their time on data center. And the third bit of the line is their time on cloud. So they're on cloud for the last 5 years, they spent 3 or 4 years on data center, and what is that, 6, 7 years on servers. So they moved their deployment options over time as this example of a customer. They've been with us for 15, 16 years at this point. And you can see at the very bottom, what applications they've bought. You're always asking, where does all the R&D go? Well, here is a good example of where they've built things that we didn't have and they bought things that we didn't over time. They're now sitting -- oh, and I should say at the top -- sorry, it's probably the most important part that I missed. This is obviously their ARR in millions of dollars at the top. So they were at $1 million to $2 million 5 years ago. They're now sitting, what is that, sort of 8 to 9, maybe 9 and change. Sorry, my eyes aren't that good anymore. But you can see, as they've moved, where those jumps in revenue related to their movements. So you can see when they move to data center, it moved from what was actually an exponential chart for the first 5 years. It is actually exponential for the first 5 years. It's just -- on the left-hand side, they're probably spending $5,000, $10,000, like near -- you can't see the bars, move to data center, they're getting to the $50,000 to $100,000. They get to the $1 million in data center range. And then as they move to cloud, they continue to ramp up. We talk a lot about why that is, but it's fundamentally they have many more things they can buy. They're expansion opportunities. Their expansion speed in the cloud is a lot faster, and they've moved to many more users. Back here, they're in the 5,000, 10,000 user range. At the end now, they have 70,000 users running on cloud enterprise. And the scale of what they're doing is absolutely wild, right? Millions and millions. In fact, more than 10 million automation rules running every quarter, hundreds of thousands of manual tasks disappearing, massive Rovo adoption across the organization, agents running AI, Rovo is their default for AI. They like it more than any of the other AI solutions, et cetera. So you can see when they've started to add strategy collection and Jira product discovery, albeit those are small and still have some growth. So one example of a customer, auto manufacturer. So I got one more. No, I forgot something. Didn't elation Rovo, I already said that. And it's all about the platform. When you talk to this customer, the platform is what has enabled them to keep going. A typical pattern that we'll see with the cloud, '21, '22, '23, they're actually adopting cloud, moving to cloud, moving users over. Once that 2- or 3-year period is done, you start to see that acceleration afterwards as they then take on JPD, Service Collection and onward to the Strategy Collection, which is now running their enterprise dashboard for this large organization. Here is a totally different customer. This is a major financial services firm. Different continent, different part of the world, different industry. Again, a second of many hundreds of customers, I can sure they look exactly the same. They again have been with us for about 15, 16 years. I actually cut this one off. I think they actually have been with us since 2004, for memory. So actually 22 years as a customer, but the first few years are literally dislinked from the scale. Very similar journey to show you this one. Again, I believe they're at 30,000, 35,000 users, something in that realm, so about half the size of the prior customer end users. However, very close in spend, and you see them in the last couple of years, again, once they've moved to cloud, a year or 2 of adopting and moving themselves to cloud, moving their data across after a long journey with us and then starting adopted Product Collection 3 years ago, Rovo 2 years ago. They again have tens of thousands of AI now. They're running with Rovo. They really, really like it. They use it in automations. In everything, it's one of their default things. They built a bunch of agents to do critical work inside compliance, in HR and lots of other places that run natively on the Atlassian platform. They're a big user of the Teamwork Graph CLI. They're in an alpha program for the Teamwork Graph CLI. So they're actually using the [ camel ] interface. They think it will unlock a whole lot more usage of the platform. They have an incredibly strong, responsible AI framework. They spent 6 months putting us through our -- their AI paces, and we're one of the first AI tools that they turned on, which we're really proud of. We love customers that have very strict guardrails and barriers, and then we pass those parts. One of the biggest goals they have is what they call the connective spine of their organization. So Rovo is part of their connective spine, automating reporting, cross-team overhead, thousands of engineers like it's a really core part of what they do. And in the last 12 months, they've adopted DX and the Strategy Collection. So you can see that nice jump as they've come those types of things. It's not entirely due to that, but just like a general movement. It's an amazing example of customer. And it works because of the platform. So two of many hundreds of customer examples, I hope they were illustrative to try to show you what it is that we deliver for our customers. And I think they show you the power of being an R&D-led company. We've always been an R&D-led company. We made a deliberate choice to build a true large-scale platform, a single platform, not taking any shortcuts. Even if that meant taking margins down temporarily as we invested in our platform, as I believe we've tried to explain every single time what it is that we do, right? We have certainly ramped R&D over the last few years in a very deliberate effort to build that multiyear period of time to build that platform, right? What have we done in that period of time? Many shifts simultaneously, taking our largest customers and moving them to the cloud. Building that singular cloud platform, which we've seen the results of today as we saw this morning, delivering all the enterprise reliability, governance compliance scale of those world's largest companies to 50,000-person-plus companies right there, 30,000, 40,000, 80,000 seats, were 70,000 seats. And moving from those point products to those integrated collections, a huge amount of work to do, all of those. What does that deliver for customers? Well, I thought I'd give you an example, one way of looking at what it is we deliver for customers today. So all of this is what we deliver for customers today. So when I say that, we look at our 3, we think about our 3 transformations, right? We think about it 3 pillars. We've got the Enterprise, the AI, the System of Work. Tried to organize the things we do into those sort of put it into the same lens that we explained in our shareholder letter. All of this stuff on the left, isolated cloud, not easy, [ Jira ] customers really want it. Government cloud, we're in Google Cloud, FedRAMP moderate. So multi-cloud is Google for now alongside AWS. All of the things on the left guard security clients we talked about. In the middle, obviously, all the AI capabilities that we've talked about, from Rovo Search and Rovo Chat, which are -- the first customer there, it's like their best AI tool. We give them the best answers in Rovo Chat of any tool that they use. The CLI, all the things you saw this morning, MCP, et cetera, we talked a lot about AI questions. Obviously, all powered by the Teamwork Graph with pulling that out. That's very important. And on the right, those collections, all the third-party apps, all the platform apps that we've built in terms of goals, et cetera, et cetera, et cetera. You see the icons, you get it. And our Forge development platform, I don't know where to put that. It's not technically in any of the 3, so I put it in a System of Work. Now we can support our biggest customers in the cloud after all of that R&D. We've built the world's best context graph at a time and context graph are incredibly important, and it's running at absolutely massive scale. Like we're preparing for trillion object scale. We have to, right? $150 billion plus at the moment, it's nontrivial size and scale of anything. Even on a consumer internet company, that's a big scale. Deliver FedRAMP, all of the things that we've talked about there for banks, government health care around the world. That's what unlocks all the enterprise expansion that I'm sure you'll hear a lot about from Brian. So I put this down for one reason because you're always asking, where does R&D go? So we look at that. Let's have a look at what this looked like 2 years ago. So 2024, it looked like this. So we had apps. We had no collections. Most of the AI stuff wasn't there. Very little of that enterprise stuff have been delivered. Everything on that prior slide, by the way, was GA, shipped out to customers being used today. 2026, 2024. That is a massive change in our delivery of product that those 2 customer examples hopefully are showing you how that is turning up. Hopefully, that was illustrious. We would explain where R&D goes. And our platform investment, R&D investment is not just going into product capability, all of those things you show. It's delivering leverage that I hope you can see showing up in our financials. What does that leverage look like? Well, you probably already know these numbers. But for FY '26, whole year is not finished yet, et cetera, expected non-GAAP gross margins of 88%, right? 