⚒️ How we built Atlassian Design System CLI Sharing a great deep dive from Kylor Hall on our new Atlassian Design System CLI. It exemplifies how we build and stress test with the same considerations that our customers have for their own AI stacks. Token efficiency, MCP vs. CLI decision paths, evaluation loops - all of these are done with a ton of rigour. This takes deep collaboration between engineers, designers and agents over many months. Read more at: Bonus - in the spirit of building in public - anyone can use our ADS MCP and ADS CLI. Collectively, they provide 90% plus coverage of what is on

🏖️ Forget the plywood-framed podcast studio - beach interviews could be an Australian design institution! It was great to hear Tania Graham-Brown and @davidhoang talk about design questions that are global - and still influenced by what is special about places on both sides of the Pacific.

📣This will be worth a watch! @TYehoshua has a unique perspective, having led at companies that shaped our experience of collaboration, enterprise software and AI. Not to mention decades of stellar product thinking and leadership in technology generally. Seeing her weave that all experience together at @Atlassian has been so impressive. It feels like the scope and ambition of what we're building for better teamwork...is a perfect match for all that she brings. And most of all, I hope everyone watching @lennysan and her on stage gets a feeling for the vibe she brings. Tamar is a wonderful teammate to work alongside and to learn from. Wish I could be there, but I'll be cheering nonetheless! 💪

This week, our designer Marco walks us through the design challenge of making Dia on Windows feel familiar to Mac users and at-home on a Windows machine. Speaking of familiar... vertical sidebar for your tabs landed in the Beta this week! Get on the waitlist for first looks.

Hey, I’m Matt. I’m a Director of Product at Atlassian working to bring agents to the millions of builders who use Jira. Prior to getting into product I was a founder and engineer. Now based in SF and grew up in New Zealand 🥝 In an effort to bring more product voices to X, I’m going to start sharing what I’m building and how I think about product craft. If that’s of interest, let’s connect. My DMs are open and I’ll be answering questions in the thread.

Sharing this work from Maxwell Hathaway on the next generation Rovo Omnibar. It showcases the level of care that goes into our motion design work for the @Atlassian Design System. Those annotations and exploded views help him and the whole team inspect coded interactions from every angle. This is helpful because the Rovo Omnibar is one of those components that is a lot more complex than it looks at first glance. It's also a great example of using multiple tools and working with multiple designers (props to Deborah Lindberg and Dan Hoolahan as well). Plenty of care for detail and no one-shots.

🔬 Design principles for an agentic Confluence We've been exploring what "agentic" actually means inside a collaborative document, especially how we represent humans and agents working together. Here’s a behind-the-scenes walkthrough from Confluence designers Nat Johnson and Eden Gorton on their application of these principles: 1) Extend what already works - leverage a common design system for both humans and agents, so users do not need to relearn anything 2) Agents appear wherever humans do - this naturally involves a ton of trade-offs between visibility and UI density, but we're aiming to find that sweet spot for comprehension 3) Make it easy to 'Pop the hood' - you should be able to drill down and see the specific moment of a particular edit Agentic Confluence is a hugely powerful and exciting phase shift for teams, and I hope hearing from the designers communicates the care (debates and all) that is going into the change.

📔 Loom Speaker Notes You know that feeling when you go to record something and end-up re-doing the intro five times? Just me? Even though I am pretty comfortable speaking off-the-cuff, I reckon we all have things that trip us up when speaking to camera - and that intro line is mine. We’re always working on ways to make the recording process in @loom feel easier for everyone. Here’s an early look of how Marc Jenkinson is designing the details of Speaker Notes. Marc shows a bunch of thoughtful interaction decisions that ultimately helps teams communicate better - by enabling you to feel more prepared, be more authentic in delivery and stay in the flow of the message. All without the 'reading from a teleprompter' vibe that comes from your eyes darting around, or the "uhhh" pauses of frantic window management. I'm still working on the existential choice of "Hey" / "Hello" / "OK" / "G'day" - but Speaker Notes has me covered for the rest.

