Building the future of modern work | Chris Davidson

Okay, building the future of modern work. Thanks for having me today. I am hopefully going to share you some stories and some insights - on how we are trying to build the future of modern work. I think we heard from Kelsey about collaboration - and about how we don't know, the future is now, - the future is the actions that we take now. And if we talk about practices, and we talk about ways of working, - and yeah, okay, there's some software and some platforms in there that help, - I'm going to share you some stories about how we're doing that - that I hope you can take away and learn something from. So, we're going to go on a journey. Build the future. What's the future? Let's see about what is the now. Where are we now? So, there was a State of Teams report. We do it annually, 4,000 respondents, - three key areas came out that said the now, - the problems of the now, productivity. Everyone here is, I'm sure, like, no productivity. What about productivity without burnout? Alignment to goals, this second one, - this tracking of progress to goals is really one of the ones - that rings very important to me personally, - because if you don't know why you're doing something or what you're doing, - less than half of people really know what or why they're doing. And stakeholder management, so many teams, so many people, - how do we just deal, like the burnout factor of dealing with that. So, that's current work, in some respects, - and you may have different experiences, or some of this may resonate. So, what's modern work? If we were to try and describe what would feel good, - we know why we're doing it, we know purpose, - the teams that we work with, we work well with, - like Kelsey's point, what's wrong with the collaboration now? And of course, productivity, right? More with less, but more without burnout, right? Productivity without burnout. I think there's a really key thing that I want to level set on, - both personally and from an Atlassian perspective, - is that going from and to modern work is an evolution. It's the actions that we take. It is not the silver bullet. Go and buy all the Atlassian platform, and you'll be there. It's an evolution, and it's your journey. But we have to touch that LL-LL- [stuttering] LLMF... I can't say it. You did it so well. LLL-elephant in the room. Atlassian. I stand here saying this is an Atlassian viewpoint. We see a future of human and artificial team members. Why is that? Because they both bring something to the party. Do not underestimate human intuition. Intuition is not a guess. Intuition is, you know something, you sense something - based on your experience and your knowledge, - and very quickly, you can form a viewpoint, - which may be an LLM, - which can look at billions of parameters, - can validate and can soak up a fair amount of energy doing it. But there is a complimentary thing there. Nothing like a human has intuition that can work as effectively as it can. What happens if we don't know? What happens if we don't have a dataset? If we don't have a dataset, humans are actually quite strong. LLMs less so. There is a role for us both. We see a future of human and computer teams. This whole concept of bringing this transformative technology, - I want to just pause for a moment, and Kelsey took us back - to Ruby on Rails and Hadoop. There's a story I've told a number of times, - and I've told it in every other location but London. And this story, it resonates with me - because it's a great story about a team, - but it's a great story because it happened, - well, the actual work happened just a few streets away, - just by Old Street, a place called the Tavistock Institute. And let's just set the context. Technology influence on ways of working. British coal mines, 1940s and 50s, we need a productivity increase. It's the post-war era, we need to be more productive. And there was a whole approach, which I'll tell you about, - that took place to raise that productivity. But a team came together, an awesome team. And the reason why this resonates is that this team here, - cross-functional team, you've got a sociologist, Trist, psychologist, Fred Emery, and Bamforth was a coal miner. Eighteen years he'd been a coal miner, right? Emery is the guy I resonate with because he's an Aussie. He came over from Western Australia. He studied here in the Tavistock Institute. He went back, I stayed, 25 years later, still here. But what did they do? What did they do? So, they went and looked at what was going on - in productivity in the coal mining industry. And the big change was going from hand-got method - to longwall method, which is where technology and machinery comes in. The hand-got method was all about a team that would have specific roles. There was a person that did the cutting of the coal. There is a person that would put the roof up - where the coal had been dug out, and then a super important role, - the carrier of the coal that would take the coal out. And they would self-select. The skills would pass between the teams. They would select who was going to join the team. They would try and aim for the best section of coal, the softest section of coal to collect. And if the coal was really hard, they would self-decide - in terms of their goals about how hard they will work. This is a dangerous place. You're in the dark. It's dangerous. You're relying on your teammates, on knowledge, on skill. They have craft pride. Craft pride, execution independence, and multiplicity of skills. There is, of course, more than one team involved. There are multiple teams, but they are essentially autonomous. They are making decisions based on what they see and how they are executing. There is not a huge amount of intermediate management - and layers over the top. Then, technology transformation comes along. Big cutting machines, - big tracks and rails for how you can move things out. And this is brilliant because we can now get engineering metrics. We can get productivity metrics, we can do mathematics - about how much coal we can get and how quickly. Brilliant. Let's break the roles down. We don't need as many roles because the machine's going to do more. So, let's break the roles down. Let's actually pay them differently per role and not by team as well. And that way, we're going to not just make more coal, - but it's going to cost us less, and it's going to be brilliant. And what Bamforth, Emery and Trist found was that - they disrupted the social construct or the collaboration of the unit. And so, in fact, because the roles were separated, - and they didn't work together, the goal