Turning connected context into AI-powered acceleration with Teamwork Collection | Team '26 Europe

0:08 Please welcome Atlassian senior vice president teamwork collection Sanchan sex center. Hello everybody and welcome to the teamwork collection keynote. Somewhere in your company right now, someone just had a brilliant idea. What should happen next is that idea should reach a bunch of other people in the organization. And those people should work together to connect that with other ideas in the organization.

0:41 They should work with other team members and take that idea to its full potential. But in reality, it doesn't happen like that. In fact, the knowledge in the organization stays siloed. People don't have the right tools. They don't know how to collaborate across different teams. And business outcomes take longer than they really have to. I'm Sanchin Saxa and today I'm going to tell you how any and every idea in your organization can actually reach its full potential.

1:09 To help me bring all of this to life, I'm going to invite my colleague Claire Drummond to come on stage with me. Thank you, Sonin. And hello everyone. So, every company here has the same models that we do. Nobody wins on the model anymore. You win on what it knows about you. And right now, it doesn't know enough. Your context is scattered across apps, stripped of the why. So, your agents are guessing. And guess what? Guessing is really expensive.

1:52 The teamwork graph gives your agents a map to the brain of your company. A map that gets richer every day across 250 billion objects and counting so that your agents can surface the most relevant answers and build the best solution. It connects it all. the people in your organization, the tools that they use, and the data that informs their decisions.

2:23 The context is yours alone, and it compounds every day. Your ideas are your competitive advantage. And the graph is how you build on them with AI. Your competitors can rent the same model, but they cannot rent your context. >> That's right. And in the last year, we have given every organization the ability to act on that context. In fact, we have enriched the teamwork craft significantly and added close to 250 billion objects in that graph. We have brought agentic workflows deep into every team's workflow. And last but not the least, we've built it in such a way that the smallest of organizations to the largest of enterprises in the world can continue to scale on our platform.

3:13 Speaking of the largest organizations, I see a lot of customers in this organization in this audience that I know are pretty large. U I'm excited to announce that we can now truly accommodate any team of any size in your organization. We've removed all user tier limits from Jira and Confluence. You can scale as much as you want. >> I am so excited about that update for our customers. But here's the part that I love the most about what we're building. You do not have to imagine it.

3:43 It is already changing the way that our customers work every day. But don't take my word for it because for the second year in a row, Gartner has recognized Atlassian as a leader in the magic quadrant for collaborative work management. And from the world's most innovative companies to the world's most established, our vision to deliver human and AI collaboration on top of the best, most nuanced context graph is already in the hands of our customers.

4:19 In fact, customers are reporting that they are able to save the most precious thing of all, time. Specifically, Robo Search and Automation enabled our customer Procore to save an average of 23 hours per person every month. And we estimate that Robo saved Atlassian customers almost 2 million hours in August. >> I was sold. >> It didn't matter though because everybody was on vacation. Today, we're going to show you even more innovations to help your teams move work forward. We're going to show you how any kind of team whether it's in marketing like yourself in product like myself or in sales or any other organization can benefit from the teamwork collection.

5:04 Now before I show that CLA let's tell the audience how any big project at last generally starts. >> Oh yes. Uhhuh. Usually with a long crowded meeting at a very inconvenient time and they're usually led by you. >> What I call that many late night meetings. Okay fine. I do call some late night meetings but what if just hear me out what if I don't do that and instead what do I do I try something new what if I record a loom and tell the team everything I want them to focus on I am just going to press play and I'm going to tell them what my strategy is what vision we are trying to achieve what should be our goals you know normally I'll have to create a deck or a presentation right to tell them about that but this time around I just want to record a loom and when I do that here's the loom that I pre-recorded, by the way. when I do that, I think I should look cool before I send it out, right? Yeah, I don't want to just send it out like that. So, what if I just hit generate AI overlays, and I ask my trusted AI to create some cool animations for me. When I do that, it allows me to choose a bunch of things like the color I like, which matches my personality, the font, etc. And then I can just hit generate overlays. Now, are you ready to see the final product?

