From Startup to Exit
Welcome to the Startup to Exit podcast where we bring you world-class entrepreneurs and VCs to share their hard-earned success stories and secrets. This podcast has been brought to you by TiE Seattle. TiE is a global non-profit that focuses on fostering entrepreneurship. TiE Seattle offers a range of programs including the GoVertical Startup Creation Weekend, TiE Entrepreneur Institute, and the TiE Seattle Angel Network. We encourage you to become a TiE member so you can gain access to these great programs. To become a member, please visit www.Seattle.tie.org.
From Startup to Exit
Gen AI Series: Developer Joy through Generative AI. A conversation with Rajeev Rajan, CTO at Atlassian
Rajeev embarked on his career at Microsoft as a developer after a sabbatical from his Ph.D. to build upon his internship experience in Seattle the previous summer. In this episode, he delves into how his journey and interactions with various leaders have influenced his own leadership style as the CTO of Atlassian.
The theme of ‘Developer Joy’ has resonated with Rajeev throughout his tenure at Microsoft, META, and now Atlassian. He elaborates on how Generative AI is being integrated across all Atlassian products, emphasizing this core theme. Given his background as a developer early in his career, this theme holds a special place in his heart. In this episode Rajeev outlines how Generative AI will shape tools for developers and users, particularly as Atlassian expands its collaboration offerings to all corners of the enterprise
Rajeev Rajan is the Chief Technology Officer at Atlassian where he oversees Atlassian Engineering, IT, Security and Trust, and the Engineering Operations teams. As CTO, he is focused on our Cloud transformation and continuing to grow Atlassian into a world-class engineering organization.
Rajeev previously acted as the Vice President and Head of Engineering for Meta and spent over two decades at Microsoft, where he held a number of roles, including building and leading the team responsible for Office 365's Cloud Infrastructure.
Brought to you by TiE Seattle
Hosts: Shirish Nadkarni and Gowri Shankar
Producers: Minee Verma and Eesha Jain
YouTube Channel: https://www.youtube.com/@fromstartuptoexitpodcast
Welcome to the Start of Texas Podcast will be bringing work class entrepreneurs and easy hardware and speaker. TI is also not offered to focuses on offer entrepreneurs. I see Add offered the First of Programs and Costco. I entrepreneur institute. We encourage you to become a tie member. You can gain access to this great podcast. Become a member visit www.seattle.ti.org.
SPEAKER_03:Welcome to another episode of From Startup to Exit, uh, a podcast that I co-host with uh my friend Shirish Natkarni. Uh my name is Gowry Shankar. I'm a member of the Thai Seattle Board and uh serial entrepreneur based in Seattle. Uh Sherish uh is also a serial author, in addition to being a serial entrepreneur, and serves on the board with me um at Thai Seattle. This episode, in our continuation of our Gen AI series, is very special to me. Uh our guest today is uh a very close personal friend, and I admire him professionally. It's going to be a fun episode for me personally, and I think it'll be very insightful for our audience. Um thanks to everybody for uh subscribing. Please share uh the link with everybody you think will enjoy this episode, and we are available everywhere podcasts are uh can be heard and seen. Uh with that, uh uh let me hand it over to Sharish. There are two books Sharish have written. Sharish's first book from Startup to Exit, from which we derived our name, and then the second one, Winner Take All about Marketplaces. Both the books are now out wherever books are sold. So hope you buy them, read them, and enjoy them as I did. Sharish, take it away.
SPEAKER_04:Thank you, Gauri. Welcome to our podcast. Uh, today we are very pleased to have with us Rajiv Rajan, who's the CTO of Atlassian. Uh, before that, he was um an executive at Microsoft and then at Facebook slash Meta. Uh, he joined Atlassian a few years ago. Uh, and we're very pleased to have him today on our podcast. So, welcome Rajiv.
SPEAKER_01:Yeah, thanks, Sharish, and thanks, Gaudi. This is uh extra special. As Gauri mentioned, we're we are good friends and go back a long way. And I've also been at uh several Thai Seattle events, and so really excited to do this with um Sharish with you and with Gaurie. Okay, great.
SPEAKER_04:So let's start a little bit with your background. Uh, as we discussed, uh, you've had a long and storied journey. So if you can take us through that journey and how you line it up at uh at least, that would be great.