was different. They had different goals, they wouldn't care about the following shift. Motivation went down. Absenteeism went up. A stoppage, when you have a huge, big, longwall mining system, - a stoppage for just one hour would cause many hundreds of tonnes of delay. And they actually proved that, in fact, it was less productive. And their job that they did just down the road - was to come up with ways of how could they transfer learnings - from the hand-got method into the longwall method. The ship had, to many extents, sailed. There was a real problem with the longwall method. Firstly, intermediate management. Because you separated all the roles and broke the independence, - you had to bring a lot more management in. Unions became very powerful and very big. They managed what used to be a team unit that would connect holistically - with representation on the top. And the other thing is, this machinery, - these cutting machines, the rails, et cetera, - hugely expensive upfront investments. So, they're too big to fail, you've put too much money - into building this mechanical technological productivity improver. And so, I won't go into exactly, there's a whole story off to the side - about what happened with British coal mines, - but not as fast, not as productive overall. And what I want to just pause on for a moment, - the takeaway here is, - when you look at an innovation in a work system, - an innovation in a work system, the takeaway here - has a technical content and a social content. And that's really the key message that when we look at - building the future of modern work, how do we mix - the social and technical content, especially in the context of AI, - to improve the way we work, improve our own happiness, - productivity, alignment to stakeholders, - and of course, a better planet in the process. So, I'm going to share some stories, - three stories about three key elements that we are seeing - when we're building a modern work culture that is working for us, right? And I won't quote lots of data. They're stories. What we're doing, I really hope it gives you some ideas - and some discussion that we can have later on as well. Social and technical content in building a future of modern work. Empowerment of teams. So, we know productivity is a big main issue. Empowering teams to make decisions, - like our teams in the mines being able to execute. Decision-making, bottlenecks in decision-making, - bottlenecks in escalation. We have done a lot in terms of ourselves as we have grown. How do we make decision-making closer to where are those that can act on it? And we've come up with a very basic framework. And that basic framework is, as a team, - when do you make a decision that you can just execute on? Because not all decisions are equal. So, what are the ones that you can execute on? You drop a Slack message, it's just within your team, - You tell the team what's going on. Maybe there's a bigger decision that has some risk involved in it, - and we maybe need to communicate more broadly, publish a page, - have some discussion on it before we execute it. And then, there's the game changer, something that's going to actually - change a whole business, and you might want to run a more formal process. And that more formal process, so you decide, the team decides, - or indeed, the whole organisation decides, - that formal process, we call it a DAICI. Decision, the actual Driver of the decision, - the Approver, the Informed, and the Contributors. A really simple model. And an example of where we use this inside Atlassian is - we had a native Jira Mac App. For various reasons, we had to decide if we would continue it or not. And visibility of a decision is a very common thing in an organisation. You can disagree and commit. You need visibility of decisions. So, a DACI is formed. The roles are very clear. People can see who is involved. Knowing your role is a very important part of modern work. But this team in particular, no one is in for hours of documentation, - but being clear on what options people have looked at, - the potential impacts, including the impacts on the team, - the scrolling of the demo is too fast, but they talk about - negative impacts on the team, when you discontinue the app, - what are they going to do next, et cetera. A really powerful way of having people inclusive - in where we're going, on what we're doing. I will now pause and say, what about the technical content? I've just said, we're going to have AI and human team members. What does that look and feel like? Making decisions. Can AI help us make decisions? A real example, we have here where we're looking at our decision - to continue the Jira Mac App or not. And we're asking essentially an AI team member, - what do you think of this decision? Is it well structured? Is it well thought out? And we are using an L-L-LLM-phant - with knowledge of decisions in our organisation - to say actually, yeah, you know, you've documented quite well, - but maybe, I'm trying to think what we've picked out, - you haven't specified who the driver is. The driver's missing. You've got to have a driver. Who's the person driving this? I wouldn't put that in front of the boss without being clear on driver. Okay. Thanks, AI teammate. That really helped. And that's the first example, when we say AI and human teammates, - of we've got a team, and we've said, could you get someone - who's really good on decisions just to come in and take a look at this? Okay, it's a Rovo Agent. You heard from Erica earlier, Rovo, that's what it is. And what we perceive, - so that was called, we call it a decision director, - but what we're really talking about is, - we call, AI teammates are specialists or assistants, right? For a routine task, you want an assistant. Please, could you remove all these feature flags for me at this time? You know, tell me when it's done. Or maybe you need to, maybe it's not a decision director, - maybe it's like, get me a business analyst - who really knows these compliance guidelines - and can help me understand what this decision is about. So, we say assistant or specialist. And I say this to think like, as you bring AI in as team members, - what role are you giving them? What is their purpose? Where do they fit? How do they form up in your team? We are engaging in that process of building human and computer teams. By the way, not everything in Atlassian you have to pay for. I mean, I know Kelsey set us up there as like a great software slinger. This is publicly available for free practice. We live it. We breathe it. DACI is actually an industry term. It wasn't invented by Atlassian. We