6:19 >> I am so ready. >> All right, let's take a look. Tada. Look how cool I look. AI Overlays was able to automatically create visuals on top of my video. The visuals highlight the most important beats that I want to with my team. Right. And look, I can now share it. >> And normally this would have taken you hours. Well, especially you. >> Yeah. >> But what's better is that the AI highlighted the most important parts of the Loom so that the team didn't miss a single thing and you can share it immediately.

6:50 >> Yeah. And remember one thing, Claire, I did not call that late night meeting. >> Oh, I'm so thankful. So, as you saw in the founder keynote, we are excited about the ways that Loom is fundamentally changing the way that people work, how they have meetings, how they communicate with each other and their agents. And if you come by the Loom booth, you can sign up for early access to AI overlays today. So now that Sonchin has shared this loom with the team, it's time for me, well Rovo and I to get to work >> finally.

7:24 >> Okay, no more excuses. So as a marketing leader, I really love to build my marketing plans in Confluence. It's where my team can collaborate on the ideas and the strategy together. Confluence is a powerful tool to centralize knowledge, but it's also a powerful collaboration hub where we can brainstorm on whiteboards or we can create polished presentations in slides. So, let me show you what I mean. While I'm in Confluence, I'm going to hit create with Robo and then I'm going to paste in Sonchin's Loom. I'm going to ask it to create a marketing plan based on the details that Sonin just shared. It also will pull more context for the teamwork graph and it will build a full plan. It knows my team, who I work with, what my goals are, and what else we are working on.

8:21 So, I'm going to go get a coffee and Robo is going to build me a robust launch plan. Coffee. This is so powerful because I'm not just creating docs with AI. I'm immediately building contextually rich artifacts that my entire team can collaborate on right away. So, I'm going to give the doc a review first. Of course, AI doesn't always get it right on the first go.

8:54 >> That's right. And I'm going to scroll through all of the parts of the plan just to make sure they're there, like the outcomes, the work streams, who the owners are, the operating model, and even it even has a built-in messaging framework for me. And then I'm going to add it to my Confluence space. This way, my team who have permissions to review the content in my space will be able to access it right away. >> And that's true multiplayer.

9:21 >> Absolutely. And next I want to add some vis visuals to show the team the design concepts for the website that we are going to build. So I'm going to go through here and I'm going to highlight the text and I'm going to appmentntion Figma directly into my plan. >> Our trusted design partner. >> Yes, we love Figma. Figma agent generates the image and then inserts it directly into the doc. Figma agent creates the image and then inserts it into the doc. So think about how many steps I have just skipped with that single prompt.

10:01 >> Yeah. >> What do you think? >> I I think this looks great. But I have some feedback. >> Oh my gosh. Of course you do. >> Always. >> I'd expect nothing less. >> This is true human and AI collaboration with agents directly in my workflow. Our customers are calling on their agents to do tasks like this more than 14 million times a month. So you can appmentntion and tag your agents directly in almost every surface you work in just like you do with your teammates.

10:29 So now the true test I need to build a presentation for the sales team. Anybody out here who is in marketing knows there is always a presentation for the sales team. >> Always. So, while I'm still in my marketing plan, I'm going to ask Robo to build me the slides. Rovo knows about the marketing plan. It knows about the project details. It knows about the goals. It knows from Sonin's Loom anything else that's related to this project. And it builds me this beautiful custom slide presentation that I can immediately share with my team. Now, let's just take this in.

11:09 It has the same visuals that Sonchin's Loom had as well. I'm gonna go ahead and further refine this directly into Rovo. That is amazing. I love every piece of that. So, what you just saw that how we are bringing agentic workflows in every team in every workflow where they are at. And by the way, it's not just ROS's agents in Jira and Confluence. If you use third party agents like Figma, Claude, Cursor, they all work the same way. And that's the power of the AMP protocol that we just announced today as well.