SPEAKER_01:Sure, yeah, yeah. Um look, I was born and raised in India um and I did my undergrad in in computer science there. Um landed up in the US, I did my master's at Ohio State University, Columbus, Ohio. Um and then actually I transferred to the University of Illinois at Urbana Champaign to do uh I wanted to do a PhD. And so um sort of because my my dad did his PhD back in the day, so I had this bug in me to do it. Uh but one fine day, you know, I went to get some pizza, free pizza. It turns out Microsoft was on campus and and they were serving out free pizza, and so one thing led to another, and I ended up doing an internship in Seattle at Microsoft, and um it's an amazing experience. And um I took a leave of absence from the PhD program, never went back, and uh spent a couple of decades at Microsoft doing many things. The first few years I was uh sort of more in the lower levels of the stack. I wrote code in the kernel and and you know, exchange server and SQL Server and things like that. Um then I moved to a group that eventually became uh Office 365. And over time I was able to build and lead essentially the foundational team that built Office 365 and took it to the cloud. Um, something that's been a great success for Microsoft, you know. And I was really happy to be part of the sort of resurgence of Microsoft, you know, when I did that. Um after that, after like a really long innings at Microsoft, I moved to Facebook to try something new and different, um, moved into product teams and ended up sort of leading the engineering team for the Facebook app, uh the Blue app, as it's called, um, as well as sort of leading Mera's offices in Seattle and the Pacific Northwest. And so was there for about five years. And then um one thing led to another, and the CTO opportunity in Atlassian opened up. And so I joined uh Atlassian a couple of years ago. Having been at big companies and seen some transformations, um, it was exciting to see a company like Atlassian that is well loved and well known for what it does, but wanting to do a lot more in the next few years. And uh they wanted someone like me who would come and help scale and build the engineering team and so um joined Atlassian. Uh, on the personal note, I'm uh based in Kirkland, Washington. I've been there for a long time in terms of my family and everything else, and um just been in the Pacific Northwest in Seattle for forever. That's great.
SPEAKER_04:So let's spend a little bit of time talking about your journey at Microsoft. Uh, one of your big projects was uh Office 365 and moving Office to the cloud. So uh it'd be great if you can talk a little bit about some of the challenges that you faced and uh what you had to overcome to uh you know uh you know what we know today of as Office 365.
SPEAKER_01:Sure, yeah. Look, um, at the time that we started off, you know, uh there was also not a lot of um internal alignment, let's say, you know, Microsoft is transitioning from a very successful and profitable from a gross margin standpoint, you know, shrink-wrapped software business to to cloud, and and the dynamics and the business model is a little different. So in the beginning, there was not even a lot of agreement within Microsoft on on what to do. But rapidly, cloud became a top priority, and I was fortunate to join uh an amazing group of people who decided to take on this challenge. Um, a lot of people thought Google and Gmail would would win that battle, and uh, you know, we surprised everybody. We showed that you know Microsoft um had it in in them at that point in time to take on the challenge and um you know actually transform the company from you know enterprise desktop software to to cloud software. So we had to do a number of different things. We had to, engineers had to learn how to get on call and you know take a take a call at 2 a.m. in the morning if your code doesn't work and uh things like that. So it was a cultural change. It was definitely um a big challenge in terms of taking some pieces of software that were pretty legacy and making that scale in the cloud. You know, we had to literally go and rewrite some old implementations to have it scale differently in the cloud and things like that. So all across the board, from culture to technology to how you operate, you name it. We had to go change things, and um, it was a fun, fun ride.
SPEAKER_04:Got it. So then you uh went on to join uh Facebook, and as you said, you were responsible for the main uh Facebook uh app. Uh uh I'm sure at that time you were infusing AI into Facebook, uh, you know, to um, you know, especially with the uh the feed, uh making sure that you know the most attractive content was surfacing for every individual. So maybe you can if you can talk a little bit about that in terms of especially in terms of how you were using AI as part of Facebook.
SPEAKER_01:Yeah, yeah. Look, um, it was really uh you know a huge, huge uh learning when I joined Facebook. You know, the scale, first of all, the scale is massive, you know, in terms of the number of users and so on uh for all the apps, Facebook app, Instagram, WhatsApp, etc. So like that was really, even though I had seen huge scale from an infrastructure standpoint in in Office 365 and Microsoft, the consumer scale that Facebook has is really, really amazing. Uh but what was also amazing was the use of AI in all parts of the product. You know, everything uh the news feed is basically a ranked, it's a huge ML algorithm that ranks the feed for each user personalized. Um so obviously heavily used across the board. Um and um you know in all different kinds of places, including in terms of which ad gets shown to you and so on, obviously highly, highly personalized machine learning algorithms. So yeah, Facebook I think has been using that for a long time, and I was using it as well when while we were there. Um and the same thing with Atlassian, like even Atlassian has been using AI ML for a long time. If you look at our products, we've been using you know AI ML to uh do different things in our in Confluence and Jira and so on. So whether it's Facebook or Atlassian or different companies, everyone's been using AI ML for years. What changed with Gen AI, I think, is that LLMs came along, and we finally were able to train a model to understand the English language and to be able to do amazing things with languages and text. And so I think that was a big sort of shift that happened when Gen AI came. And um, everyone obviously is onto it now in terms of what you can do with that.
SPEAKER_04:Yeah, and we'll get to uh have discussion around that uh a little bit later. Um, so um, you know, at that time when you were at Facebook, uh, you know, Facebook was getting a lot of criticism around uh, you know, surfacing stuff that was not relevant or you know, harmful to individuals or uh teens and so forth. Um how did you feel about that?
