Atlassian On Why AI Must Deliver Measurable Business Outcomes

[00:00:04] What if the real measure of AI success has nothing to do with novelty and everything to do whether work actually gets better? This was one of the quiet but consistent messages coming out of Davos this year. Because beneath the headlines and bold predictions, leaders were clear that artificial intelligence only earns its place when it delivers outcomes that people can feel.

[00:00:29] Whether it be creating better work, improving decision making and making everyday systems function more effectively. These things matter far more than shipping features for the sake of momentum. And software developers, they offer a powerful way to understand this moment. And over the last 20 years, their role has already shifted well beyond just writing code.

[00:00:53] And developers now shape products, influence operations and increasingly carry a responsibility for how an entire system can behave in the real world. And AI also fits into that evolution naturally. Because it takes on repetitive execution and accelerates delivery. All while raising expectations around judgment, design and accountability. And today we have the customer CTO at Atlassian joining us.

[00:01:21] And we're going to talk about how he is seeing this shift up close through his work with customers navigating AI adoption today. And why he believes AI is not going to replace engineers, but it will demand more from them. And as machines absorb some of that routine work, developers will inevitably gain time to focus on higher value engineering. From system design to orchestrating solutions that solve real business problems. This is what we need to get back to.

[00:01:50] So in this conversation, we'll explore why AI is likely to create more developer roles rather than fewer. Why those roles will require deeper thinking, yes. But how engineering careers are already shifting towards reviewing AI output, designing resilient systems and integrating automation into dependable solutions. A lot of good news ahead, which hopefully means we can restore a better balance of optimism in the universe. There's enough doom and gloom out there.

[00:02:18] So as AI reshapes how software is built and how teams work together, the real question becomes, are organisations truly preparing people for a future where responsibility will matter more than speed? But enough scene setting from me. It's time for me to officially introduce you to my guest. So thank you for joining me on the podcast today, Andrew. Can you tell everyone listening a little about who you are and what you do?

[00:02:48] Yeah, thanks a lot, Neil. So excited to be here. I'm the Customer CTO at Atlassian. It's an organisation that was founded over 20 years ago with the mission to unleash the potential of every team. My role within Atlassian focuses on advising and helping our customers to accelerate delivery, improve collaboration and drive enterprise-wide transformation. Well, it's a pleasure to have you on the podcast. Last year, there was that infamous stat of 90 or 95% of AI projects that are failing.

[00:03:16] And there's a big focus on ROI of tech projects now. And at Davos this year, that theme continued where leaders were very clear that AI only matters if it delivers very real outcomes, delivers value, measurable results, etc. So from the conversations you're having with your customers, and you get to speak to so many different organisations, your customers seeing genuine impact from AI today rather than experimentation for its own sake or stuck in the dreaded pilot purgatory.

[00:03:45] AI is fundamentally changing how we do work. And it's a huge opportunity for all enterprises across all industries. The organisations that seem to be experiencing the biggest gains from AI are the ones who are using it to solve specific problems. So they're not giving all their employees AI and just telling them, here you go, here's AI, go be more productive. They're generally starting small, they're experimenting and iterating quickly.

[00:04:13] And they're really focused on team level adoption, not just like individual productivity. So in these companies, we see them treat AI as a teammate, not just a tool. And they invest a lot in upskilling their people to work alongside AI. So the combination of experimentation, openness, continuous learning, these are common patterns that we see in companies that are using AI really well. And a great example of that is actually Mercedes-Benz.

[00:04:42] So if you take a look at Mercedes, they've got 35,000 engineers and staff, and they used to spend a lot of time going through duplicate tickets and creating different reports. So it turns out that nearly 90% of all of their bug tickets were duplicates of issues they already knew about. So they built a duplication detection agent that helped them tackle this problem. They cut duplicate and wasted tickets by around 85%.

[00:05:10] And that saves their team, obviously, a lot of time. So if you think about why did that work, they had a well-defined problem. They identified a great AI solution. They experimented and they implemented it, which ended up freeing up time for their teams to spend more time on more important work. Such a great example there and a metric that everyone can get behind straight away.

[00:05:35] And I do think we take for granted or underappreciate just how much the workplace has changed in the last two, three years. I mean, software developers, they've already evolved from just writing code to now owning products, operations, and even infrastructure. So if we look further ahead, how do you see AI fitting into that long-term evolution of the engineering role? Because it's not about replacing people. It just feels like it's evolving, isn't it? Yeah. I mean, it's such a great call out.