11:41 >> That's right. And we're not just creating more AI slop. These are quality outputs that I can actually use. Okay, Sanchin, I have done my part for this new project. I built the marketing plan. I've already finished my sales deck, which was probably the most difficult of the tasks. What about you? What are you going to do next? >> I guess you're going to make me work, too. All right, I'll get to work now. So, so far what what we've shown you is that I as a as one of the individuals in the team was able to communicate my plan without calling a late night meeting but through Loom and then Claire in marketing took that message and took that vision and and strategy and was able to turn that into manageable tactical items like a marketing plan, a slide deck for sales etc. Now everything she did and I did was in Atlassian surfaces but I know just like me many of you like other agentic surfaces as well.

12:29 For example, I use cloud all day long. So the next part of the demo, I'm going to show you the exact same thing but using Atlassian MCP in other agentic tools as well. All right, so let's do this. I am going to use cloud to do three things. First, I want to connect it to Atlassian MCP. Next, I want to ask it to create a spec for the website we want to launch, right? And then I want to pull that context from teamworkcraft so that the content that cloud produces is really really rich and and exciting for us to use. All right. So to do that I am going to move over here and jump into cloud. So let me go back here. So if you look at my screen here I have cloud opened up over here. And instead of typing what I want to do is use the voice prompt. By the way I don't like typing. I like like voice a lot.

13:24 Hey Claude, why don't you create a Vantage product spec for me in Confluence with many subpages for the different features like home, benefits, maybe the card studio, FAQ, plus can you create a whiteboard for the journey map for the customer journey and also a database with all open action items that I must get approval on. Once you're done, create one Jira Epic per page each with a single task. write it fresh and check any real numbers against the source materials in teamwork rap.

13:57 I'm going to pause it otherwise it's going to listen to my jokes as well. And I'm going to hit enter now. Now to cloud is going to take some time here. So I'm going to switch over here to a pre-recorded video of what Claude is going to do. Generally takes a little bit of time. so I'm going to show you what it actually did. And this is the loom that we recorded of what Claude will do. So here I'm going to press play on the loom. And what you will notice is it does a couple of things. First, it connects to Atlass and MCP. Then it's able to process all the source pages from teamwork graph. Next, it creates the confluence docs that I wanted it to, a whiteboard, a database, and also the Jira work items. Now, in the end, it created the final spec that I wanted it to do, but it use all the context in team for graph to be able to do that.

14:40 So, shall we take a look at the spec? >> Let's take a look. >> All right. I'm going to click on it and jump into confluence. As you can see, the spec is pretty rich. Why is it so rich? Because it was able to reference all the material that is already sitting in teamwork craft. And I'm going to show you in a bit what that material was. Not only that, it created as I instructed it the different specs for the different parts of the website that I want to build. The homepage, the benefits page, as well as the frequently asked questions. And by the way, remember these because when I show you the final product, I'm going to walk you through all this product. Now, I asked it to create one more thing which was the whiteboard. and it's able to reference the user research we have done about our customers because that link is stored in teamwork graph and it's able to create the whiteboard map using that as well and last but not the least I need to track all the open items that I need to get approval from right and all those items need to be tracked very easily so I can go to the right teams and get the approval as you can see as you can see all those things were created easily now in this demo you can see it also listed the owners so here the owners are in finance team. Some of them are in legal and compliance team. Now, let's take a look at what Claude used to actually create this document. I'm going to click in Confluence on the context tab. When I click on it, it shows me all the documents it referenced from teamworkcraft to actually create this output. Not only that, it tells me the people who are working on it and the agents who helped create some of this content. And the best part, I can visualize this all. I can see that the doc that I asked it to create referenced multiple other confluence docs from teamwork graph the jira work items the loom and it knows that sunshin created some of these things as well all of that was possible because of the power of teamwork graph in teamwork collection >> yeah this is truly remarkable >> you basically went from a voice prompt to a detailed product spec and database similar to how I built my marketing plan but from one of your favorite surfaces in cloud.