3 points year-over-year, by continued optimization and scale, and a lot of R&D investment that goes into creating that scale and optimization of our client infrastructure, which now is running at massive scale. We are hosting more customers. We are hosting larger customers. We are running more workloads. We are powering far more usage through AI. We're doing an absolute boatload of AI traffic, and our margins are going up, not down when it comes to COGS in R&D. And that is an incredible effort, thanks to R&D leaders. I don't know if we have any in the room at the moment. I think we have at least one. And all of the R&D team that have done that over the last few years. That's 2 charts you would expect you in opposite directions. And that's the leverage I would argue you get when you build a real platform, right? Now in the front run some of the questions that you're going to ask. I know a few of you very well at this point. How are you going to moderate that R&D expense? I get it, that all sounds wonderful. What does that mean? Well, quite simply, like we've done a lot of the heavy lifting. So we keep trying to get our language correct here. The unified platform, all of the enterprise capabilities, the unique data graph, that has to be built upfront. That is an upfront spend to get all that. We can't move a single one of those 2 customer examples without all the things that were built in the enterprise. That takes years, 5, 6, 7 years to build properly and to do it at scale and for them to test you and trust you to get on to that. We believe we have done most of that work in the enterprise capability and in the System of Work capability. A lot of the heavy lifting is there. We will keep investing in all of those areas. But in terms of proportionality of revenue, it will moderate and trying to explain the reasons why it does. Means R&D doesn't need to grow at the same pace as revenue going forward. We're obviously going to keep investing in AI. AI will likely to continue to grow. Enterprise will moderate. We're balancing between those two, but moderate as an overall share of revenue. Right? And that's about the leverage we're getting from the R&D investments that we have made. Every new capability we build in the cloud and as we migrate more customers to the cloud. It's on a singular foundation because of the investments we've made, 1 platform. Not lots of different platforms. We don't have lots of different apps we built in M&A. We replatform everything. It's a long-term R&D story that -- I'm going to be here in a long time. It makes a huge amount of technical sense for us, and it gives us that customer advantage as well. Means every incremental R&D dollar we do now benefits the entire cloud platform. And a huge credit to R&D team. I can't explain to you how different this is to the vast majority of other SaaS companies out there, especially anyone with our breadth of product and customer portfolios. So we talked a lot about the platform R&D investment, so people can ask about AI. Look, every SaaS company has an AI story right now. And so we are trying to work out how we explain why ours is different, and we truly believe it is differentiated and different, right? I can tell you, having talked to that customer a couple of days ago, they probably have 7, 8 apps that have a chatbot, chat sidebar, chat thing, chat app, chat something. Majority of them are just utter c***, right? They don't work very well. They don't give you what you need. Why are we winning in that organization? Because our chat is awesome because it's backed by the graph, it's backed by contextual. Massive amount of work in architectural intelligence, smarts and understanding of what it is. We built it properly. We built it really well. We adopt modern models, a whole lot of other things. It is a big investment. I believe it will pay off in time, and it's very unique. That's the way to say about that. Context, you probably heard a lot about it this morning, so I'm not going to repeat any of that. I'm happy to answer any questions on that. I do think in a world of Agentic capability, agents flowing around, personal agents, team agents, et cetera, our context is going to become really important. I have a little video coming up, which I think will be hopefully interesting. But we do think that is something that is both durable and expanding to our TAM. If I had to simplify from an investment point of view, it's really tricky to work out how to do this. We have tons of AI advantages. I try to get down to 3. Enterprise is really hard. One of the reasons is that platform we've built, Rovo, Teamwork Graph runs on top of that platform. So for those customers, we can operationalize that at quite some scale. When we have tens of thousands of AI users in some of these companies, that's a nontrivial thing to do, and most other vendors actually can't deliver that today. And so we've got that. We continue to add more and more permissions and compliance and controls alongside our customers as they're learning about AI alongside us. We're really on the cutting edge of doing that, and it's very nontrivial and backed by the Teamwork Graph and the platform that we've built. Permissions and AI don't go super well together. It takes an incredible amount of really hard work. Obviously, you've got the Teamwork Graph. We believe we have a big differentiator in our context and ability to context, not just across Atlassian, but across all the apps and work that we do. For those companies that have moved from the cloud to data center, the 2 examples I showed you, 15 to 20 years' worth of workflow data, code data, understanding of their knowledge, which we can give them back in a graph instantly that is accessible to all of their AI tools is a really unique differentiator that I don't think anybody else has in availability today. And we're not stopping. We're going to keep moving that advantage. And hopefully, you saw a whole bunch of examples today of this whole like, right, Jira is going to disappear. I call bulls*** on that. But as a control plane as a part of the Agentic and human collaboration, we see more and more customers using it. We've given you a bunch of stats. I've got a few more coming for you in a second. The Teamwork Graph, as we talked about this morning, what can I say differently. Most SaaS companies see only a slice of data. If you look at any of the people who live in a vertical, they don't see all of the information, which means they can't give you great answers. We see more. We have more links. We have more connectivity because of our breadth of workflows than most other SaaS companies, the vast majority of other SaaS companies. And I think we built a better engineering system underneath that and a better platform. So we believe that's different, and we have a singular platform. If you've got a lot of different platforms and your data is all scattered right now, and I wouldn't work there for the next couple of years, it's going to be miserable. And as we try to show this morning that are compounds, the more usage we have from customers, the more AI mail grows, the more that compounds the data they have, and we create a wonderful data flywheel. We're seeing this already working. Models are continuing to improve. That's wonderful. We have great relationships with all of the foundation model vendors. It's not our job. Our job is to take these wonderful models that are built and deploy them as quickly as possible and to use cases that customers want. We're down to sub 24, 48 hours on full model transitions where we have it running in our gateway. We often get early access to all foundational model vendors now. We get it running in 24 to 48 hours of public usage, and we can start running 1% of traffic through it to see how it performs really, really quickly. And if it works well, we can move it. If it cost us more money, but the results are better, we make these trade-offs. It's incredibly advanced orchestration system that we have, but it's really important at the scale we're doing. And we think it's a moat. We talk a lot about AI, sorry, I'll keep moving quicker here. We've made the Teamwork Graph a lot richer, as I hope you can see from this morning. Lots of new connectors, lots of new types of processes, a lot of new categories of content. Code was a massive lift. It is a huge differentiator to us, the code intelligence. Internally, it is getting used everywhere because the understanding of code is so critical to business process. It's not just technical processes, and it's not just about writing the code. We see it used in all sorts of different ways. We try to show that with the example in Service Collection today. We're an IT ops scenario to fix an issue, understanding the code makes a huge difference to how quickly you can fix an issue. Again, a true differentiator in that Service Collection space. And how it does reasoning over all the connections, et cetera, the graph is just amazing, and now we're opening it up with MCP, with CLI and with Dia. So again, if you want the breadth of what we're trying to explain here, as someone pointed out this morning, we showed everything from a terminal to a browser. There are a few companies in the world that can actually show you that and actually deliver all the R&D in between and huge customer benefits at all points. So a lot of excitement there. I'll stop myself before I go on too long on AI and the core things that we're doing there. Instead, I wanted to get 1 of our engineers to tell you about this. So there's about a 4-minute video and it's kind of nerdy, let you not hear from me for a second. Trying to understand what is the ROI of the Teamwork Graph for our customers. What is the opportunity for us if we can get this out there and show customers how it works? So watch this video. I hope you follow along and it doesn't go too deep. One second. [Presentation]