[00:06:02] The role of engineers has already been evolving for decades now. And in my opinion, every evolution has been hugely positive for engineers. So if you ask most executive leaders what they're hoping to achieve with AI, it really boils down to productivity. And they don't mean more lines of code. They want to deliver outcomes to their customers faster. And engineers are really well-placed to support organizations with that goal.

[00:06:30] Engineers are already the most productive group in most organizations because as a craft, we spent decades designing how work happens, not just what work has to happen. So if you think about agile, think about DevOps, developer experience, all of these things are about improving flow. And the principles from these are applicable to all types of teams. And they really are the blueprint for enterprise productivity.

[00:06:55] So most organizations already have in-house experts in productivity, their engineering team. So if you think about how does AI fit into this evolution of an engineer's role, AI has already significantly shrunk implementation time for most software teams. Which means that over time, engineers can spend more time on solving business problems and helping the rest of the organization improve their productivity. So the day-to-day tasks that engineers have historically focused on might evolve.

[00:07:24] But overall, we're moving closer to a world where engineering teams can help support their whole organization to be more productive and they can get a deeper understanding of business problems. And before you join me on the podcast, I was doing a little research on you. And one of the things that I read was that AI won't replace engineers, but it will raise the bar. So in practical terms, what kind of work do you expect to fade into the background and what new responsibilities move to the forefront?

[00:07:52] Because I think that will also help engineers listening understand what they need to be training for, where they need to improve to. Yeah. I mean, that's right. AI is not going to replace developers, but it will give them time to work on higher level and more rewarding work, in my opinion. So if you look back at last year, our state of developer experience report in 2025, we found that coding itself is only around 16% of a developer's time.

[00:08:17] Now, with AI, the role of an engineer is evolving even further because AI can generate code or spin up prototypes really fast, which gives engineers more time that can be spent understanding complex business problems and identifying engineering solutions in new and more creative ways. So reflecting on my own career, I've had a similar shift. So after 20 years as a technology leader, I completed an MBA around three years ago that gave me

[00:08:44] a business perspective that I now apply to my role. So for example, I had previously very little exposure to how marketing teams worked. And through my learning a lot about marketing, the types of technology they use, which as a side note is very similar to CICD and generally what a marketing team's goals are. So, and since then, I've been able to partner with lots of marketing teams to help them improve the way they work and how they use technology to achieve their outcomes.

[00:09:12] So over the next few years, I think we'll find more efficiencies within software development as we have over the past 20 years. But we're able to use our logical, our systems thinking approach to solve business problems we historically haven't had capacity for. So as this evolution happens, we're going to have to rely more firmly on our skills of communication, collaboration and problem solving to work closer with various different business domains.

[00:09:38] So in some ways, we're going to need to rely more on our skills that are uniquely human. And even as AI takes on those repetitive tasks, you refreshingly argue that, hey, yeah, this is great, but engineers must always stay firmly in the loop. And why is it that human judgment remains so important or should remain so important across the entire software development cycle? I suspect we've both seen what happens when they are not.

[00:10:04] But for people listening outside of the tech space, can you expand on that for me? I think we need to start considering AI as a teammate. And AI takes on the repetitive tasks and humans in the loop provide context, direction, and they make important decisions at different points throughout that software development lifecycle. And this is really important because at the end of the day, the humans are responsible for outcomes, not AI.

[00:10:30] So if something goes wrong, we're not going to have a meeting with an AI agent to understand what went wrong and how we're going to avoid this from happening again. This happens with the humans. So when you think about how great software is built, it generally means deeply understanding customers and their feedback. And AI can help us iterate faster towards customer delight, but the process of understanding context and priorities is still largely human. AI can boost productivity, but it's not a silver bullet.

[00:10:59] It's only as effective as the quality of its inputs and the expertise of its users. And to this extent, human involvement still remains essential. AI agents can't just be left unchecked and accountability for outputs are going to remain the responsibility of human engineers. I love that. Humans are responsible for the outcomes and the accountability as well there. And we are very early in this transition still, but the direction does at least seem incredibly clear.

[00:11:27] And I love to bust myths and misconceptions here. So continuing on that same theme, can you just expand on why AI will actually probably lead to more developer roles overall, even if those roles look very different to what we see as a developer today? Yeah, there's a hypothesis going around that AI is going to reduce developer jobs. And I think it's based on the assumption that by using AI, teams are somehow going to finish their backlog.