16:42 >> So Claire, let's tie this all back together for the audience. All of this was made possible because of the power of new and updated Atlassian MCP server. Please give it a try. It gives your teams permission and secure access to all the data in Confluence Jira teamwork graph very easily. >> I know for myself that the type of work that I do every day is changing. I personally want to get my hands on all of the latest and greatest of pretty much everything. I really want to push the limits of AI in my workflows. The improved Atlassian MCP is magic for that. For example, I had Rojo write this very talk that I'm doing right now. So, we've seen some really fantastic workflows for all types of teams. And when you look at the broader AI landscape today, much of the conversation is still stuck in multiplayer mode. one person sitting with a chat prompt generating isolated text or code. But the real business impact is multiplayer work happens everywhere.

17:43 >> That's right. Exactly. And also remember we're many of our customers are already using some of these things. The Atlassian MCP confluence context teamwork graph etc. Let's take Rivian as an example. Rivian has saved over 750 hours annually per person in the company. That's a lot of hours saved that can then be deployed against better quality work or better outcomes for the business as well. And it's not just Rivian. There are hundreds of thousands of customers around the world who are using teamwork collection and teamwork graph to power business outcomes as well.

18:17 >> That's right. So Sanjan, so far we've had some great ideas and concepts, but what happens next? Well, next we get to build the website and I want to show you what Claude did in the first time when it created the Jira work items. If you recall, I did not only asked it to create the confluence docs, but I also asked it to create a Jira work item. And what you're going to notice over here is I'm going to pick one of those work items that is about designing the website. And I'm going to just move it over to in progress. When I do that, the GR coding agent picks it up automatically. I'm going to click on that work item, expand it to see what Jira coding agent is doing. I can very easily see that it's busy working. But let's expand that to see what exactly it is doing.

19:02 As I see the agent is looking up the code repository. It's checking all the automations. It's pulling in the context from teamwork graph and it's writing the actual code that I asked it to do. And in the end, it generates some output for me. And one of those is a PR spec as PR link as well as the actual website as well. So let's take a look at the PR and the website. So I'm going to switch over.

19:28 >> I think I just want to see the website. I don't know that I care about the >> Okay, fine. We won't show you the the PR instead. Okay, so let's switch over here. Can you guys see my screen? Perfect. So here I am on the website and the website has the benefits tab as I asked it to create. There's all the details. It has the FAQs that I asked us to create and it has the form for applying for this credit card as well. All of that created from the spec that it first created through the Jira work item and confluence as well.

20:00 >> So, this is not just demo magic. The Jira coding agent actually developed this live website that he's clicking through right now from that product spec. >> That's right. And then one of our favorite features is >> I would love to see the card studio. The card studio allows you to customize the credit card. And you know, I'm interested in that for various reasons. >> There you go. So here you can customize the credit card that you want to receive and you can select the color, the name that you want. So Claire, what's your favorite color?

20:28 >> red, obviously. >> All right. So I'm going to get you a red credit card delivered to your home very easily. >> That's amazing. This is this is incredible. So, I think it might be one of the most exciting things I've seen today because it's really about taking something that was a spoken word using the teamwork graph and nothing gets lost in between. And this was all thanks to the Jira coding agent I just spoke about which is now generally available.

20:57 Powered by the Atlassian teamwork graph, Jira coding agent leverages the organization's context while it builds and it keeps your work visible, while keeping it connected to Jira. >> That's right. >> Okay, so let's zoom out for a second. We've just shown you a lot of great value today. We announced the removal of user tier limits so any size company can scale. We showed you some fantastic capabilities like AI overlays in Loom helping power the spoken word into valuable outputs.

21:31 We showed you agents and Confluence working side by side with human teams. >> And we also showed you the power of the new and updated Atlassian MCP server that connects your tools where you prefer to work. And also the agentic workflows are in Jira like Jira coding agent and other agents are also in there to turn your context into action. Thank you so much for spending time with us to learn about all of the amazing innovations that we're doing for every type of team especially through the teamwork collection. Thank you all so much and have a great rest of your team.

22:05 >> Thank you.