SPEAKER_01:Yeah, look, um, you know, uh it's always tricky when you're you're dealing with a new technology, right? Newsfeed in many ways is a new technology. Social media was a new technology. Um you know, I'm sure when uh television first came and radio first came, there were things on it that you know people were wrestling with in terms of you know types of content and so on. Um I think like Facebook, like every other company, tried um through AI, through ML, through uh community integrity and things like that uh to fight the battle in terms of making sure that the content on the platform is good and helpful and that sort of thing. So I'm not going to get too much into sort of like you know the the pros and cons of of those kinds of things, but I can tell you from an engineering standpoint, there was a massive effort using AI and ML to manage the kinds of content that came in and to make sure that you know people got positive content.
SPEAKER_04:Right, got it. Okay. So after you said four to five years um uh for the opportunity uh for Atlasian uh CTO position, um what made you consider that opportunity?
SPEAKER_01:Right. So I've always been an engineer. I in fact, first few years I did a lot of work as an IC engineer, um, and then sort of, you know, I accidentally got into management, if you will. Like in fact, the first time uh I was I was doing uh performance assessments, I didn't even know I was a manager. My manager comes to me and says, Hey, you're doing performance assessments, and that was a little bit of a, you know, I didn't really realize I had gotten into engineering management, but I try to stay an IC and sort of like not be a manager for a while until until I found that um, you know, you get a lot more ability to take projects forward, and I started to enjoy the coaching aspect of it as well. And so those things led me eventually into engineering management roles. Um and so you know that was sort of a an evolution of things um you know that led me there. Correct.
SPEAKER_04:I'm sure that you were speaking to the two CEOs of Atlassian uh when you interviewed with them. Uh what were they looking for?
SPEAKER_01:Yeah, so you know, um in terms of Atlassian, like in terms, like I said, I was I was I've been an engineer all along, done a number of engineering leadership roles. Um and so I care a lot about the engineering function and engineers in general, right? And the Atlassian leaders, the the CEOs, we have two, um, we're looking for someone who could come and take Atlassian to the next level. So Atlassian is a fairly well-known company. We we we produce Jira, which is what we are known for, but we have other products like Confluence and such. And uh, Jira is actually a very well-known name with respect to technical teams. So any any software team, any technical team uses Jira for issue management, bug management, things like that. Right. Um, but the company's aspiration goes beyond that, right? So Atlassian wants to essentially do make it so that you can use Jira for non-technical teams as well. So an HR team, a finance team, legal team, etc. Because at its heart, Jira is a project management, issue management software. And people in any department, any company need to manage projects, right? If you're building a rocket that you want to put it to Mars, like you you want to be able to track issues and projects and things like that. So the company's aspiration is to essentially unleash the power of teams. You know, Jira is one piece of software to do that, but we have a whole what we call system of work um set of solutions that you can use to manage work in any company, any any technology, any vertical, you can use that. And so if you uh if you achieve our aspiration, really Atlassian would be something that would be used by all companies, small, medium, large enterprises, wall-to-wall, to get work done, to do knowledge worker work, right? So that's a big mission. And you know, Atlassian therefore has the potential to be in your top tech company. And to do that, you need world-class engineering. So, what Atlassian was looking for was can we have someone come in who can lead engineering as a CTO, take it to the next level in terms of world-class engineering, someone who's been in places where you know they've seen world-class engineering, and that was definitely true with me. And so that was the pull for Atlassian to uh have me come take the role. And for me, the attraction was to be able to come and lead all of engineering, the whole function. As I told you, I'm passionate about engineers and engineering, and to be able to have a vision for where I want to take engineering, and so to be able to come and lead the whole function and help the whole company go towards the next level of growth was uh super attractive to me. Right, right.
SPEAKER_04:So um uh what would you describe your mission as a CTO to be? I've talked I've heard you talk about uh developer joy.
SPEAKER_01:Um is that the key to your mission? Well, I I would say the key to my mission is is really how do I build a world-class engineering team, right? And so there are many elements that go into world-class engineering. You know, I actually break it down into three pillars. You know, the first is whatever software you build, better have trust with customers. And by that, what I mean is it should be reliable, right? We are we build cloud software, so you want it to be four nines, you don't want it to go down. So reliability is important. Trust with your customers with respect to security is important. You know, customer data is super important, so you want to make sure you're a good guardian of customer data. Um, you want to make sure if customers have problems that they get support quickly and the product is supportable. So those are all things that go into customer trust. The second pillar for world-class engineering is you need awesome people, you need great engineers, you know, you need 10x engineers. Um, and so how do you find the best engineering talent? How do you make sure that your existing engineers in your team are learning the right skills to be world-class engineers, right? So it turns out, even though you're I'm here as a CTO, it's not that I you know only solve architectural technical problems. I'm also building the talent in the organization so that we can have a world-class engineering team and be known for that, right? And that helps us build amazing products. That's the second pillar. And the third pillar is what I call developer joy, which is has many components to it, one of which is developer productivity. But let me explain to you what developer joy is. All of us who are engineers who you know joined, who learned how to write code, wrote our first program, hello world program, you know, saw the did a little printf or s or some kind of thing and saw the text appear on the screen, you know, that excitement that you felt, that joy that you felt that you could go write something on a keyboard and it and something happens on the computer, right? That that's the joy we want in all our engineers. That's the joy we want in engineers in general. Now, as you know, when you actually write code, you end up uh with a lot of frustrations because you have bugs and you know the code doesn't compile or something doesn't work well, right? And what happens is you end up with like, or you're writing code in the at 2 a.m. and suddenly you need to use a different component, but you don't understand the APIs and you're blocked, right? So the thing that really upsets engineers or like comes in the way of developer joy is when you get blocked and you can't make progress, or you need to like find another person at a different time zone. And so one of the things we've done in Atlantic is how do we make sure that that flow of writing code, that flow of creativity doesn't get blocked? And we have done a few things to help engineers unleash that joy. In fact, one of the tools we have built that we sell as a product is called Compass. And Compass is a place, it's sort of like a registry where we have all our components and microservices and so on. So if you are that developer at 2 a.m. and you get blocked, you can go to Compass and find the self-serve APIs of a component that you're trying to use and unblock yourself and keep going. So developer joy to us is about the joy of writing code, the flow of creativity where you don't get blocked and you're able to do the thing that you want to do, which is write code. And so developer productivity was a side effect of that. If you do the thing for developer joy, then your developers are productive. If developers are productive, you're happy. You have happy engineers, you get amazing products. You know, so that's that's a that's the third pillar. Got it, got it.