Okay. So I think hopefully, [ Kunis ] has given you some idea. That might go into it, if you wanted to just zoom out, the exact same Agentic kinase in this case, [ Cloudone ], probably the most popular at the moment. $2.68 to run that task, a real world, quite complicated engineering task. Without Teamwork Graph, $1.32 with Teamwork Graph. Don't have to be a Brian-level genius salesperson to go to a customer and explain why we can help them. Secondly, we actually finished in a minute in 10 seconds less, and we got a better result. The code needed less tweaking afterwards, it was more clear because it did a lot of searching of the context first, it understood business and technical data to get a better result. Exact same coding engine, exact same harness. The only difference is we added the TWG COI. And we obviously have a lot of our own code index, et cetera, et cetera, et cetera. So hopefully, that is a different example of the graph in action. You can multiply that across all the business agents that are coming across all the other areas. This didn't use people data at all in this case, but we have great access to that, et cetera. So when we say cheaper, better and faster with the Teamwork Graph, it's a very simple message to customers, and it's a very true message backed up, and nobody else can do this right now. So it's not theoretical. It's already showing up in the numbers. Let me show you what we are seeing with some AI stats. Some of you know, some that I think are new. So as we said in our shareholder letter, Rovo customers AI credit usage, their Rovo credit usage is growing at 20% month-over-month. That is a pretty good number. We're early in the curve. But we're already a very large player in terms of AI [ mal ] and usage, and the density of usage is growing very fast. Secondly, when customers adopt Rovo, they grow their ARR at twice the rate, twice the growth rate. And our expansion rate, as you know, is quite good NRR. They're growing at twice the growth rate of customers that haven't turned on Rovo. And lastly, it is delivering value, right? They are stickier. They move more. Again, we've given all these stats. I don't know if the 75% of the Fortune 500 companies that have turned on Rovo is a new stat, but it's there. And you can see our Agentic automations are growing really, really strongly. One way of using agents again is through automation, so putting them into various rules that happen automatically and can access all of the Teamwork Graph. So some AI stats. And when you look at the Teamwork Collection, which is our primary AI monetization vehicle right now, again, north of 1,000 customers, 1 million seats inside the first 6 months of upgrades of existing customers. Some new in there, but mainly upgrades Choosing to upgrade, why? You get roughly 10x the AI or Rovo credits when you upgrade. We have a lot of other economic advantages. And I would argue now advantages for the customer, you get the same number of Confluence, Jira and Loom seats, which increases your collaboration broadly, but again, far more AI credits. TWC customers use twice as many, who use, consume twice as many Rovo credits per paid user as non-Teamwork Collection customers, and they have twice as many active agents. Now they have more credits. So they've got more credits, they've opted in. Awesome. Give people more credits. They use it more. As long as that keeps growing, it's a really good wicket to be on. And I'd say this often enough, because sometimes I don't think you have differentiated. These aren't road maps. These are delivered in market features being used by customers today. We delivered the Teamwork Graph CLI today. The Teamwork Graph is used by all our customers and available to all our customers. They can log in teamworkgraph.com and see it today. Almost nothing we showed this morning is not available today. It's like 1 or 2 features that are coming in the next maybe 2 months, 3 months, everything else delivered on the day, right? That is our amazing capability. So obviously, you are saying, okay, well, how do we capture the value in that model? We're going to get asked about our pricing and monetization. So I put this together, I hope it's useful. We have multiple pricing models. We're trying to be customer led, we're trying to meet customers where they are. Most of our apps are priced per seat. I still think that's the way most customers want to purchase software. When I talk to customers, that is what they say they want. We have a very generous number of Rovo credits involved in those per-seat accounts. So we have the price per seat. I'm sure you're familiar with that part of our model. And these are sort of illustrative down the bottom. There's other things we could fit in. Secondly, we do have usage-based pricing, consumption-based pricing. We have a growing set of consumption-based offerings, whether that is Rovo credits, whether that's customer service management resolutions, Bitbucket Pipelines. We have assets, which is increasingly growing with some customers. We have Forge and Forge Compute and Forge Sequel in our extension framework. So we have a lot of usage-based models that are growing, right, and established revenue streams there. And we have our hybrid pricing, so that is things like the Teamwork Collection and quite a way to put that, where you can buy these apps individually. Most of them, you can buy the usage-based pricing individually. And then when you buy something like Teamwork Collection, you get a set of those usage meters together inside the collection pricing, right? So it's a different way of thinking about it. And we have a flex-based model that we are working on with customers at the moment that -- we're actually talking to a whole bunch of customers this week, which is more of the overall sort of commitment-based model, as I would say, is the way to think about it. So as our bigger customers that get well north of $1 million, so here's what I want to commit to your platform. I want to be able to move it around seats, move it around applications and collections, and I want more usage-based certainty. That is another model that we have. Again, we're here for the long term. This is us being customer-led, as they ask. We have the ability to move around this. And again, I would say the early signals are strong. Customers using MCP grow their ARR at twice the rate of those that do not. So the people use a run away from Atlassian, and I'm like, Oh, certainly not seeing that the data is, saying quite the opposite, I would say. Customers using our MCP server. That's before the CLI. We obviously don't have any stats on that. Early signals. MCP is a pretty advanced organization. So we need a lot more data to have that growing, but at the moment, twice the rate in ARR. And as we said, our Rovo credit usage growing at 20% month-over-month, moves you between different pricing models if you're seeing value. I think I've covered all that. I didn't know where to put this stuff. So AI isn't just adapting how we price, right? It's also creating new product categories. So here, we have a focus on talent and DX. Reason I put these up here, these are different applications to what we traditionally sell. They're really only enterprise applications, sold at large scale. If using DX to measure the productivity of an engineering organization, you probably have 500 or 1,000 engineers to start with, right? You're not buying this for 10 people. You don't need Talent for 10 people, you need to look around the room. These -- as workforces adopt AI, where we're seeing those customers like the financial institution I showed earlier, they are using us in their how do I think about myself moving from AI novice to AI native. How do I look at my talent, my skill mix, who can help me see that. DX is all about understanding how AI native my engineering force is. And which teams which services, which parts of my organization are faster and slower, which are using more AI, which are quicker and not quick. That is a phenomenal business. I wanted to make sure I put DX down there. It's the head of our projections. We're feeling really good about that business. It's only a few months in. But again, all in that area, it's not necessarily a different pricing model, but a different type of application for us. Again, as we have these customers who've come in and opted into us as a platform, it's those customers that you saw earlier that are likely to open it up. And these are all used by the C-suite. As we continue our enterprise journey, I spend a lot more time with CEOs than I ever have and CIOs. They're using these applications day to day. They don't probably sit in Jira so much, but they certainly sit in Talent. They certainly sit in DX. They certainly sit in Focus. They're very important to them, helps continue our journey on that. And there's a prerequisite for all of this. This one is going to kill me, we should probably talk about cloud migrations. A little bit of a topic. I haven't gotten there yet. I'm taking too long. Customers are moving to the cloud. We're doing really well. I think you know that and you've seen it. It is the ultimate destination. It's the value of Atlassian, where customers get the most value, can't get AI, can't get the Teamwork Graph, can't get anything. This, I think I should say, is illustrative, by the way. So we've been very thoughtful and systematic, I hope you can see, about our multiyear, decade-long cloud transition, both in terms of the building of the future ahead and moving customers I think we've executed extremely well across the hundreds of thousands of server customers. Now we're getting the data center customers at some scale. But the migration still has a ways to go in terms of the number of seats we have in the data center. That's a good sign. What does this look like? What is an illustrative example of customer pricing? It's always hard to pick an example. 93% of our data center customers that are upgrading to the cloud are landing in Premium or Enterprise. And I know this because in some of the models and some of the things we've seen, you can think, "Oh, well, a data center customer could go backwards moving to cloud standard." That doesn't happen for 93% of those data center customers. Can mathematically happen, but I thought it was useful to say this, this is a -- what did we put, 5,000 -- sorry, 5,000 user data center tiers, so that's 4,200. Again, in the cloud, you buy what you need, you don't buy sort of bigger tiers infrequently. So we've talked about that in the past. The interesting part that probably isn't news. The interesting part is what happens after they migrate. So now we have many years of history of moving large-scale customers. So let's look at multiyears, let's take a 3-year period. Let's take 1,000-plus user customers. So only the large customers have moved, 1,000-plus users that have upgraded to the cloud, what happens in their 3 years after they moved to the cloud? Well, they grew about 1.75x, 1.5x to 2x growth in the 3 years after they migrate. Why? All those things I talked about before, seat expansion, cross-sell. They can get Focus. They can get Talent. They can get DX, they can get the Product Collection. They can get the Service Collection with all the capabilities. So you saw that in some of those customer examples earlier. It's saying that we're being successful at moving those customers. And when we move them, we have a good couple of years after they move. They continue to expand their footprint after they move because of all the reasons we talked about. And their ARR growth continues strongly after they migrate. We're not collecting all of the value at the point of upgrade. In fact, we're trying to trade that off over that 3-year period. And you've probably seen this chart before, but the cloud is not a flywheel, right? Sorry, it is a flywheel. What am I talking about? It's not a single project to upgrade. It is a flywheel that moves into the cloud, and they continue to grow. That's what we're trying to show over time. It allows us to continue to ship our AI faster, monetize all those new use cases, capture those workflows, right? Compounding our momentum with consistent 120%-plus NRR growth at scale in the cloud. That is what we're seeing. We had this rainbow chart when we went public. I love that we just keep adding years to the rainbow chart, and it keeps it going. So we have a huge opportunity. We're crushing it with customers, voting ever more for us as a platform. We feel incredibly strongly. Brian will come up. You will hear from him. Our cloud growth is extremely healthy. Our AI progress continues to be extremely strong, and we're -- hopefully you can feel for me and everybody on stage, we're so bullish about what we're doing in that area. NRR, 120% strong. You might be asking it can be difficult to see this in your financial statements, on the face of your financial statements, with ASC 606 through this migration. So what can we talk about that? Well, I want to cut through a little bit of noise before I hand over to Brian. And again, is anyone familiar with this logo? Extra points for investors, if you can tell me what it means. Open company, no bulls***. Love it. Yes. This logo means open company, no bulls***. It's 1 of our 5 company values, and we want to be direct, right? So 2024, 2 years ago, we issued expectations on a 3-year, 20% plus revenue CAGR through FY '27. Since then, we've had a few major changes. In September of 2025, last year, yes, that's right, calendar year, we announced [ Ascend ]. So the end of life of data center by March 2029. First thing I want to stress, that is a great thing for our customers. It's a great thing for our R&D organization and our business. The challenge was made for us, we're a year ahead of schedule. We delivered Ascend in September 2025, not September of 2026 due to all the enterprise progress we had made and the R&D we invested in. That was ahead of our schedule. That's great for customers. That's great for our business, enabled us to accelerate our customers' journey to the cloud. However, as a result, as you saw last quarter, that creates 3 interesting dynamics that affect the data center revenue line. First, with ASC 606, we moved from 20% upfront revenue to 50% upfront revenue. And you model that with the 1-year acceleration, creates a challenge for us on the license line and data center subscriptions. Secondly, we moved our customers, which was not predicted 3 years ago, from multiyear deals to 1-year deals. A couple of reasons why we did that. Firstly, when you have a 3-year end of life, you can't tell someone a 3-year deal a year later. It doesn't make any sense. Secondly, we want to have multiple opportunities to talk to that customer. So we've moved the vast majority of our data center customers to 1-year licenses. Model both of those together, you see what the effect is. Those 1-year licenses give us 3 opportunities to talk to that customer over time. They often say, "I get it, but not right now. I've got busy things." Business AI, I think have you not heard about it. We have a conversation. Yes, we're good at that, et cetera. And lastly, data center customer purchasing patterns have continued to emerge and accelerate as we move to cloud. Again, all these are good things. We saw that play out in a pretty big way last quarter, as I know a number you heard about, with a lot of pull-forward activity into FY '26 from FY '27, which results in greater licensing revenue in a DC in '26, far more than we had expected, which is a good thing, in general, timing dynamic creates a problem. So we continue to see strong cloud growth rate. Great for our customers, great for our business and data center retention, but the greater term revenue in '26 in data center, we now expect negative data center revenue growth in '27. We've pulled all that forward, which is a great thing for us. So I tried to explain this, both overcoming no bulls***, and I'll give you some illustrative charts. I thought I'd say all that to we're listening and I said we're running before, I showed you the charts. So this is an illustrative revenue growth. Number one, we're expecting a trough in total revenue growth in FY '27. Number two, we expect that to significantly reaccelerate in FY '28 as we lap that data center effect going through. You can see why that 1-year acceleration is a credit to the engineering team, creates a challenge for the finance team. And given these dynamics, the 3-year, 20% CAGR that we set [indiscernible] years ago is no longer a relevant target to anchor on. What is relevant and a much better measure of the strength of our business is [ ARR ]. Again, this is illustrative. I get at your rules. So why is that? Well, ARR normalizes the effect of ASC 606, that revenue recognition, and helps everyone, I think, much better understand the overall health of our business. So if you take ARR subscription business, so data center and cloud, put it together, to a lot of great mathematics that our finest team is able to do, you see a very different pattern. Here is our subscription ARR for the last 7 quarters. What you can see in the last 3 quarters is a pretty good pattern of acceleration of ARR at the overall level of the business. One of many indicators, but I think a really strong one, showing how strongly, we feel about how our business is operating and trying to explain the dynamics between all of these different things. I'm sure we'll have time for questions on that. Last one I wanted to talk about, 1 more thing before Brian comes on. What do we mean by the phrase durable, profitable growth. Well, I hope we've shown that we have a durable growth story both in our results and in the go-forward opportunities we have in AI, System of Work in the Enterprise and all the things we're doing in cloud, the customer stories, the customer scale, all the things we've done. The profitable word is also really important to us. We've been a capital-efficient business for 24.5 years. And I think we will make sure that we are continuing to be so. We also expect, as a result of all the changes that we've made that was not taken into account 2 years ago to accelerate and grow our GAAP operating profit, beginning in FY '27. I'm sure we'll have more on that in the future, but I wanted to put that profitable and the durable and the growth, all 3 words incredibly important to us. So we've made a lot of the great investments. Enterprise, AI, System of Work, hopefully, they add together to see why we have that durable, profitable growth over time. But you probably want to hear a lot more about our large customers. Let me bring up Brian, our fabulous CRO, who has lapped 1 year and a little bit in the seat. So he has a lot more to tell you than he did a year and a bit ago. Here's Brian.

