The Exchanges

Q Did you love the problem of JIRA? You said there about kind of working on software and bouncing around different products to find what works. Did you love the problem of JIRA? I'm always trying to understand whether you do have to focus on something you love.

A Look, I'm a certain believer that Two things. One is that you've got to love what you do, uh, otherwise someone else out there who's just as smart as I am, who is more passionate about what I'm doing, will beat me, like, every single day of the week, so I'm a firm believer you need to be passionate about what you do. At the same time, I think there's an aspect of you, when you get good at something, well, you learn to love it more, and, uh, I think you need to, you know, try something on for size, and so, for me specifically, I've always loved building software. Like, I love the idea that you can create something out of nothing, Uh, write some code and it, you know, controls this thing called a computer. So I've always loved that. And, you know, if you said, Hey, would you, you know, is bug tracking like your, what your life is going to be, you know, like for the next 20 years, you know, that's what, which is what Gio originally started with. I don't think I would have said, yes, that's my last calling. But when I look at what we've done around, you know, turning that, you know, initial starting point into a great company that people will have to work at. And our mission now is to unleash the potential of every team. And so it's, Super jazzed about that mission.

AI assessment note: “is bug tracking like what your life is going to be... I don't think I would have said, yes”

Q Is it difficult with children? I have David Velez on the show from New Bank, and he said his number one concern right now is bringing children up in a world of Huge wealth and retaining humility and ambition and hunger.

A I think it is a very difficult thing to do. I don't think I've cracked the code on that. You know, even if we try and isolate our children from that, um, you know, I bought an expensive house in, in Sydney and, uh, my kids, friends at school, talk to them about it, how much it costs. And so that, like, no matter how much you try and isolate your kids from that, it can, uh, be, you know, something that you just don't have the opportunity to do that. And so for me, it's just trying to teach kids The value of money, and like, um, and also just being a good human being. I definitely haven't cracked that, and my kids are still young, so I don't know if I've done a good job for another decade or so.

AI assessment note: “I think it is a very difficult thing to do.”

Q Is Atlassian able to be fast enough to move with the speed of AI changing today? Just out of respect to stature, public company status, size of team, Are you able to move fast enough?

A There's some interesting things there. Look, obviously, you know, as small, the smaller the company is, the faster you can move. There's a saying that, like, if you want to go fast, go alone. If you want to go far, go together. And I think that's the difference between small companies and big companies. Like, if you want to go far, if you want to have a big impact, you know, we have a whole bunch of advantages that we can do when we run AI experiments. We can run them with 250,000 customers. Like, we want to find, hey, how can this work with a, you know, a search problem work? We've got customers with 20 years worth of data. You know, and terabytes worth of data that we can, you know, work with them, uh, to run experiments on. And so, yes, we've got, you know, more overhead than a normal company or a small company would have. But if you look at, you know, can we move fast? Yes, we have these other advantages that allow us to move fast, particularly in the AI area where it's about data. If you're an AI startup and you don't have access to customer data and customers to test with, that can actually make you move slower.

AI assessment note: “can we move fast? Yes, we have these other advantages that allow us to move fast”

Q Is it difficult with children? I have David Velez on the show from New Bank, and he said his number one concern right now is bringing children up in a world of Huge wealth and retaining humility and ambition and hunger.

A I think it is a very difficult thing to do. I don't think I've cracked the code on that. You know, even if we try and isolate our children from that, um, you know, I bought an expensive house in, in Sydney and, uh, my kids, friends at school, talk to them about it, how much it costs. And so that, like, no matter how much you try and isolate your kids from that, it can, uh, be, you know, something that you just don't have the opportunity to do that. And so for me, it's just trying to teach kids The value of money, and like, um, and also just being a good human being. I definitely haven't cracked that, and my kids are still young, so I don't know if I've done a good job for another decade or so.

AI assessment note: “I think it is a very difficult thing to do. I don't think I've cracked”

Q To what extent do you aggressively allocate towards AI today versus appreciating we're still in phase one. A lot of the money will be burnt in, in, you know, productive ways, but still burnt because you need to burn it for progression to happen often. So what extent do you aggressively allocate versus do, but tentatively in the knowledge that we're still in phase one?