SPEAKER_04:So we'll talk a little bit uh about AI in your product line, but let's talk about AI as it's been utilized uh to increase developer joy. Um, you know, have you used tools like uh GitHub Copilot or you know, how are you using AI in your tool set to make your developers more productive?
SPEAKER_01:Yeah, so like everyone else, you know, we are experimenting with a bunch of things. You know, I think generally speaking, when you think about coding assistance and AI for developers, we are still very early in that journey, right? So everyone is experimenting with different things, and we are too. So we are using coding coding assistants, and we find some of these coding assistants can can save you know one to two hours per week per engineer, which is quite a decent amount of savings. And so it is definitely something that we are using and playing with. And uh, in many cases, it is giving us good productivity wins and gains. I wouldn't say it is like nonlinear, it's not like, oh, suddenly I can do, you know, cut down my I can increase productivity 100% to 100%, but an hour or two a week is is decent. Um and it's more in the realm of a teammate or like uh an assistant who does a few things for you or maybe solves the blank slate problem, right? You open an editor or you open a uh you know file to write some code, and if you can do a few things and it writes some things for you, you're past the blank slate problem, right? And then you can start writing more code and you can get in the flow of things, right? So that's where we have found it to be helpful. Uh, but we are also working on things ourselves that we think because we are a number of our customers with Jira are engineers. You know, Jira is used to open a bug or an issue, and then you have to go and try and fix that issue, right? So we are working on technology where we're using AI to go and try to convert an issue into a pull request so that you know if we have enough information in the Jira issue and we are able to convert that into like a pull request, then an engineer can go and review that code. They may have to make some few changes and things, but at least they get a head start on going and fixing the Jira issue, right? So we are focusing on use cases that tie into the products that we make for engineers. Got it. Okay.
SPEAKER_04:Uh one final question before I turn it over to Gauri is um, you know, you've had the opportunity to see different CEOs in action. You know, you were there when Bill Gates was a CEO, and then Steve Bomber, and then Satya, and then with Mark Zuckerberg, and then now with the two CEOs at uh at Lishin. Um how do you characterize their leadership style and how has that affected you?
SPEAKER_01:Yeah, look, um, anyone who's like a founder CEO, you know, they have you know unique energy and charisma. And obviously, you know, they built great companies, including you know, Mike and Scott, who are the two co-CEOs for Atlassian. You know, that's a it's a very great story of how they started Atlassian, you know, in Australia 20 years ago, um, and and built it from, you know, building a company outside Silicon Valley, as you know, is tough, but building it all the way in Australia. And then today uh Atlassian has 300,000 customers, you know, across the spectrum of all businesses, it's a pretty amazing story in and of itself. And so obviously you learn from each of them different things. Um, I guess one of the things that's sort of a common thread across all of this is it's really all about people. Like, you know, talent is is is key here in uh technology, right? And so finding uh 10x engineers who can do amazing things and being able to tap into great people to go and do things that nobody thought was possible, um, I think is a is a common thread, right? You know, with with all of these leaders, you know, you'll find uh, you know, they'll they'll push you to do things that you know you didn't think you could do or that you didn't you didn't think was possible. And I think that's Something that's also a founder, CEO, you know, trait that's that's important. Um, the other thing is you know, creating an environment where amazing people can come in and do their best work and you know having a learning mindset, having a growth mindset. I think those are some things that I find comes through, and that's that's an important learning. Um the other thing is really having a customer focus and a mindset around shipping, because you know, you could end up in many situations where you you try to sort of build a multi-year plan and it goes nowhere. I've seen those projects too, and you know, sometimes, you know, sometimes some of the some of some folks can get into that um area too, where you know you dream too big and it takes too many years and then it doesn't ship and such, right? So you some of the learnings for me also has been to be focused on shipping, you know, get shit done, like GSD is very important. Otherwise, you know, you can be on a project for a long time. And it's never fun to be on a long project that never ships. I've I've had some of those movies as well. Um, the other thing is really being um principled but pragmatic. You know, I one of the lessons I learned from one of the leaders I worked for is you know, have your head in the cloud, but right? Um if you if you dream too much and you don't ship it, that's not good. Uh so you you do need to dream, but you need to ship and get it out there, get it to customers, get the feedback and iterate. Otherwise, you can go for too long without shipping anything. So those those are some sort of lessons you know I picked up. Uh, and then one last thing I'll say is, you know, there's a lot of lessons you pick up around scale leadership, right? Because when you're running a big company or running a big organization, I've learned a lot about how to scale your leadership. You know, at Atlasia and I lead a team of about 7,000 people. Um, it's a big team. So, how do you scale your leadership? How do you connect with everybody in the organization and be an effective leader? That's something I've learned, not only from seeing some of the CEOs, but also seeing many great leaders at some of the companies I worked with.