Thank you, Mike, and good afternoon, everybody. Great to be here with you. So I -- as Mike said, I've been here a little over a year. And I guess I'm still relatively new at this, I'm running with us. . And I'm regularly asked, why did I join Atlassian? And the answer for me is pretty simple. And that is when you look at Atlassian, we are a very unique company. We obviously are known for the incredible PLG motion that we have. And we have scaled the business to be billions of dollars, as Mike has laid out. And we also have 350,000 customers. Now we did all of this without having a mature enterprise sales team. And many other enterprise organizations would say that they have somewhat similar ways of acquiring customers, but in reality, they don't. And for me, when -- the reason for me to join is because we have the opportunity to take that massive frictionless flywheel that we have where customers really love the product. And when I joined, Mike notes that I regularly said I'm surprised by how much our customers really love Atlassian and love the product. And they are expanding with this organically. And now the opportunity is to layer on top of that, a world-class enterprise sales motion. But importantly is that at the same time, we're not moving away from our PLG heritage, and that motion is still running at the same time. Instead, what we're doing here is we're going to be spinning a new gear, and I'm hopefully going to show you exactly how we are aiming at doing that. So for me, really, this starts with the foundations, and you can't build a great enterprise sales foundation if you don't have enterprise-grade products to ultimately back it up. And if you look at the landscape today and you see the slide behind me, Atlassian is recognized as a leader across all of our major product segments. And we are putting the necessary sales muscle behind these products in order to capitalize on the opportunity ahead of us. And the good news is, as you saw on the slide in terms of the ARR growth, we are already seeing the positive momentum. And over the last few years, we have been evolving the business itself. And that shift is still underway, and the transformation is an evolution. But in the 16 months that I've been here, I've been very pleasantly surprised with the progress that we have made and the wins that we have under our belt. So I want to go in a little bit deeper in terms of some of the numbers that prove that the motion is working. So in Q3, our volume of deals over $3 million surged by 79% year-over-year, which we are very proud of. It's not on this slide because we just decided to add it last minute. But over the past year, we saw a 54% increase in deals over $5 million. And that obviously is something that is significant and has a strong contribution to our ARR and RPO. Now not only are we closing more of these larger deals, but the ASP at the same time has also jumped, and that's jumped by 22%. The best part of all of this is that we're not feeding the larger deals at the price of the smaller transactions. Because what is happening is that our smaller deals are also growing at a clip of 32% as well. And all of this is happening with our retention rates, like Mike said, at 99%. So when you step back and you look at the business, you can say we're firing on all cylinders, from the top of the enterprise into the middle of the pyramid and then at the bottom of the pyramid as well. There's no piece that we are sacrificing. And at the same time, it's not just about the size of the deals that we have. It's also about the durability of those relationships. And something that obviously is hugely important to us is the time frame of our commitments. And our customers are now committing to us for longer time frames. As we shared in Q3, our RPO grew to $4 billion, and that is an increase of 37% year-over-year. Now I know you regularly ask, as these customers migrate to the cloud, what happens thereafter? The good news is that they're not sitting still with us, and neither is my team of salespeople. They're driving customers to expand in the cloud and expand across the Atlassian portfolio. Our team is focused on uncovering new opportunities for our customers and driving product adoption. Also cross-selling into new collections and at the same time, closing complex, high product deals as well. Now our customers, it's fair to say, have more options than ever before. And they are clearly voting with their wallets and they are expanding seats across our core products and adopting our additional offerings. In terms of when we look at our collections, that certainly is being led by Teamwork Collection. And as Mike said, that is our primary motion for AI monetization. And then being led by Service Collection, which you saw the numbers there in terms of the size of that business, and we're continuing to make progress there. And if you look at the slide behind me, you can see the outcome of the strategy. And these are just some of the logos that have placed their trust in Atlassian over the years. And many of these have been customers for 20-plus years, and their footprint has grown with Atlassian. But one thing that I would stress is that these brands grew their footprint, firstly, by themselves without having an enterprise sales team in place, which speaks highly to the products, first of all, and then also speaks highly to the ecosystem that we have. So I thought it would be interesting, similar to what Mike did earlier, but with a little twist, to show you 1 of the brands and the evolution of that particular brand. So the brand that you can see behind me is a semiconductor company, and they have been a customer of ours since 2015. They were a low-touch customer, so there was no accounting associated with this customer. A story we've seen many times over, they were a Jira and Confluence on-premise customer and they were experiencing hyper growth. This particular customer had internal systems that does not fit for purpose. And then their management made the decision that they needed to transform and that they needed to modernize. In 2023, you can see that we made the decision that we would assign an account team to this particular customer. And we consciously made the decision that we would build relationships across the entire business. At the same time, as an enterprise sales team, we would also build relationships with the C-suite within this particular customer as well. And we were not pitching a lift and shift to this particular customer. We were pitching a complete transformation and reimagination of their business, and we were doing that in conjunction with our partners as well, all obviously to fuel their hyper growth that they were experiencing. Now once an enterprise account team was in place with this particular customer, you can see what happens, the investment here. It multiplied because obviously, they bought Strategy Collection, they bought Teamwork Collection, they bought Guard Premium and they've invested in advisory services as well. Now this customer is just 1 customer journey, and I could have shared many others with you. But what's particularly promising to me is that we are hiring and I'm building out the organization, but there are many stories like this to come where we will have a similar journey and a similar investment projections, which are obviously going to change. So we will see many similar stories like this as we move forward. Now even with these wins across our enterprise base, it's fair to say that we've only scratched the surface. We have a $140 billion total addressable market, and we have an incredible runway in front of us with our 350,000 customers. And we also have a massive opportunity to go out and attract net new logos as well. So I walked you through the opportunity, but internally, we are obviously looking at how are we going to go after and capture this opportunity. And we have a transformation, and there's a lot of work happening behind that transformation. But today, I wanted to walk you through just 2 particular pillars, which is customer-centric selling and our partner ecosystem and give you a little bit more detail around what we are doing in terms of those pillars. So when it comes to customer-centric selling, we are evolving the go-to-market motion within the organization to become a trusted strategic partner to our customers. And obviously, coming from a PLG motion, this is a complete transformation of the sales team and the [ VAT ] team and the presales team around us. We are now engaging like I said, in the example, beforehand with the C-suite to deeply understand their business challenges so that we can help them address those business challenges and we are now delivering business outcomes for our customers. In many respects, we are slowing down in some of our deal cycles so that we can actually grow the transactions, and then we can speed up and then we have a better result at the end. This is translating into stronger customer wins as we said, in terms of the logos. Secondly, in terms of the value of the deals and then in terms of the length of the commitments from our customers as well. Now secondly, we've also realigned our sales incentives to reinforce our strategic priorities. We are now enabling our AEs to focus on the outcomes that matter most strategically to Atlassian. We are incentivizing them in a big way to uncover new value within our customers. Obviously, focusing on our collections and expanding our footprint within these accounts as well. We now have a specialized sales force for those collections, which are focusing on new demand within each one of those customers. And as you can see from the numbers that we've put on the board, uncovering new value and new demand is also playing out in our favor currently. Now we are appointing the team directly at our biggest growth levers. And I touched on this already, which is customers are expanding across our collections. And our team is looking at ways to grow accounts, which is great. We are in a very fortunate position that we have a lot of white space around the world as well in terms of countries where we are not present, and we have an ecosystem that is present. We are also operating in high-growth markets as well where we are making investments where our customers are present. That is in countries like the U.K., France and India and a few more. It's very clear from where we sit that the demand is there. We are expanding our growth horizon. We're prioritizing new markets and cross-sell opportunities through our collections in order to unlock that incremental revenue and opportunity at our customers. Now to capture this opportunity and to sell to these enterprises, we're obviously making investments and to ensure that we have the right amount of resources, quota carriers. Now what you can see on the slide here is that 4 years ago, believe it or not, Atlassian had 117 quota carriers. So the sales force was 117 people. And now since I have arrived, we have grown that to 400. So I constantly remind Mike and James of our productivity, which you can do the math, is best-in-class in the industry by far. And we are clearly punching above our weight. So obviously, as we moderate the spend in R&D and we look to make the investments in sales we obviously have an opportunity to continue to make significant improvements here as well. But we aren't just adding head count blindly, I would say. We are tightening our coverage ratios when we look at where our customers are. We are increasing our quota carrying capacity relative to the overall sales headcount that we have. And we ultimately are building a very mature sales organization. And what's exciting is we have the opportunity to learn from the mistakes of many organizations who have gone before us, which is an exciting spot to be in, and that gives us the opportunity to scale it very efficiently also. Now even with the 400 sales reps that we have, we know that in order for us to capture the opportunity, that we won't be able to do it by ourselves. And that brings us to the second area that I said and mentioned, which is our ecosystem. And for the first time at Atlassian, we are now partnering with the GSIs. And I think that's particularly important because obviously, if the GSIs are involved, this speaks highly to the opportunity that they see. So we are now partnering with Accenture, Deloitte and PwC in a big way. They are giving us access to the C-suite at a scale that is unprecedented for Atlassian, which is great. Earlier this week, we had Mercedes on stage. Mercedes is a customer where we partnered heavily with Deloitte and would not be in the position that we are with this customer if it wasn't for the partnership there. At the same time, our partners are building dedicated solution to drive collections and AI adoption. Earlier this year, I was with Infosys in India. And they have just launched a center of excellence in order to support their booming Atlassian practice as well. And now when we look at Accenture, where we have a very deep relationship, they are also driving significant pipeline for us personally. And I would add that we're tracking all of our GSIs in terms of the incremental new pipeline that they are bringing to the table for Atlassian. And we're continuing to see an evolution of our relationship with those GSIs. Now at Team '26 this year, we have over 1,000 partner attendees. And I spent time with them on Monday, and I explained to them how we are committed to helping them evolve their business also. You all know that in the earlier days of Atlassian, we would reward those partners really on a transactional basis. And we are obviously now changing that. And we are now shifting our incentive structure really to be towards high-value strategic services. We want to help them grow their services from a one-to-one to ideally a 1 to 3 or 1 to 5 ratio, and we are making the necessary investments to help them grow that business as well. And we are seeing a return already because we are seeing by incentivizing them in terms of the right behaviors, that, that is leading to a higher adoption for us and better customer outcomes as well. This is obviously a big evolution for the ecosystem overall. Now this is obviously a -- I've walked you through the combination of our product-led foundation that we have and our enterprise sales muscle that we are building and evolving and the ecosystem, which is critically important to us. And I could share with you a lot of numbers and a lot of slides around how we're very successful and how we're knocking everything out of the park. But instead of doing that, it's best if I could welcome onto the stage, two customers who can actually share with you what it is we are doing. So if you could please join me in welcoming Jason Andrews from Cisco and Stephane Baumier from Canal+ to the stage. All right. Well, thanks for joining us, first of all. I appreciate it. So maybe we will get started with you guys quickly telling us about the company and then what you're responsible for. And Jason, we'll get started with you.