truly live it and breathe it. And the collaboration of who gets involved, - what you do, freely available. DACI play, Atlassian playbook. If you don't know about it, jump on it. It's really, really powerful. And we genuinely use it every day. Empowered teams, I started there because I believe in teams, - and there is no greater connection when you get a good team. Empowered teams. Got to empower them, right? It's just, we can't lose that as the wave of technology washes over us. Then, most people in this room I know probably work - in organisations where there's more than one team, right? And getting teams to work together and to collaborate, - this is the secret sauce. How can we help unlock whole organisations, - teams to work together, connected teams? There's an org chart. You're in this department. We all know that's kind of, okay, right, I know whose boss is who, that's great, - but actually, work flows across an organisation, - and it's constantly evolving, and it's changing, - and it's adapting to that, and how do teams adapt to that, - that causes cognitive load. It's a really hot area for us, - because you have to move as a whole, - you can't just move as one team, or indeed, one individual, right? AI productivity, it's going to make your productivity great. Excellent for the person. We need everyone to move forward because of this. So, how do we do it? We talk about networks of teams. The first thing we do is we start to identify - and thinking of that inner source example that Kelsey was talking to, - who are all the teams we collaborate with? So, we just start by laying out who all the teams are that we collaborate with. We use some colour coding to say, are they a software team? Are they like a service team, a support team? Are they like a leadership team? Just lay the land out for me. And then, what we start to do is we go, okay, - we've got the lay of the land, but then which ones - are really, really important to our success? And we start to organise by importance of who do we really depend on - to have a good relationship with, to be successful. And we actually get quite, dare I say the word, deterministic about this. We say, "Pick five." This is the fun bit. Pick your five most important teams - that are going to be critical to your success, your goals. And once you've got those five, have a really hard conversation. Do we talk to these people? Do we coordinate with them? Do we have a healthy relationship with them? Do we even know who they are? Do we want to maybe improve that or get it better? And so, that drives, we do this quarterly or six monthly, - run this process to come back and say, - okay, have we improved our relationships? Now, I've run this many times, - and I've run it virtually, and I've run it face-to-face. Works really well in quarterly, six-monthly cycle, excellent. Both are as good as each other, but there's one thing - that always stands out to me. This outer perimeter, you ask a group of teams, - sit down on a piece of paper, what teams do you work with to be successful? I am blown away in modern work just how many teams there are. And you look at it, and you go, no wonder everyone's burning out. Oh yeah, but there's this team, and there's that team, - and there's this team, and there's that team. It is really tough out there. And I can't help but think, - I've put these AI team members on the outer banks, - I don't have the data, I don't have the exact story to show you right now, - I'm a solutions engineer, I get a lot of questions every day - from a lot of different people, I am building a sort of a Chris, - I'm not going to call it a Chris agent, solution engineer agent, - so that other teams that are not in my sphere of critical importance - can still have access to me or how I think, what I know, my knowledge, - so that they can be progressive without necessarily having to come to me - and reduce the cognitive load on myself. It's early days, it's early days, but I will share this story - next time I see you on how that's going, - because that outer perimeter, and using AI agents - to give people what they need without necessarily raising the load, - is an important one. Network of teams, super cool, super fun play. You can take it really seriously, or you can use it as an icebreaker. I highly recommend it for opening up a lot of discussion about - how do we collaborate, how do we actually work. Alignment of teams. We probably should start here. I spoke about goals, goal independence, right? So, we've got our autonomous team, our empowered teams - that are connected, we're working together, - but are we all going in the same direction? So, alignment of teams, how do we do that? Coming back to our Jira Mac App, the true story, right? There was more than one team involved, particularly because it was, - we ended up, it was discontinued. So, you've got users that are affected. You've got to communicate. We've got marketing, we've got a whole pile of teams that were involved. How did they stay aligned across teams to achieve this? Well, they got a spreadsheet. And every day, on a Thursday, they'd update the spread-... It's a joke. It's a joke. I don't tell jokes as good as Kelsey. Hey, we've all lived those spreadsheets, right? I know some of the audience here has. We've all lived those spreadsheets, and we've all lived those pings. Can you give me an update on what's happening with the Jira Mac App, right? We've lived it. So, how are we doing it, right? First thing is common language. Got to have a common language, very simple language, - and maybe to take again, we all stand on the shoulders of giants, - but with Kelsey talking about reinventing the wheel, - maybe we are reinventing the wheel a bit here. It is intentional. So, we say common language. What is the common language? What are you doing? Why are you doing it? What will success look like? Who's leading it? And who are all these stakeholders - that want to follow it and have visibility to it? No matter who you are, no matter what sort of a snowflake you are in Atlassian, - those four things, what, why, who, how, - you've got to be able to communicate it, and it's got to be visible. But then, even more so, I want to dig in, - and I want to understand what's happening. We simply have an update process where on a Friday, - by Friday afternoon, you've got to put in 280 characters, - and yes, you can put a smart link off to some data, et cetera, - 280 characters into an update. You can see when the updates are flowing. On a Monday morning, those updates go out in a sequence. And so, you get a sort of a summary view - of everything that you're following and