SPEAKER_04:Great. All right, over to you, Gari.
SPEAKER_03:Oh, great. Thanks, Shirish. So, Rajeev, let me kind of start with this, right? Uh, you're the classic case where uh you're eating your own dog food, right? Because you're based, as you said, in Bellevue, Washington. You, the company start in Australia, has a lot of presence there, many, many time zones away. I'm sure that every day for productivity of knowledge worker, put yourself in that shoes, not somebody running a large organization. How do you uh conquer that? Because our audience, who are all startup uh entrepreneurs, have this problem every day. They won't find talent just in Bellevue, Washington, or Kirkland. They're going to find talent wherever they can go. But your product is a glue, as you describe it, to keep the teams in sync and uh producing every day. How have you embraced it? Because you went from fairly large companies to, in theory, a smaller company because uh, you know, Atlasino, smaller when you joined. Um, how have you embraced this and brought that entrepreneurial spirit uh to your team who are spread arguably all around the world in many time zones?
SPEAKER_01:Yeah, no, look, great question. Like one of the things we have done as a company is um, you know, we have really embraced the concept of people working anywhere. In fact, we call it team anywhere, right? So we we really have a philosophy where you know we are fine with you working in any part of the earth, um, whatever makes sense to you, as well as in the office. You know, I'm I'm right now in Bellevue in one of in the Atlassian office. It's a new office, actually. We just launched it uh, you know, in the last month or so. Um so we have offices you know in Seattle now, in in the Bay Area, in Sydney, in Austin, Texas, in Bengaluru in India. And and those are places where we get people to meet, and because you need the human connection. So we we do things called intentional togetherness. So every team does an off-site or a meetup, you know, every every quarter or so so that you get to meet people in person, you do some brainstorming in the office, and then you when you are on Zoom or whatever and you're collaborating, you have that human connection to uh to to build from that, right? Um, but at the same time, we embrace you working from home because you know that we believe is the future in terms of distributed workforce and flexibility of work, right? Now, where it becomes interesting is our software is uniquely set up for that, right? So we have, for example, we use Confluence. Confluence either really helps you collaborate across distributed teams. And so, to your point about eating our own dog food, we use Confluence a lot. Um, everything is documented in Confluence. Confluence is something that you go to to figure out what's happening with the project or what's happening with different aspects of what you're working on. We did an acquisition where we bought a company called Loom that's now part of the Atlassian family. And Loom is amazing because you can do async video, right? So instead of having a meeting, you know, you could you could spin up a Loom, record your thoughts, walk through a conference document, walk through a Jira ticket, and publish the video. Boom. And then that goes to whoever you want, you know, to send that to. And if they're in a different time zone, they wake up at a different time, they can read the loom or they can view the video essentially. It's an async video and get up to speed. So sometimes it's hard to coordinate meetings across time zones, right? So Loom, async video, we believe is a great medium to go do that. And we have integrated that you know into our products like conforts, right? So definitely our products go a long way towards distributed work. Um, of course, when we situate teams, you also look at uh what we call sort of um, you know, you can think of the earth as having a few halos, right? So, for example, US West Coast and Sydney is a zone across which you could have a team, or you could have a team across Bengaluru, India, and Sydney, you know, that's time zone friendly, right? So we don't also randomly take a team and distribute it across all parts of the globe, otherwise the team won't be able to have a meeting at any you know conducive hour. Uh, but certainly the confluence and loom mediums help you do distributed asynchronous work, right? It's all about async, right? If you can do async distributed work, then you don't have to have that synchronous meeting, right? Um and so those are the elements that come together. For engineers, we have leaned into developer joy and developer productivity through products like Campus, which is something we offer to other customers. And um, compass, like I said, is a registry, it's a place you can go to. If you are stuck, you don't know what this microservice is, you don't know what this component is, you're trying to call some API. Instead of waiting for somebody to wake up in another time zone, you can go to Compass, you can find the documentation, you can find the API, and you can unblock yourself and get going.