Yes. So I'm with Cisco Systems. I run the engineering operations function across our product development, specifically for the networking business. That results in a like kind of helping guide the -- how engineers work for around 25,000 users. I also have a couple of other functions and program management process ownership, which is a real asset. And then on the back end, having our global app services, which is like a large data center environment.

So thank you. So my name is Stephane Baumier. I'm the CTO of Canal+ Group. So perhaps you don't know Canal+, but in U.S., it's not so famous. Canal+, it's a media and pay-TV company, a French company, which is based in Paris and headquarters in Paris, but we broadcast pay TV in Europe and in Africa. It's a EUR 7 million of turnover to give you an idea of the size of the company, we have more than 40 million subscribers. And so I would give more information in the link.

Okay. Great. So Jason, Cisco, obviously, the company, everybody in the room knows has evolved over the years and has 1 of the most complex and operational footprints in the world. And when you look across the organization, our relationship with Cisco has evolved. We initially had been, let's say, viewed as a tool that you were using. And I know from the conversations that we've had, we're now viewed as being a mission-critical to Cisco. So maybe you can share with everybody, how that has evolved over the time?

Yes, it's definitely evolved fast. I think it's evolving the way we work, right? A lot of ways, if you view it as a tool, you'll only have a tool when you look at it as a platform to deliver product. As we start to leverage things like the Teamwork Graph, again, bringing all this work we're doing across 25,000 to 30,000 engineers making collaboration easier. And a large company, if you work in a large enterprise, collaboration is the hardest thing you're going to do. Building a product in a single team, super easy. But as you start to say, I'm going to blend security into the fabric of our networking. I've got to get the security group. We're going to get the networking group on the same page. It's extremely hard because they have different models of working. And as we started to align those groups, we've seen great results in terms of how fast we're able to deliver product to our customers. And we've actually seen -- what's also about this is we've seen huge gains in productivity as we leveraged a Teamwork concept or a System of Work concept. We saw as much as a -- for about between 18,000 and 20,000 engineers, an estimated 5% uplift in productivity just by adopting the System of Work. On our own side, we actually saw like a 40% reduction in our TCO of actually managing the platform as a whole. So it's been a great journey.

Awesome. And Stephane, anything you want to add? In terms of the journey that...

Yes. We have the same idea. In fact, as Canal+ is 42 million subscriber, we would like to target to 100 million with organic growth, but also with acquisitions. So acquisition means M&A. And at this moment, convergence and synergies are key. And Atlassian is really a part of the suit to be able to manage this convergence and to adopt all the same process and the standardization that we can have in the different countries in Europe, it can be in Poland, in France, but also in South Africa.

Great. And so Stephane, you had shared with me how you moved to a unified System of Work as well and change your global decision-making and collaboration system. So what happens within Canal+ if, let's say, that was to go away?

That's a good question, especially in technology. I work in a 30 years now in technology, and I have a different job in the world. And not in Europe, in Africa and in Asia. I was the CEO and CTO. So I have to manage both side. And I was appointed as a CTO to create a global technology department with 6,000 people. So standardization is a key to have exactly the same. So to answer to your question, if it's going away, we will continue to stream. We'll continue to broadcast some live sport to use the satellite and OTT towards. But the key is to take some decision for the projects. Because we can have some project in Poland, in South Africa, in Senegal, in France or even in Myanmar, where we broadcast. And from this, if we don't have any more the different system to do it, we will not be able, and we'll have to do what, to have new Excel file, spreadsheet, hours of meeting, discussion is finance. And to take this decision will be so complicated, which is not the case today.

Awesome. So we know the criticality of Atlassian within both companies. And we know that AI is top of mind for everybody, including both of your companies. So maybe Stephane, to you, can you share with everyone how -- and I know you're a Rovo customer as well. So you can share with everyone how Rovo is helping accelerate the decision-making within Canal+ and how you're utilizing Rovo?

Yes. AI is a key. It's a revolution for everybody. The result of promise since few years that we'll have 50%, 60%, 70% of productivity. But when you have to manage a worldwide company in technology, you need to have a P&L also and to have some results. So each choice that we have to do with AI, I have to answer a different subject, cybersecurity, finance and usage. For cyber and for finance, we use Rovo because we know all and we validate all. After for the usage, it's not just a key for the tech department. It's a key to deploy in the different departments. For example, if you have to manage projects, the CFO of the group, it's EUR 7 million of turnover for Canal+, for example. When you have some questions, what he did in the past. He asked me to say, okay, what is the cost of the program? Where is the program? What is the CapEx? What is the OpEx? And all some meetings as I said before. And now with Rovo is able to ask directly for any project, any programs and to do this kind of validation. That's a very crazy situation for us and very useful. And the dynamic can be that the CFO of France, of course, but of Poland or South Africa can have the same question for his own project.

And just 1 thing so it's clear for everyone, like Atlassian's footprint within Canal+ and then how Rovo is being utilized, it's not solely within IT as well?

Yes, exactly. IT, in fact, at the beginning, we use like a lot of company confluence because we have all the documentation. And after we start to use the different suite of Atlassian and last year, with Team '25 when we discovered Rovo, we said, okay, this is a key now to be sure that we can deploy all the suite for all the departments, which are not technical.

Great. And Jason, anything you want to share in terms of Rovo?

I think we're seeing a lot of value and really focusing a lot on the time to revenue aspect. A lot of times, a lot of people are focused on engineering productivity because it is obviously a huge win in kind of early focal point of AI. But really, what we're trying to look is leveraging Rovo to actually run a different business process for us, right? How do you actually standardize and automate the reporting? How do you make these things faster? How do you stop spending time pulling status together from different pages, meetings, everything to actually focus on solving the risk and issues? We're seeing the team really lean in heavily to it. Our Rovo usage is going through the roof. It's great, right? I'm getting great feedback from the team and how it's making their job easier, how it's easier to collaborate across the organization, even ones that aren't on our system will work. Because, again, we're still getting everything together the outside teams. But so as those teams like to bring features and data that are attracting a slightly different way from this instance to this instance. It used to be that was a very manual spreadsheet-driven process. Now leveraging Rovo, we automatically pull that data in. It creates dashboards, it does all the things you need to do so people can focus on managing risk in issues and not bringing status reporting as their primary job.

And maybe in terms of the transformation from a change management, and maybe both of you, if you want to show the change management a difficult process? Or...

I think it's impossible. I think at the speed of AI is going, people I keep on saying that the change management. And I think we're spending a lot of time kind of democratize AI and make it very easy for us to leveraging the Atlassian platform, prebuilt agents and things that people can get value about, but they're having to understand system architecture, codes on all these kind of crazy things. But I laugh when it talks about a change management. What I thought we were going to be doing with AI 60 days ago is different than today and definitely different than a year ago. So it's a tough ask, I think.

That's tough. And change management is human, in fact, human relation to be sure that people can adopt what you would like to do. What we need to be able to change the way that we manage this is to go to create a clear strategy that we can adapt for each country. And for this moment, as a human adapt the strategy, we will be able to do exactly where we go and where we would like to go. And at this moment, they will be ready to change. But it's not so easy, for sure.

And then maybe, Stephane, do you want to share with everyone your vision in terms of the future of exact reporting?

Yes. So [ means ], is that in fact, you use for example Loom, you just record a video because you have an idea. So you can imagine, you are in [ Jonesboro ] in [ Deca ]. You have an idea, you record your video. You have an automatic translation of this video to create the project in the system. And everybody, all the countries are aware of this new project and can start and to expand it. This is really our future, what we would like to target, especially when you are in different countries. Different culture to have 1 common product where it's easy to speak your language because we have many languages. It will be easy to speak, to translate. And everybody have this information. And at this moment, we can create also process to create it with the Atlassian suite. This is really key for us for the virtues where we go and battalion is going and the discussion that we have during the different conference is really the past in fact, of Atlassian. So it's interesting for us.

And how do you see our relationship evolving in the future?

With this, very good. As I said before, we define a global strategy for the group to be able to manage M&A. And when we define the strategy, we have defined a list of trusted partner, where we have a good relationship. So it means that we are confident with partner, not only in terms of legal, in terms of finance, but also in terms of product. It means what is important for us is to have some partners like Atlassian or [ Datadog ], for example, we have many -- we have 10 partners what we select. When we can exchange about the product and to give some advice, if we can say advice or what we would like to do, what is our remand for the moment, each time, we have this kind of discussion, we saw that it's coming after a few months or 2 years. We are calls, okay, we have these features. You asked it for 1 year ago, and we have it. So the relationship is really important, not only in the contract and the finance, but the relationship is the trust that we have and to create some -- to create the future of this because the evolution is very big. AI is a revolution for everybody, including for you. So we have to be confident in cyber, finance, legal and to be partner intact more than to be just a customer and a reseller.

Great. So then maybe, Jason, to you as Cisco's business continues to evolve and change and grow, how do you see Cisco's relationship with Atlassian evolving and then the reliance that Cisco has with Atlassian as the business is going to continue to change and evolve?

I think it continues to be this massive partnership, right? There's a lot of 360 relationship, we go back and forth. Obviously, Atlassian really steps in and lean in when we have problems. So I think the relationship couldn't be better from an enterprise service standpoint. I think when you look at the future and how it scales with us, it really is I think doing a great job. It's literally leveraged, I think Mike Cannon-Brookes today and the things said something that really was impactful to me is he said, acceleration is context times intelligence, right? I think that is really where we're heading. As we start to develop AI native applications, things that are written 100% in this vibe coting method. The System of Work in the frame is kind of required to do that, right? You're going to need those the product requirement do built of the feature link the dependencies track in the program management space. We're leveraging Rovo to go out and look across our entire stack to help us identify dependencies we were not aware of, right? These -- all of these programs and these motions are going at a constant pace. We deliver around 15,000 features a year in networking alone. That's -- these are top-level epic. We have about 6 engineering epic. So you're really talking about a platform around 80,000 to 100,000 features, it's impossible for a human to track and connect how they work. But something like Rovo, it's really made our job in driving visibility into the work we're doing is super impactful.

Awesome. Great. Well, maybe to close the side, maybe -- and Stephane, we'll go to you. Maybe you want to kind of wrap it up in a short answer in terms of like the ultimate business value that we drive to share with everybody the ultimate business value that we drive for Canal+?

In fact, I think we have 3 business value. One is let's say, more but compliance. We are a listed company in London Stock Exchange and very soon, in Johannesburg. So we need to have some partners that we trust to be sure that the process will run for a daily basis and 24 hours per day, 7 days per week, so which is really a key success for us in terms of business. The second one is really operational. What I said before, Atlassian suite is more for technology department, but more and more for all the departments of the company. So this is the business value that we'll bring when we start to set up the application on the smartphone of the CFO to say, now you just have to check for the project and to have the cost of the CapEx and what we spend on what is the status of the project. And the last one is more for the development. As I said, we would like to continue our development with M&A. We have some shares in [ Viaplay ], for example, in Nordic Europe with a view in Southeast Asia company who -- around 30%. And each time we do some M&A, each time, we bring in our suitcase or partner to say, guys, now we know exactly how to manage a convergence in the synergy, which is a key for all the M&A. And Atlassian suite is a part of it. So if you do the addition of the business value for this operational, this compliance and the organization is really key for us, and it will be the value at the end of the day for us.

Awesome. Great. Jason?