A We have, um, you know, multiple thousands of engineers in Atlassian and we have Multiple hundreds of engineers working on, you know, AI, uh, you know, things at the moment. What we do, if you look at, um, our data, our data is our customers' data. Like, we don't control our data. We're not like Bloomberg where it's like we have, you know, 50 years worth of trades that we can, you know, put into a large language model. Like, we have our customers' data and we need to help our customers access it better. So we don't have to make the huge list and investments of, uh, you know, building our own large language model or building, you know, our own thing across So across, you know, data sets. So what we do need to do is integrate them in interesting new ways. And so I think that's not wasted work because, you know, if we can work out what the most useful way of doing that is, I think that's going to be, um, you know, time well spent. And yes, you could wait, I guess, till that paradigm has been, you know, sorted out that, uh, you know, I think by the time like Uber worked out the exact way to order a car and that it was cars and not limos and stuff like that. It was hard for other people to catch up because they had an advantage and it's not just understanding what works. It's often understanding what doesn't work. Um, and, uh, you can't really buy your way into knowing what doesn't w…

AI assessment note: “Multiple hundreds of engineers working on, you know, AI”

Q remain full remote. When I spoke to Mike, he said the biggest risk decisions that have been a success have been made by Scott, not by me. When you reflect on, I thought it was very humble of him, but when you reflect on like the biggest risky decisions that you've made that have worked out, which one is the biggest for you and what did you learn from that?

A There's a couple of ones that we've made that are risky. I think Team Anywhere is obviously Huge one because, you know, it's not always a one-way door, but like we would burn a whole bunch of, you know, employee trust if we changed that now. We've got about half our employees don't leave near an office. So like it's pretty close to a one-way door decision as you can make. And, uh, we made that relatively quickly. And so that was one. Look, another decision we made was to shut down Stride, which was our competitor to Slack. That was a product that we had been, you know, operating the market for multiple years. And, uh, Um, if I could share some of the lessons of things that we, you know, made mistakes on, we acquired this product, it was growing really fast, it was used by us, it was used by devs, and then it shared, you know, went through the entire company, and there's a couple of things that I would do, have done differently with that. Um, one is that, uh, Stride was two dollars a month per customer, like, which is cheap, like, enterprise software is two dollars a month, a user a month, that's incredible, but Slack was free and 10 dollars a month. And free is a lot cheaper than two dollars a month, even though on average, Slack made more, more dollars from us because they eventually converted people. And so, you know, the freemium model basically, um, you know, can win over t…

AI assessment note: “There's a couple of ones that we've made that are risky. I think Team Anywhere”

Q And how much were you going out asking for in the race?

A So when we did this, we, we told people that it's one shot caused envelope bidding, because I believe that was the best way to get a result, because the way most VCs do it is they give you an indicative offer, and then they sort of talk amongst themselves, and then they kind of agree on the price, and there's a bit of a collision that can often happen there, and They sort of do a price that's enough to get in the door, and then, you know, they get a second chance to up their number, and I felt that that was not the right way to do it. It was not enough game theory involved in that, and I tried to fight with my CSO at the time about this, because he said it's not the way these things are done, and I said I don't care about the way things are done. So we actually got closed envelope bids. We said, hey, we want to raise, you know, I don't know, you tell us how much we raise, you know, tens of millions of dollars, but you tell us how much we want you put in. And then you tell us what the valuation was. So we're on sixty million dollars of revenue. I think we were doing 20 or 30% profit margins. So we're making decent money at the bottom line. And, uh, the valuations we got back varied between, there was one sort of high hundreds. There was three of them ended up in the two to three hundred million dollar range. And Excel, who ended up, uh, you know, winning for us, uh, came back at…

AI assessment note: “you tell us how much we raise, you know, tens of millions of dollars”

Q this as a leader, say, with AI? Like, bluntly, as a public company, everyone's told you have You have to have an AI story. You have to have an AI narrative. I get it, but you're also a very successful sustaining business with great customers already. How do you think about that need to have an AI narrative, an AI story, but also just remembering the amazing business we have?

A The history of technology with the way I would write it is that, um, it's a very much a winner takes all market. And once there's a winner established in a market, it's hard to disrupt them except when there's a technological change. And if I go back through the ways of change, I would say we went from minicomputers to, or mainframe to minicomputers, which was before my time, then we went to desktop PCs in the nineties, and you saw, you know, you see companies emerge and see companies decline, right? Like you saw, you know, DEC, DEC, and a whole bunch of, you know, minicomputer people didn't make it to the PC, and instead you saw sort of IBM and Microsoft rise, and then the internet came around, and you saw Microsoft wane a little bit, you know, and you saw a rise of net Escape and Google and eBay and PayPal and so forth. And then mobile came around and didn't really displace anyone, but Apple suddenly came back into the floor and Microsoft continued to sort of miss that, that, uh, that area. And then you saw, I guess, cloud, public cloud, which, you know, changed a lot of enterprise stuff, but not much for consumer. They didn't care what, where, where things were running. Um, and now we're seeing AI. And I think again, in each of these changes, you see some companies continue to survive. Some companies, you know, rise from the ashes, like, um, and some companies don't make the…

AI assessment note: “as much as it's the next greatest buzzword... I don't think you can avoid”