SPEAKER_03:Great. Congratulations on opening the new office. I think you have the unique distinction of doing it twice back to back, once for Meta and once for Atlassian. So you may you may have a new skill that that's uh that was not a CTO uh dribble. But let's uh if you could take a few minutes and describe to our audience the uh AI initiative within your product and or products you have created with AI baked in uh first and how what the use cases may be.
SPEAKER_01:Sure, yeah. Look, um, first of all, let me just step back and say that that you know the AI moment we are having right now, I think is a big deal. This is a big, big moment for technology. If I go back in time, probably the two other big moments are when the iPhone mobile revolution happened, and then before that the internet, right? So I do believe that this is a big deal for not only tech, I think the difference between uh AI and sort of like the previous ones is this is going to be even more far-reaching in its impact on all kinds of industries and and all people across the globe, right? So it's a it's a big moment for us to lean into. Like I said before, AI has been around and ML has been around for a while. What's really revolutionary is the LLMs and Gen AI and the ability to like really train a model to understand language and then therefore transpose you know text between um computer things and and languages and so on, right? Um, and so uh when I look at it, there's kind of two different things happening. There's the LLM wars, if you will, there's a whole bunch of companies building LLMs and so on. That's similar to the cloud wars, excuse me, or the browser wars of before. And you know, certain companies will operate in that mode, and that's what they will go after. And so there'll be some winners and and such. Um we are not in that game. We are we are in the game of applying those LLMs to our end users to make our end users more productive, right? So Atlassian is in the knowledge worker game, right? So if you think about what knowledge workers do in any kind of company, right? They do work. How do you define that work? You know, today a lot of that work gets defined in spreadsheets or you know, bespoke tools and things like that. Whereas what Atlassian is selling is a system of work, all the way from OKRs and goals at the C level down to uh team goals at the departmental level, down to Jira tickets at the individual level, right? So if you really, for any given company, this is what they want. They want to be able to track their system of work across their talent, across their people, across budgets, and manage all of that. And that's what Atlassian offers. And so for us, our main thing is how do we use AI to make that better, right? So if you're a company building a rocket that's going to Mars and you have a bunch of Jira tickets, what we would like to say is, hey, just come in and say in natural language, when is the rocket going to launch? Right? And imagine it goes and grinds through all the tickets and says, you'll be ready in 23 days. Okay. That's value, right? And so that's what we are working on is how do we bring not only natural language but also AI productivity to your to the work that you do using the products that we offer you, right? You have hired us to set up the system of work. How can AI make you more productive, right? Um, another example is we have something called Jira query language, JQL. And you know, a lot of people go and do a lot of JQL queries to find things and so on. But we have worked on something where you can do a natural language query and that converts it to JQL and then does the job for you, right? So basically, we are looking to infuse AI into all the things that people do with our products today to make them a lot more productive. And we launched that as something called Atlassian Intelligence last year. So that is the first wave of our products is Atlassian Intelligence in Jira and Confluence. You know, you can have a summarization feature, you have a sidebar chat that allows you to do things more productively. Um and so that was the first wave. The second wave is what we call Robo, which we announced in April and we are going to launch later this year. And that also brings in something called agents. So now you can have agents that work like teammates that do specific tasks. For example, if you have a marketing agent that wants to do a press release, well, you can you can build an agent for that. So now we have a whole platform that allows you to build agents to do specific tasks. These agents aren't super smart yet, but they'll do one sort of uh you know narrow task really well, and that helps you with your productivity, right? So um, so we have agents now and an agent platform. And then also we have improved search because search continues to be something that's really important to be able to find information in an enterprise, and that's part of our Robo offer. So that's our sort of second wave. So we're already into our second wave of uh AI, you know, from the Gen AI wave.
SPEAKER_03:Got it. So um let's sort of step out. We we've talked uh in our Gen AI series to a variety of investors, startup CEOs, etc. Um, one of the challenges that everybody faces is the current LLM, whether it be closed or open, it doesn't matter where the what the ilk is, they could do things and turn on a flip, uh flip on a switch, and then suddenly what is a company could become a feature because they should they just have it, right? Um, you guys are at the core of a knowledge worker's livelihood. This is it. I mean, they if they kind of lose some product of yours, they they could be lost for a long, long time. But the attack uh of AI or its use cases in at least in mid to large enterprises has been uh to protect the workers from what is called routine tasks. You're now talking about uh eliminating those routine tasks and giving them uh results that can make them uh joyful, as you said, right? Uh as you um think of AI, you are thinking differently because you're a company that has to think about the human as opposed to the AI, right? Because LLM players are thinking about making LLM better. They just want to maybe I could be a poet. I never have been, but I could be. But that's not what you're thinking about. So as you, as the uh leader of a classium, are in a unique position to uh uh both impact as well as uh uh inspire humans to do better tasks. How did you uh go about in your two ways that you've done to say, hey, let us see how to make somebody more productive as opposed to the fear that you're eliminating something, either a job or a task or anything. You you're in this unique uh leadership position as a company and as a leader in that company. So if you could talk to our audience saying AI doesn't have to be scary or it is scary, then tell us.