I think the ultimate business value, if you break it down, is it becomes a platform to deliver product, right? Our time to revenue improvements that we kind of mentioned earlier. I think the real power is it gives you the context in the system. And it also helps you orchestrate it, which, again, builds a platform a collaboration for a team to centralize that work on and it really helps them stay in touch. It's incredibly hard in organization. I can't reiterate that point enough. And they've really done a good job of kind of helping us lean into that and get this sorted.

Great. Well, I want to thank you both, firstly, for the partnership and the trust. I appreciate you coming here and sharing your story with everybody. I have a great job because I got to spend time with awesome customers like this every day. So if you could join me in giving these 2 gentlemen a round of applause. All right. And with that, I'm very happy to hand over the clicker to the new guy, James. So if you can give James a warm welcome.

Good evening, everybody. I consider myself really lucky to be able to kick off my Atlassian journey right here at Team. This is my first one. And it's been really inspiring to spend time with our customers, our partners and so many practitioners and users that are super passionate about the Atlassian platform. Before we get to Q&A, I know a couple of you asked me last week during our call back, what drew me to Atlassian. So I thought I would just share a couple of thoughts. First, I would say that I've long admired Atlassian. This amazing culture of innovation and this very powerful mission of unleashing the potential of every team. Right? It's such a powerful enduring mission in many ways, sort of an infinite TAM. I've got my GC back there waving and say, "You can't say infinite TAM." And we know there's a lot of noise right now in our market. But if you punch through that noise, what I see is an incredibly well-positioned company, right, one that has a diversified and durable set of businesses, building a platform that's serving over 350,000 customers and hundreds of millions of workflows. I've been in the role now for a little over a month, and that conviction has only grown. There's a couple of reasons for that. And the first is that we are incredibly well diversified, not just across tech teams, but non-tech teams, business teams, it's allowing us to play in these massive expansive TAMs with line of sight to $140 billion of opportunity. And you saw some of that momentum play out in Q3 with the outperformance that we saw in TWC while also expanding our Jira seats on our stand-alone offerings. You saw Service Collection grow to over $1 billion in ARR, growing 30% year-over-year. And that value, that ROI that we're delivering to our customers right now is positioning us really well across all of these segments. The second is how well Atlassian is positioned within that enterprise workflow stack, right, serving not just developers, but HR teams, finance teams, legal teams to help them collaborate and manage an ever-increasing amount of workflows. That's only more amplified with AI now. And it's early, but we're still seeing a lot of amazing traction. Mike talked about it a little bit earlier, AI credit usage growing 20% month-over-month. Customers that have unleashed Rovo, they're deploying 2x more agents. They're using 2x more credits, and Rovo customers are growing their ARR at 2x the rate versus those who haven't yet, the keyword being yet. So that's off to a great start. You heard from Brian just now, the progress that we're making across our enterprise motion and with our partners and a ton of headroom still with us being penetrated into 85% of the Fortune 500, but that only represents 10% of our revenue. So a lot of headroom still ahead of us here. And maybe finally, you heard Mike talk about it a little bit earlier. We talked about this in our shareholder letter last week as well, which is alongside AI, Enterprise and System of Work, we're elevating key strategic priority of driving durable profitable growth. Right? You heard me talk about it a little bit earlier. This culture of innovation, that's a big part of what drew me to Atlassian. It's what drove all this enterprise value over the last 2 decades. And so we're going to continue to reinvest in areas that's going to drive the pace of innovation and scale that's required to win in this era. But we're also going to do that while driving the culture a strong fiscal discipline while we accelerate our path to GAAP profitability. So just a ton of opportunity here. Really excited to be here, the privilege to partner with Mike and the entire Atlassian team to really steer this next stage of growth for us. It's a wonderful time to be in our market right now, and it's an incredible time to be at Atlassian. So thank you, all. All right. With that, why don't we go into Q&A? I'll ask Mike and Brian to join us.

All right. We'll open up for Q&A. I see Gregg in the front row, so we'll start with Gregg Moskowitz.

All right. Thanks very much. Gregg Moskowitz with Mizuho. Mike, the fact that Atlassian customers will be able to use MCP and other external tools to query the Teamwork Graph, add to it as well. Super interesting. What does this mean though, not only for your relevance over the long term, but also your ability to continue to grow at a healthy rate? Why is this going to be a winning move for Atlassian?

Why is it going to be good news for Atlassian. Let me give you 3 reasons. It's probably many. First, again, we're seeing the current -- we've had our MCP Server market for 6 months or so, something like that. Quite a lot of adoption already, growing very, very fast. The customer is using it, and we're quite a medium to high level of adoption. It's hard to know. There aren't a lot of MCP stats around there, but we believe we're certainly in the top handful of used MCP servers at scale anywhere in the world. Customers using it. As I mentioned, we're growing significantly faster than customers that aren't using it. So if they're AI-native organizations or the most forward thinking, the early signals we have is that, that really helps accelerate because they can do a lot more on the platform. Those customers are in the cloud. They're using Rovo. They're using the MCP server. That generally will result in more workflow. So the data we have so far, it's very early, it's very good. Secondly, it allows them to do many more workflows, right? They can use other apps. So the example we showed on stage, Figma, a lot of customers, Figma and Atlassian overlap, a good partner of ours. We use a lot of Figma things, they use a lot of Atlassian things. And the customers that use both of those is a live subset. It enables both platforms to play with each other in terms of data exchange and usage. It makes Figma make better to be able to access the Teamwork Graph. It also will result in write-backs to the Teamwork Graph. So it grows our graph, which then makes the next operation. Maybe you then move to Claude Code to code after designing or something like this. That's going to help us from a network effect point of view, if you think about sort of the data network effect of the Teamwork Graph, right? People that are reading and writing from it, the more access there is, the overall the graph size grows. It also keeps it fresher and newer because it's the active workflows that are really important, right? The latent graph is not. It's how it's used over time and continues to evolve and grow. And lastly, there is a good angle for us in terms of being more actively used by customers. If we are more actively used by customers in a generalized sense, I always believe that is better for us. We don't have a lot of data on this yet, but if you look at the very, very small amount of data that seems to be true in like now and all sorts of other signals as well. So I think there's a lot of reasons why I would generify MCP to like app and Agentic usage of our AI stack, if you like. There's a huge amount of AI that goes into making the Teamwork Graph. And so we obviously have a Rovo credit story. Again, too early to see how that one pans out, but we feel very confident about it. But overall, both the CLI, the MCP and other usages, I think, are very, very positive for our overall story.

I guess we'll go with Peter Weed in the front.

Got to keep the mic up here close. Peter Weed from Bernstein. Maybe I continue the conversation around AI, obviously, both the hot topic for your users and adding a lot of value, but investors have really been kind of wondering how can we see the upside beyond kind of like the strengthening of the platform like directly in revenue. And I think there's kind of two sides of this question. One is you're showing kind of the amazing use cases of plugging the MCP in, delivering a tremendous amount of value. But it's not as clear to me how you might get directly paid or monetize that. And I think the other piece that's been really kind of an investment, maybe you talked about it, which is on the credits that you provide, you're not really enforcing people having to pay for those. And I think that's like, hey, we want to get people like to a certain level. We want the flywheel to get working before we want to crank that down. As you think about those two things, you're clearly adding value. What's the line of sight for you to participate in some of that financially, I guess?

Look, we obviously struggled to answer these questions by continuing to answer it. So I'll try another attempt at doing it. Teamwork Collection strength is due to AI. Like it's a huge driver of that, right? It's a very obvious customer path to upgrade to Teamwork Collection to get far more AI Rovo credits. It's not the only reason to upgrade to the Teamwork Collection, but we're trying to make sure for customers that they see it as a simplified package. So there is some amount of that cloud growth that we are seeing and increased revenue from Teamwork Collection in the AI story. How much, it's hard to pass because there's a number of reasons where we've moved to Teamwork Collection, but it's always in the top handful of reasons. Why? Also, as we've talked about many times, the upgrades of the cloud, increasingly, AI is this is 1 of the top -- 2 reasons that customers will move, right? And so that -- the economics of customers moving to cloud for us is extremely good over that multiyear period. So both of those are actually leading to a positive economic outcome that's a result of our AI investments. We obviously have an AI Rovo credit story in the longer term. That requires that usage, that requires that flywheel to be moving. We're seeing it moving, which is great. We want to be very careful to grow, very patient, looking at that long-term story and making sure that it comes through. There is a story there for sure, but we're trying to make sure that, that is the case. And again, we'd rather have customers consuming and using and seeing the value of the platform. We will learn about the platform and really early in that journey as an industry, I think, everywhere. Lastly, I would say, there's a sort of interesting competitive dynamic, right? The things that we're investing in building and the amount we're investing in R&D in AI, the platform we've built is very, very hard for us to replicate. There will not be 20, 30 SaaS vendors that can do this. They just won't. It's too hot. It's too expensive. It's too difficult. We're determined to be one of those handful of SaaS vendors that can do that. And I believe that's the long term of very good position for us to be in, to be able to do that. The stuff we're doing in the Teamwork Graph, this is nontrivial massive consumer scale stuff that very other few other vendors can do. Again, the video I showed of [ Kun ] is one of many examples where the Teamwork Graph can accelerate the customers. We have to continue to tell that story, get that out there and grow it and scale it, but we feel really good about where we sit at the moment.

And Mike touched on this a little bit earlier. But just maybe to put a finer point on it, TWC is the best AI monetization vehicle right now. And a big part of that is because of the credits that we're giving 10x credits relative to the stand-alone SKUs. But I think what we hear from customers is twofold. One is they want the access to, they also have within their enterprises, and I'm sure you guys see this within your organizations as well. There's lumpy usage. You got some hyper users, others that are provisioned but not using it. So they're able to load balance those credits across their organization as they're getting better and better at driving value from that. So it's giving them that access, it's giving them that predictability from an invoicing perspective. Then Mike shared a little bit earlier, all the different commercial models that we will adapt to as our customers are evolving as well. So I think we're seeing that. One of the things I don't want to get lost here is when we showed the video a little bit earlier there. I'm not sure if we ended up showing the stat, but in the keynote, we showed a stat that talked about how we were getting 44% of better quality in terms of output and 48% cheaper rate, right? You're using 48% fewer tokens to get to that same or better outcome. That's a conversation with your CIO. That's a conversation with your CFO that's going to resonate. And so I think both things are true. We're seeing that show up today in our monetization, and we're still building the platform in the flywheel to continue to compound on that.