[00:11:53] And the reality is there's always going to be more work, more problems to solve, more solutions. And in fact, by doing things faster, software teams' backlogs are probably going to get bigger. And yes, AI is going to enable us to do more and to do it quicker, but it will also create more of that substantial deep work that we were talking about earlier. So more creates more. So there's always going to be a need for developers. And this is already happening.

[00:12:21] Within Atlassian, we're hiring more grads this year than we have in previous years. And even if we reflect on history, like every major technology shift has resulted in requiring more developers. So it's looking like at this moment, AI is going to follow a similar pattern. Yeah, I completely agree. We've seen it with every tech trend, every emerging technology. And as you said, more equals more. And you can get rid of that backlog. But all it means is there's going to be more work coming in. So yeah, completely with you on that.

[00:12:49] And as engineers spend more time reviewing and refining AI output, how should teams be rethinking skills, training and career progression to prepare for a shift like that? Because we will have people coming into the workplace. We need to maybe improve the entry level roles there. And also people that have been in a role for a particular amount of time. They don't want to be left behind. They need investing in as well. How do you get that balance right? What's the secret here? Yeah, such a great question.

[00:13:16] We've gone through multiple stages already with AI. First was an assistant. And now AI stepping up in some ways as a true teammate. Leaders need to recognize that they actually play a key role in helping their teams prepare for that shift. Some research that we've let out recently, in 2025, we did an AI collaboration index. And we use that research to help us classify different types of AI users.

[00:13:43] So when people first start using AI, we call them, I think it was a level one and they're labeled as simple AI users. So these are people who think AI is a tool. I use it every now and then. It does a specific thing for me. Or maybe they're using it like a personal assistant that helps with hard parts of my job. And then at the other end of the spectrum, we have users of AI that we call strategic AI collaborators. And these people talk about AI very differently.

[00:14:11] They describe it as a creative partner with its own skills, its own insights. Some people describe it as a team of expert advisors who can enhance my decision making. And that mindset shift has a huge impact on the value people get from AI. So on the first part that I explained, the simple AI users, the early stage people, they're saving about 53 minutes per day using AI, which is still a lot. That's still like getting a whole meeting back per day.

[00:14:40] But when then you look at the strategic AI collaborators, the people who are, let's say, more advanced, they're saving around 105 minutes per day. Now, these figures were actually that we did two research reports. These figures are from 2024. So I'd expect that these figures are much larger now. But you can see there's a big difference between those people who use it as like a task manager and ones who are collaborating with AI.

[00:15:06] And from a leader's perspective, it's very tempting to think people can just figure this stuff out on their own. But our research actually shows it's not the case at all. Leaders who encourage their teams to experiment with AI have teams who say, 55% more time per day than those who don't. So leaders aren't just influencing how much AI gets used. They're shaping how much value their teams get from using it.

[00:15:31] So leaders have a huge role to play when it comes to getting the benefits of AI for their team. Again, so many big stats there in your answer. And as we continue designing systems and orchestrating solutions becomes a bigger part of the engineer's job role. How do you see AI changing the way that engineers maybe approach system design and integration inside a modern organization? Any big changes occurring there? Yeah. Yeah.

[00:16:00] One thing I think we forget in all this AI hype is that every system we use is still designed by and built by humans. Yeah. Whether it's a customer journey, a workflow, a piece of software, a model pipeline that AI runs on. Humans still decide what the system looks like, what it optimizes for, and where it fits in the business. So what's changing with AI isn't who designs the systems and the solutions, it's how we spend our time on them.

[00:16:26] So with AI, we're able to offload chunks of that build work, and that frees up time for us to spend on the part that actually moves the needle, which is deeply understanding the system and how it maps to real business problems. And so there's a hidden upside that we don't really talk about enough, which is what this does to the quality of the solutions we produce. So over time, we've learned that the best systems really just arrive fully formed.

[00:16:55] They emerge through iteration, through prototypes, feedback, course corrections. And now with AI in the loop, we have the opportunity to compress that whole learning curve. And it means that we have a real shot at ending up with better systems in a much shorter period of time. Yeah. Humans decide everything and also responsible for outcomes, as you said earlier. Yeah. And speaking of humans, I know one of the things that you're passionate about at Atlassian is the community there.