Some improvements to pinning tabs in Dia: - You can now drag individual tabs to the pinned area to start a new Tab Group - The divider line shows up on hover, rather than drop, increasing discoverability of the pinned area Shoutout to @Davidimel for the feedback

🎙️@mcannonbrookes speaks to an important point in this interview on @etnshow. We have been making software cheaper and faster for decades, and the net result of the efficiency curve has been...more people a) making more software and b) making it a better experience. Where the big efficiency gains made from AI will go in the Software Development Lifecycle - represents a choice. Every company will make that choice based on its' own unique context, so the outcomes are going to vary a ton. But I share the (increasingly widely held) belief that the waves of new AI capabilities means the 'ceiling is rising' in terms of what customers expect from the experience of software. And if you scroll deep enough across on-line design communities right now - that outlines of that new ceiling are emerging. Especially when it comes to designers really pushing limits on functionality like personalisation and proactivity - but also the level of expression that's possible through motion, input handling, sound and better graphics frameworks. The shape of the ceiling is not just about the new stuff that is possible to DO, it's about it being easier to create a sense of flow by increasing comprehension on WHAT to do next. I think @scottbelsky once wrote that "making something easier to create doesn’t make creating quality and originality any easier." So - on top of that all that new inventiveness in design - there (as always) needs to be the necessary curiosity, clarity and conviction about what to make as team - in order to actually get close to that ever moving ceiling. It's all about choices.

While working with external teams, we have to discuss bugs, edge cases, change in contract as requirements evolve etc. My team regularly uses slack for all these discussions. Having to organise the work discussed in the thread by re-reading the entire slack thread isn’t something software engineers are meant to do. These days I have been heavily using @Jira in slack that creates very accurate Jira tickets from the entire slack thread. There is a lot of satisfaction in using a tool that saves your time away from a mundane task.

At @Confluence, we are building for agents with an important key ethos: 🤖 the UI is the API 🤖 As a result, agents in Confluence can do nearly anything a human can, with seamless co-working from @claudeai @OpenAI @GeminiApp or any other agentic tool that uses MCP. But, let's be real: the UI still matters. And Confluence is the shared visual canvas for knowledge for teams of humans and agents. Today we rolled out: 📣 the ability to @ mention agents in Confluence documents and comments to assign tasks or have them co-edit directly in context 🧑‍💼 👨‍💼 agent avatars that show presence when they are actively working on a document 📈 enterprise-grade analytics on agents that have viewed Confluence pages and live docs to promote transparency Read about everything we launched today here: 🚀 much more to come - hit me with your biggest wishlist items for agents in Confluence!

🎙️ Giving Rovo a distinct visual dialect We’ve been working what it means to balance the scaling power a cohesive design language, along with the need for AI to have its own distinct "dialect" to help users build comprehension of the system state. I wanted to spotlight the work of Deborah Lindberg, who’s been exploring that space of intersection between the Atlassian Design System and Rovo. Two aspects that I particularly appreciate - first, is the how the team is executing motion to signal the "active listening" state - and secondly, how we're using variables to prototype different paths.

AI usage and productivity is up, yet only 6% of executives point to real ROI. Everyone is speeding up but not in the same direction. It's the "fragmentation tax". @fosslien, @molsands_, and I explore how the best teams are stay aligned to achieve group flow in this webinar.

As far as I’m concerned, this is what peak product design looks like when operating at the highest caliber. Incredible thoughtfulness over details, beauty, functionality, and most importantly: people. Put this in the MoMA, RISD, etc. Required viewing!

⚒️ Close to the metal - the engineering behind Rovo Long Horizon Sharing a great engineering deep dive to how we've built Rovo's Long Horizon Architecture to handle the future of long running and rich context agent work. For designers, these type of technical write-ups are very useful for informing how to build experiences that handle those type of tasks. Dig in to learn more about context compaction, adaptive reasoning and latency optimisation.

🔈 Prototyping the Sound of Atlassian Sharing some cool internal building from @stuthesmith and Eli Mishkin - where they talk through their perspective on exploring the sonic character of the @Atlassian brand, using a prototype that let's you compose using the core elements of our sound catalogue. You can see Stu Smith's musician background and playful vibe coming through in the proto 🎸. Messing around with the pads is just a bunch of fun. You can some make beautiful noise. And it's a good way for me to remember the range of sounds that we already have across the System of Work apps. I would call out that A11y is always a foundation in our designing, and sound should not be something teammates depend on to get work done. But where it makes sense, sound is both hugely evocative and helpful for interaction flow. The 'ding' of the Dia Morning Brief, the Loom recording tones...there is just so much territory here. Stoked to see Stu and Eli continue to explore what is possible. 🎶