And maybe one thing to add from a go-to-market perspective, obviously, our approach is quite different from others in the market in the sense that we've given our customers access to Rovo when both of them have said, get more access via TWC in terms of credits. That approach removes friction that might exist if we were positioning this in terms of how much it costs. So it's a way easier way to position it to our customers, first of all. And secondly, this is, as Mike had said, a more patient approach that we're taking. However, this is an approach that is -- historically, this is something that Atlassian has done before. However, the one difference compared to previously is we now are making considerable investments into our customer success organization in terms of helping those customers adopt even more so. And not only into our own customer success organization, but also into the partners as well to help them stand up similar adoption organizations. And a combination of all of those is where we are now seeing an increase in terms of our monthly active users.

Let's go to Jason on the front.

Thanks. It's actually a good segue to my question. So the side-by-side of the Claude example, and James, your comment about utilizing tokens more efficiently or cheaper. I think a lot of companies right now are still pretty early in maybe their token utilization. You see this means online of people making fun of how many tokens they're trying to deploy. Right? So if the value prop for players like Atlassian is to utilize tokens more efficiently, but the user is not quite asking that question yet, what needs to happen to get there? Or how long might it take for the tech community to rationalize their token usage mark?

I'm not sure most CFOs really like the token maxing meme, to be honest. I think they are starting to see and pay bills that this magical revenue pile is not coming from nowhere. So I do think that story resonates. I can give you multiple customer quotes already today that have been sent around some of the internal rooms and groups I'm on where customers are loving it. They're talking about it. They're already saying, "I'm installing Teamwork Graph," we read to get someone at work that did it like in this morning. And these are large-scale big customers. So I think that is a relevant story the cost side and token usage. I would not dismiss the other two, right? We've done a ton of work making sure that it's both higher quality and faster. So the example we showed finished a minute and change earlier. Again, that's one thing. But imagine that when you're running suddenly more and more parallel agents, you're doing more and more things and your workloads get more and more complex. The more complex this gets, the more saving you get. For these smaller trivial examples, the math's way, less good actually. But that's a good thing for us. If we believe the trend is towards more complexity, more complexity of agent use and agent work, we're heading into a better and better world as that happens. I believe that's absolutely the world that we're going to head into. But even if we just sit down with a customer and explain, we can make every AI agent you use, have a better quality answer due to the work we've done in your Teamwork Graph, the search that runs across that, again, those answers coming out of a search engine, kind of a search and graph combination like answer. It goes to reason that the better answers we can give the better information we give to any AI application, whether that's Figma or Claude Code, it doesn't matter, anything in between. The more you put into the Teamwork Graph, the better quality answer you will get. Even if you take out the cost story all together, that's a really compelling story to customers, especially if they already have a huge amount of data in the Teamwork Graph. They're going to connect that, that's going to write back. That's going to grow. So the quality in itself, let alone the speed, people still want the stuff to finish faster. So it's a really, really compelling pitch to customers. I expect that it lands very, very well. And if they're not having a token cost problem today, 3 months from now, 6 months from now, 12 months from now, they will probably have that in some point.

Yes. And ultimately, Jason, we're not in the business of selling tokens. We're in the business of selling value, right? And that customer journey that we walked you all through a little bit earlier in those slides where customers start with a couple of SKUs, they realize the value, they go to collections and then they go to multiple collections ultimately buy into the platform. And that reads and writes right back to that Teamwork Graph. So that is that compounding value ultimately. And if you're a customer and you're getting 40% better outcomes at 40% cheaper prices, you've got a competitive edge. And I think that's partly what we're seeing right now in terms of the consolidation and the conversations that we're seeing with some of our customers.

And maybe, Jason, just to round it out, I would only add that -- like it's top of mind for from the customers that I speak with every day in terms of the usage, not only the usage, but the return on investment. And when we look at DX, which obviously is operating in this space, 75% of the transactions that they are doing are associated with AI and the return on investment in terms of AI or agents, specifically. So it clearly is top of mind for our customers.

Why don't we go to one online that I thought was good from Thomas Blakey. Maybe good for Brian. Customer examples given were excellent in terms of continued growth focused on cloud and the power of one platform, noting the consistent growth in millions of dollars, sometimes $5 million or more in a year or so time is meaningful. Could management talk to penetration rate of these accounts at these spending levels and maybe cite where this spending is coming from, i.e., is it consolidating from other vendors or other IT spending pools? Brian, I know you and I talked about this the other day.

Yes. So obviously, we have considerable customer still that we need to move from DC to cloud, and we are continuing to make progress there. We have also established a considerable specialized sales organization. Now that specialized sales organization allows us to focus more in terms of selling into specific buyers into our customers. We historically have landed in IT, so tech. And from tech, we will expand into new buyers. So when we now look to sell to new buyers within the organization, we have to position ourselves differently to those particular personas. Therefore, we are selling into -- if it's marketing, if it's into HR, et cetera, we have persona-based selling that we have in place. So we are seeing and making progress there. Where from a collection perspective, obviously, Service Collection is where we are making considerable progress. We have -- as you saw in our Q3 results, we have made considerable progress in Service Collection. And then also, we have had seen a pickup in terms of our win rate against [ ServiceNow ], specifically and also some replacements as well in terms of our competitors, which is what we like to see. And we're hopefully expecting that that's going to continue in that vein.

Why don't we do Rob Oliver, right here.

Great. Thanks. Rob Oliver with Baird. Mike, I was struck. In the 2 charts you put up, the financial services company and the audio manufacturer, Jira Product Discovery was kind of right at the crux of the hockey stick there. And obviously, that's a very important part of the Teamwork Collection, but would love to hear how you guys think about that as a driver when I think about some of the most durable companies and software, they are those that are around system of record, system of truth, design. So thinking about that JPD opportunity for you guys as a component of the Teamwork Collection, how important is that as a land factor?

Look, there's no doubt, the newly grown Product Collection. So we have the Feedback app that got launched today. We have Jira Product Discovery and that Product Collection. It is a few things. Firstly, it's an excellent cross-sell opportunity to a Teamwork Collection customer or a classical Jira or Confluence customer because it is a very different way of looking at the data that you have and a very logical part of the workflow. So there's no doubt it is a good cross-sell opportunity. I would say that's more of a classical Atlassian business, right? It's still quite small, it's growing quite fast. We are making sure that the product is just freaking awesome before we start scaling it. I think we've only just added a premium edition 3, 6 months ago?

Yes.

Not long. We don't yet have an enterprise addition of Jira Product Discovery. So again, I want to make sure that it has that momentum, the seat growth, people are loving it first. And then we can start working out what we need in the premium enterprise that makes sense as the value grows. So classical cross-sell opportunity for us, a great team doing really well. Feedback is a really interesting addition, I think, because it moves us into some other spaces we have not classically been in beforehand. At the same time, it leverages our strengths in AI because I do think the volume of Feedback is going to go up, the product manager's job to take all of that feedback and distill it down and turn it into ideas that become the pipeline for engineering. If you want to think about it that way, you've got to first start with a set of insights and sort of scraps of text and other things from customers. You distill that into a set of ideas. Those ideas become road map items, and road map items get built by engineering. Engineering is -- as [indiscernible] ceasing to be the bottleneck in that process. So that creates other economic opportunities when you go outside of that. Now a product manager's job isn't going to be as classical. It's going to have a lot of AI and automation and other things associated with it, which is exactly what we're trying to do in the Feedback after show some of this at a large scale in a category where we don't have much footprint. So I think it's all upside for us, which is going really well. I think it is also a completion point because more and more marketers, other roles in the business are going to start looking at that feedback and working it out. And it somewhat completes a cycle. You saw a customer service management as one of the ingestion points for Feedback. Think about customer services, I think you do at the very end of a chain. You've built the thing, you've marketed it, you've sold it to someone, someone's used it, and then they have some challenger feedback that comes from customer service. Inevitably, it's a problem that they also add some feedback into and it kind of runs around the cycle. So it's us continuing to show that we have a system that can take those customer requests with a really deep AI and Teamwork Graph core. So that's a really interesting point for us, as well as a new point of data connectivity to some of those other systems we've shown, whether it's Pendo or passive sales force, other sources of Zendesk, whoever of access, all partners of ours to say, look, if you've got the connector to vendor ex installed, now look what the Feedback app can give you, look what the Product Collection can give you on top of the existing connector that you already have installed in your Teamwork Graph, right? So you don't need more data. You don't need more connectors. You have that already. It creates a higher multiplicative AI sort of ROI opportunity for that customer, and again, makes our graph more central to that customer. So super bullish on that area of the business, but I hope in multiple years' time, we're talking about the Product Collection heading in the same direction as the Service Collection is.

How about DJ, right here.

DJ Hynes from Canaccord. Two quick questions. I'll ask them concurrently. I think they're straightforward. Mike, you talked about, we're now at the point where you can start to moderate R&D spend. And it sounds like Brian's team will be a beneficiary of that. What does that look like in practical terms and kind of over what time frame are you talking about? And then the second question, some really interesting data and context connections to the Teamwork Graph today, right, code, assets, which is apparently your favorite skills. Are there any other major areas of data that still need to be connected to the graph that will kind of improve AI outcomes?