SPEAKER_01:Yeah, yeah. Look, uh I think that um, you know, any any new technology, right, um comes with these kind of sort of fears of things, right? So long story short, my view is that AI is gonna help human beings be more productive at a at an aggregate level. I don't think it's it's gonna like take away jobs and things like that. I think you know, at least my view is that it's gonna help us actually create more jobs and uh perhaps have human beings go up the value chain in terms of being more productive and doing higher value things, right? To the extent you have teammates and agents that take away some of the routine stuff, uh that allows you, therefore, you know, the human mind and the human brain is has uh enormous amounts of creativity and talent, right? So we can then uh reserve those brain cycles for those more creative kind of flows, right? And um, so I think I don't so that so I that's my view, is that I think it'll be net positive. Uh doesn't mean that there won't be like certain, you know, uh uses of the technology that we need to worry about. There's obviously things we have to worry about with AI as well, but largely I believe it'll be positive and we have to manage that transition carefully as a society, of course. But um I do think you know you can you can get more productive through AI. So that's our view. Um, but also we are very early in the AI adoption. I don't think, notwithstanding all the hype and everything else, uh, you know, we are still very early. We 50% of knowledge workers, you know, we've done some surveys and things like that, and there are many studies out there, but I'd say, you know, 50% of knowledge workers don't use AI on a on a weekly basis. You know, maybe they use it a little bit, there's a little bit of a novelty effect, and then they they don't use it. So um, but those teams that are using it regularly are finding it to be about 1.6x more effective, right? So we are seeing that if if you actually have a use case that that that lines up and has you know product market fit, so to speak, with that with that set of people, they are being more productive in those use cases, right? Um but also a large proportion of knowledge workers and executives are struggling with how to use it effectively, right? So that that's some of the data we have seen and some of the research and so on, right? And so um again, I think we are early in that in that evolution. I think we are seeing some use cases where it's super productive, um, including for engineers. We have seen that you know, like I said, some of the coding assistants help save you one to two hours per week. Um, and we are experimenting with them and also building some of our own. Um, but also I think um, you know, we're all trying to figure out what is that sweet spot and where does that productivity gain happen, right? Um so I think we should not get too caught up in the hype. We should understand what is working, what is not working, and that's the business we are in, right? So we are looking at hey, how do you get your job done? What is going to help you be 20% more productive, 30% more productive? And in some cases, there will be a breakthrough where it would be a nonlinear jump as well, but we're all still experimenting to find out what that is.
SPEAKER_03:Got it. So do you um when you build the uh let's say an AI agent, or do you then go back to your core tenants and say, hey, is that does this give me, say, for example, developer joy? It how do you guys decide? Because you you're you're building around what's already being used every day. I mean, you your products as you describe it is an everyday, you can't live without it, right? It's it's as core as email or my phone. That that's how core it is. If that is the case, how do you because I don't want my phone to change things because I've gotten a I've gotten used to things. I wanted to do things easier, but not change the way I'm using it, right? This becomes tricky for you because everybody's using Jira differently than what you what you have decided. Uh how do you keep to the developer joy as you infuse AI so that it is not a hurdle as opposed to it's uh it's some way helping them achieve that joy? Because that's your that's your that's the reason why you exist. AI could be useful, but these are extremely smart engineers who are using your product. They know everything about it.
SPEAKER_01:Yep, yep. So, you know, um with with engineers and developers, for example, I gave you the example of how you could take a Jira issue and you know almost come up with a pull request, you know, for it. So imagine an engineer who goes in to fix a bug, right? They have a Jira issue, they go read the issue, and then they go into their source code to go and try to write write the code to fix it. Well, if you had an agent, like a coding agent or coding assistant that understood enough about the issue to be able to create the starting point coding, you know, for it, sometimes the code could be almost nearly done. And all you have to do as an engineer is go in, make sure it's okay, and then submit it. In some cases, maybe it's half done, and you go in and you have to do a little bit of work on it and then get it done. But in all cases, it's boosting your productivity, right? So we are trying to sort of keep it in line with what people are doing, but boosting the different stages of productivity that they have so they can do more. And so it's more like a helper or an assistant helping you get things done. The other part of it is you don't want the AI to be a black box. So even with the example I gave you, you can the software opens up saying, hey, this is what I'm trying to do, or this is what I'm trying to do. And if the human wants to go and tweak that, they can they can go and do it. Like it's the analogy is you know, if you have autopilot in your car and it's driving itself, you as you're still there as a driver, you can still go in and take control and do a few things. So it's important to not build the AI in a way that it's a black box that humans can't you know understand what's happening, um, because you do want that level of control. Now, over time, certain elements will become so sort of routine that you don't want to worry about it you know anymore, right? You know, so for example, like like cruise control in cars, right? You don't necessarily once you set it in cruise control, you know that that's going to work a certain way, and then you're fine, right? So over time this will become better and better where we will trust more of those things, but in the beginning, you don't want to have a certain level of human control in some of these things, right?