[00:17:23] And Atlassian, you are looking ahead to your Team 26 event in Anaheim. I will be there, same as last year. And I'm going to use this as an opportunity to try and tease a few themes or areas of focus that will let me and the listeners find out a little bit more about what you're thinking about, especially regarding the future developers and AI and a few other things we might be looking forward to seeing. You're probably locked down. You can't share too much.

[00:17:51] But are there any teasers you can leave me with? I'll give you a couple of teasers, maybe two. Let's see if we do. So the first one is AI impact and measurement. So we've spoken a lot about AI today. So we're going to show how Atlassian becomes a system of record for understanding the real impact of AI on software delivery. Powered by the team at Graph and insights from Rovo. And RovoDev provides leaders with end-to-end visibility into flow, into quality and outcomes.

[00:18:22] And this theme anchors on our return on investment on AI. So it's not just shipping AI features, but proving how they reduce bottlenecks, how they improve developer experience and accelerate value delivery across different teams. And the second one I'll share with you is agentic software development. So we're sharing more about moving from AI-assisted tasks to truly agentic software development. RovoDev embedded across software collection and Jira Service Management orchestrates work

[00:18:51] across planning, coding, reviewing, and operating software. And when you combine that with DX and the team at Graph, we're going to show how Atlassian helps teams define golden paths, how to automate different parts of the SDLC with guardrails, and safely scale AI agents across their organization. Oh, lots of teasers, lots to be thinking about. And I think one of the big standout announcements of last year's Team 25 event was the Atlassian entering the world of F1 now.

[00:19:20] And I know this is something you're quite close to. So any updates you can give around that and that relationship and what you're working on? Yeah. So I'm leading the transformation that we're running with Atlassian Williams F1. We're really excited for the beginning of the season. So it's starting next month in March. And within Team 26, we're going to share a lot more about what we've been doing with Williams F1 since we last shared some of what we've been doing. Fantastic.

[00:19:46] And for anyone listening wanting to find out more about the event, exploring some of the announcements as they drop, keep up to speed with all things Atlassian, connect with you and your team. Where would you like to point everyone listening? Yeah. So for Team 26, you can visit the Atlassian website. I'm also very active on LinkedIn. I often share details about the transformation that we were just talking about, the one that we're running with Atlassian Williams F1. And one more thing to look out for, we're releasing our next data developer experience in mid-May.

[00:20:15] So good to keep an eye out for that one. Excellent. Well, I'll add links to everything that you mentioned there. And we did cover a lot from how AI will create more developer jobs, not fewer, but they will also demand higher level thinking. And also how the engineer's role will increasingly focus on reviewing and refining AI, designing robust systems and integrating AI generated tools into solutions. So many big changes coming. It does feel an exciting rather than a scary time.

[00:20:43] So just thank you for putting it all in a language everyone can understand. And looking forward to maybe meeting you at Team 26. But thanks for joining me today. Thanks, Neil. I think one of the strongest themes running throughout this conversation today is that very idea that AI does not remove humans from the equation, but it does sharpen their role. And as Andrew explained, even as AI accelerates development and reduces repetitive effort, people remain responsible for outcomes.

[00:21:14] Engineers will still decide how workflows behave and where technology fits within their business. And it's that responsibility that might bring new pressure, but it also brings a chance to do more meaningful work. And yeah, we are still in this transition, but that direction feels impossible to ignore. Teams that treat AI as a teammate rather than a shortcut seem to be moving faster and have more confidence. But at the heart of all this, they're investing in skills.

[00:21:43] They're encouraging experimentation. And remaining focused on real problems rather than surface level efficiency or the next shiny piece of technology. And looking ahead to Atlassian Team 26, that event feels like another important checkpoint. With that growing attention on measuring AI impact, system level visibility and agentic approaches to software development, all these things collectively signal a future where AI is becoming embedded across

[00:22:12] planning, building and operating software. All with humans firmly accountable throughout the process. So before we wrap up today, will you be attending Team 26? If you are, remember to look out for me. Give me a shout if you are there. And also the work ahead. What emerging themes are you most excited to talk about and share and get involved with? And how are you preparing your teams for an agentic era of work? I'd love to hear your thoughts and experiences right now.

[00:22:42] So please, techtalksnetwork.com. There's a myriad of ways you can get hold of me over there. I'd love to hear your thoughts on anything we talked about today. Good, bad, indifferent. If you disagree, this isn't a monologue. It is a dialogue. And I'd love to hear all sides of the story. So hopefully I will have an opportunity to speak with you all again tomorrow. Speak with you then. Bye for now.