Two very different questions there. Look, the moderation of R&D spending is going to continue to happen over time. We've seen some of it already, and that will continue to happen over time. I think it's probably the best answer that we can give. I think what we're trying to explain is the why behind it, right? I really like to -- people to understand things from first principles. We need to build this enterprise platform. We need to build the AI infrastructure. We think those are all strategic long-term pieces for us. There are only so many enterprise controls and compliance. There is a limited number of those eventually that you get to. And we've built them into our platform. Again, our FedRAMP, our isolated cloud, commercial cloud, all run on the same code base, which is quite unusual. We don't want any engineer to be slowed down, and we continue to keep rolling those things out. Same as our multi-cloud offering, our data residency. There is an upfront cost to pay for all of those things. It needs to be maintained, it continue to build and evolve and all those things, but that is a different price over time than the price we've paid to get here, I guess, in terms of the smart teams. It also frees up engineering resources to be able to put towards some of those other things like the Teamwork Graph. So you'll see that coming through over time. Secondly, on the graph, I don't know if there's any major categories that we're looking at. Code has been a long-term investment. It's technically very challenging, both to index large volumes of code. Again, as we said, we're north of 1.5 billion lines, 11 million files. It's a nontrivial sized code index when you're looking at it semantically. And breaking it all down into the component parts, all different languages, everything else, that is a huge piece, but also because code underlies a lot of the other things that we want to do. Secondly, the people aspect was a big investment. The assets in the graph we've had for a long time, that's really taken the asset graph we already have, continue to scale it and connecting it into the Teamwork Graph, right? It's really connecting 2 graphs in the same graph language running across both of them and all sorts like that at the moment. That's been a sort of longer-term thing. People has been a new investment, right, as we continue to grow really strongly in the HR function broadly. So we have the Talent app now. HRSM continues to be a big aspect. People, chief people officers increasingly responsible for culture. So things like the Loom and other things become a big part of how they're communicating. Hubs and Confluence are big for HR usages. Again, we partner with all the major HRIS vendors, and we have a very symbiotic relationship with them, which is really helpful. So I think that's a big category, right? If we look at what we need to answer questions fulsomely and well better than anybody else in the world can answer said questions, it's about the people, the code and the physical objects connected to the sort of knowledge or the classical knowledge objects that I guess you have. So we feel pretty comfortable where we sit actually at the moment in terms of the major categories of things.

Adam Tindle, right here.

Adam Tindle from Raymond James. Really appreciate the presentation today. James, just wanted to ask for you, Mike mentioned a helpful comment expecting data center to decline next year. I wonder if you might put any parameters on the magnitude of that decline and the potential knock-on impact to be mindful of on margin and cash flow given the upfront rev rec? And the flip side of that question would be -- as we've talked to partners here today, customers are not looking to churn. There's no real viable alternative to Atlassian's products here. So if data center is going to be in decline, I wonder what that might mean cloud growth into next year?

Yes, I really think about this as sort of timing, right? To get a step back, Mike shared a little bit earlier, cloud remains really healthy for us. We're seeing retention rates as high as we ever had, in fact, actually outperforming expectations in Q3. So I think you hit on all the points there. As it relates to cloud -- or sorry, data center growth being negative next year. Again, this is more in that spirit of clarity. I think a lot of folks were sort of asking for that. We're obviously not giving out FY '27 guidance quite yet. We'll be more clear about that in August earnings, but obviously, you need to still see Q4 play out as well. But again, like the dynamics that Mike described a little bit earlier, I think was worth reinforcing, right, what we're seeing in terms of strength in cloud, strength in retention on the data center side. And then some of that customer behavior, right, as end of life is nearing, they're making decisions about what that more complex migration is going to look like, being able to pull up some of those commitments. Again, these are good things. These are customers that are committing to Atlassian whether it's through data center or through cloud. So we'll see that play out over time. We'll have a little bit more guidance, I think, in the August time frame for that.

Koji?

Koji Ikeda from Bank of America. Mike, during your presentation, I thought you had a pretty cool slide on your AI advantage with enterprises hard, Teamwork Graph and Jira as the orchestration letter as the 3 advantages. And so if we fast forward 3 years from now, what do you think will be, out of those 3, the most important advantage for Atlassian? And I asked the question because I think that inherently implies that might be the hardest to defend.

I mean, I suspect all 3 of those are important advantages. I think it's always important to remember the world is an incredibly dynamic place. So right now, our advantage is in AI. We want to make sure to magnify those advantages with customers in terms of getting that adoption. The flip side, maybe if the question asked earlier around, we'd much rather get that AI adoption right now while we have an enterprise advantage, going to harder for other people to invest in building it. And we know it's very hard and very difficult at the moment, right? So it's a dynamic world. You might argue some of those advantages relative over time, and we will find new ones. It's about us continuing to evolve. I don't know that I could say that any of them are more or less important. We love our Jira franchise. It's an amazing franchise. It continues to grow really strongly and we ask a lot to make sure that it is the case. At the same time, in the Agentic world, what has been done, who is doing what, to which piece of any workflow is -- needs to be tracked, managed and understood. I think if you want to think about it generically, I believe you'll have way more actors doing that, right? The idea that a 100-person company might have thousands of agents running around, that creates a lot more management overhead and understanding overhead, which I think is really good for orchestration platforms like a Jira that already have workflow and automation and all sorts of other capabilities to handle or manner of business workflows and processes. So I talked on a podcast and someone sort of -- I think 1 of the hosts explained to me backwards or I figure which way it went, but you're sort of talking about system of record and system of record. And that always sounds to me like a very static database-driven type of thing when you're a system of record. We were sort of -- maybe not even accused or pointed to being, I forget how it happened, to a system of process. And I'm like, that's a great description of what it is that we actually do, right? These dynamic business processes, you've heard from both Cisco, Canal+. You heard from customers today, we talked about Mercedes-Benz yesterday on stage. We showed another handful 2, 3 customer examples between Brian and I that aren't those other 3 customers. They're all running processes through Atlassian, right? And those processes that they have, 10 years of data center and server history that move to the cloud, they only grow the number of processes they run through us. Agents are going to yield many more processes and much more of that difficulty. So I think Jira is going to continue to have a really strong place, especially if we can give all those tools to allow people to manage and understand that, which we keep trying to do. There's no doubt that Teamwork Graph has amazing potential, right? As customers use it as it grows, we're seeing it grow very rapidly, create some fascinating and interesting engineering challenges for our team to grow it at the scale that it's growing. But again, things that are hard to do and hard to scale and hard to deliver on are great because it's very hard for other people to do that, and we have a distinctive customer and data advantage already.

Keith Bachman?

Good afternoon. Thanks for doing this. I'll ask two concurrent questions in the interest of time. Mike, just to pick up on that thread. When I walk the show floor, it just seems like Teamwork Graph is one of the key competitive advantages that Atlassian could have over multiple years. And so the inverse of that, though, what is the risk you see from a competitive dynamics? In other words, if somebody -- when your Atlassian customers is not deploying Teamwork Graph, what else might they be using? Or might they just be using a stand-alone cloud or something along those lines? Which you've shown and demonstrated it's not as efficient, but just curious on the competitive dynamics that you see on Teamwork Graph. In other words, what somebody might choose to use instead? And then just to ask a second question, Brian, you indicated you had 600 -- or excuse me, 400 sales reps or quota-carrying reps. I'm not sure how you refer to them. It's still very subscale, right, on almost every respect. And your -- 10% of your revenues are Fortune 500. So -- and you're moving to collections and these higher value-add sales. How should we think about the ramp of sales as we look at '27 when you're trying to, I think, get more revenue, greater revenue from existing customers?

I guess I'll take the first one. The Teamwork Graph competitive question is an interesting one, actually. I think -- look, firstly, customers don't have to buy the Teamwork Graph, they get the Teamwork Graph. We've made it a core part of the platform. There's a reason that we don't sort of separately charge for it or sell it or any of this sort of stuff, which always makes it harder from an investor point of view to communicate that. You'd love to see a separate line and that's going up into the right, and it's looking really, really great. But from a long-term strategic point of view, I don't think that's the right thing for us to do. What we want is anybody using Jira, anybody who's using Confluence, anybody using Loom, anybody using Bitbucket to have the Teamwork Graph. And the more of those things they're using, the better their graph gets, the compounding exponential effects, right, [ Metcalf ] lore of context graphs. I'm not sure if they can kind of take 2 different eras of technology and put them together there. But that is a really powerful thing because it means that the next collection they get, the next app they add, that graph gets more powerful. The next connected they add, it gets more powerful. Now we have a lot of connectors already because we've had various graphs for -- the reason we built the graph 8 years ago was nothing to do with AI. It didn't really exist. It's machine learning. It was a very different world. It's to do with us connecting to other customers' applications and them wanting to share data and us continually managing the connectivity of that data. In such, it doesn't really have a competitor in and of itself. You know you need to buy it as a customer. What we're trying to make sure that customers understand is the more Atlassian products they use, the more connectivity, the more their data grows and more processes and workflows they run, that graph is just going to keep growing in value. Now you don't need to connect only the graph to Claude Code in that example. You can connect the graph and other -- something else, great. It's up to Claude Code to determine where it's going to get the better answer. Now in the long term, it's a really interesting question, whether the agents themselves will learn that we are a great place to get great answers. I suspect they will. And that's a great thing for us. We're just going to make sure that, that is the best graph that you can get off the type of data that we want to serve. Now we have a lot of partners that do different types of data, and that's great. If you go look at Databricks or Snowflake or someone doing sort of relational numerical data at massive scale, that's not what we're trying to do in the Teamwork Graph. And so you're going to have multiple of these types of knowledge graphs, data graphs, whatever you call them. Most of your agents are going to have a number of places they can go and get information to help them. We want to make sure that we're building the best for our use cases, I guess, our competitive vector, and make sure that customers understand and that we can bring all of that past history into the graph. So today, you can have 2 customers do it already, download the CLI and be like, holy c***, look at all the stuff I have. I'm like, that's the exact reaction that I want them to have. And then I'd say, "Great, just go plug that into Claude Code and watch it get better." That will -- I suspect lead to the graph continuing to grow and then getting more and more value from us and them seeing us as a more valuable partner to them, which is an amazing position for us to be in over the next few years. And we can continue to invest in the R&D to make that both scale and the smarts.

And as you said, we have 400 quota-carrying reps. And we certainly punch above our weight when you look at the revenue. Over the course of this year, we have increased our reps. And the way we look at it is we prioritize both from a geographical perspective, and also from an account to a sales rep perspective to look at where we'll be the highest return on our investment. As we look at FY '27, as Mike had said, there will be additional investments made into my team, and we will continue that approach of ensuring where we will make those investments from a return on investment to ensure that we're going to maximize the return that we will be getting for the company. That includes from a geographical perspective, entering into new markets where we do not have people. So just earlier this year, we put our first reps on the ground in India, where we did not have any reps before. In addition to that, we're also going to be leveraging the ecosystem. So there will be certain parts of the world where we identify that we have an opportunity, but we won't be going in there yet. So we've already leaned on our partners to ask them to come into those markets with us where they are doing that. And they will serve as an extension of our sales force, and we will continue to make investments and look at it prudently as we move forward.

That's actually all the time that we have today for questions. So I really appreciate everyone for coming out, learning more from us, talking to our customers, learning about our announcements. And for those that tuned into the webcast, thank you for joining.

Thank you, all.