SPEAKER_03:Right, right. So what's fascinating from what you're what you just said is uh the idea that it's not a black box, so you are transparent in what it's doing, and then allow for uh humans who are vastly at the mo at the moment more intelligent than any AI tool out there today, can do it better. You you almost like a leverage tool, so you can get get a lift as opposed to just saying, hey, uh it'll do your job. I I find it uh sort of rooted in empathy more than just saying we have a better tool, because you you have a very difficult, I mean, from my from our perspective, a very difficult position because um developers are you know power users of everything, but they don't want job to be done for them. That's core being an engineer.
SPEAKER_01:Well, this is not uh an alien concept to developers. You know, if you we've had for a long time the ability in in IDEs, you know, to have more and more sophisticated tools like tooltips and you know, uh code gen, uh autocode gen templates and things like that. So for developers, this is a good thing. Like anything that helps you write code faster, get started faster. Also the blank slate problem, right? You open your editor, and it's like, where do I start? Right? There's a little bit of a blank slate problem that that you know the coding assistants help with that because you can put in a few comments and you can put in a few things and it puts some code for you. And then once you see some code there, you kind of get started and the flow comes kicks in, right? So we very much see As being helping human beings, you know, do some of the routine tasks and then therefore unleashing their own creativity and productivity. So that we talked a lot about developers, but let's back up a little bit and look at knowledge workers who are non-technical and are using our products like Confluence and Jira and so on. We have really infused AI in with a little AI bubble, right? So there's a little thing that sort of sits there on the top along with the all the other edit controls. And if you click on it, you know, it allows you to go and either summarize the document or learn more about simple things like three-letter acronyms. You know this, every company has its own set of TLAs, right? But what we've done in Confluence is in your company, the three-letter acronym gets highlighted with an underline. And when you click on it, it actually goes and searches your enterprise context and tells you what that three-letter acronym is. You know, that's a simple sort of tip or helper that's so valuable, right? Imagine anytime you're new to a company, the first thing that drives people nuts is what is that three-letter acronym, right? And so simple things like that, I think, really help people get that extra time back, that extra context, that extra understanding of things, right? Um, and there's so much potential in improving productivity for people. So that's kind of like the mission that we see ourselves in.
SPEAKER_03:I think it that probably you built it yourself so that you can understand that Latin TLS, I think. We'd have to check with your with your team. Was this done for the is it the Rajiv special? Right. Um, but uh let me ask this. Uh uh, we all ask all our guests, and this is uh thing that we have uh been doing throughout the series, is uh a prediction, right? As you said, and it's early. So what would you predict say a year from now the most likely thing that um that based on everything you see is uh to happen? So just a prediction. Uh we'll hold you to it because this is our way to get guests back to our show. So uh give us a prediction for um for uh genai uh from Rajiv.
SPEAKER_01:Yeah, look, you know, um predicting is a dangerous game. But um I I I'll say this, you know, we are we are probably um overestimating you know uh how much AI will do, right? So one thing I will say is a year from now, we'll have made a lot of progress, we'll have a lot of use cases, uh, but we won't be anywhere near things like AGI and like you know, human beings being replaced by AI and some of those kind of like uh you know cases that people are saying. Uh but I think we'll be much further along on the LLMs, especially with modalities. You know, today a lot of it is text. I think video and and audio and other modalities will become more commonplace as well in in some of these things. Um and then the LLMs are making a lot of progress, so I think the context windows will become much bigger for sure. And um with that bigger size, you know, you start to then be able to do more interesting things with code. So I think that revolution with respect to um coding assistance, I think will will continue to happen at a at a good clip. Um and then I think like a lot of things will become table stakes. Things that we thought of as AI features today, like summarizing something and so on, I think will just be expected. You know, just like spell check. Remember, spell check, grammar check, you know, you take that for granted, right? Um but so some of those things will become like it's no longer AI, it's just you know, it's part from the course, it's there. Um I think agents will become more ubiquitous and we'll start to trust and use more agents in what we do, uh, little agents that do little things, but then make your day more productive, but not like complex agents that you can give an entire like go uh do this you know interview with Rajiv for me. Like uh here's my Gauri avatar that goes and does the interview for me. I think we're a little ways away from that one. We still need Gowri to be Gowri, you know.
SPEAKER_03:I like that. Hey Rajiv, this has been fascinating. Uh I like uh what you have uh uh what you said about AI, especially how you are impacting Atlassian with uh with uh your approach, rather unique approach, I should say, uh with AI. Thanks a lot for making the time and uh best of luck in your new Bellevue office. Uh, we would like to visit just to get that straight, once your cafeteria is up and running.
SPEAKER_01:For sure, for sure. Thank you, I really enjoyed uh the conversation and the questions and what you're doing with uh startup to exit. And for sure, we would love to have you over at our office. Thank you very much, Rajiv.
SPEAKER_02:Appreciate us.
SPEAKER_03:Thank you. Thank you for listening to our podcast from Startup to Exit brought to you by Dai Seattle. Assisting in production today are Isha Jen and Mini Varba. Please subscribe to our podcast and rate our podcast wherever you listen to them. Hope you enjoyed it.
