Team US '26: Scale confidently with new AI trust and governance controls
|
This is the final installment in a 5-part series highlighting the biggest Atlassian Platform announcements from Team '26 US: The Atlassian Platform - Start your AI and cloud journey with the Atlassian Platform Teamwork Graph & Rovo - power AI with your organization's connected work context Platform apps - connecting the dots for your organization Administration - master the four shifts of modern administration 📍 Enterprise-grade infrastructure - scale confidently with new AI trust and governance controls |
|
TL;DR ⚡ AI is changing the game, and we’re introducing new AI trust and governance controls to give you flexibility as you roll out new AI-powered capabilities to your teams.
|
👋 Hi Community!
I’m Nat, and I’m a product marketing manager on the Enterprise Trust team.
Over the past few weeks in our Team 26 recap series, we've covered tons of exciting updates, from opening the Teamwork Graph to the broader AI ecosystem to providing you with a single app to visualize your work.
But here's the question I kept hearing from admins and security leaders at Team: "All of this sounds great, but how do I know my data is safe?" That’s a fair question, because none of these features matter unless you trust the foundation it’s built on. Your data needs to be secure, whether you're using AI or not, and having the right controls in place makes it a lot easier to say 'yes' to AI when you're ready.
To close out our Team ‘26 series, we’ll dive deeper into three main announcements across enterprise-grade infrastructure, the bedrock of it all:
-
AI trust and governance controls to keep data safe as you adopt AI
-
Data loss prevention at scale to help address security blind spots
-
Infrastructure investments to meet your strictest compliance and scale requirements and help you succeed on the cloud
Let's get into it.
🔒 AI Trust and governance controls
AI moves fast, and security teams are being asked to keep up without a playbook. If you're already using Atlassian Guard to manage security across Atlassian cloud, you've got a solid foundation, but now there's a whole new set of "what ifs." What if Rovo pulls in something confidential from a connected source? What if someone accidentally pastes credentials into a prompt? As AI adoption grows, these concerns — managing what it can see, who can use it, and whether it's used responsibly — become much harder to manage manually.
You shouldn't have to choose between enabling AI and maintaining control. That's why Guard Premium is going a step further with new AI features so you can decide what data your AI tools process.
Connector data security
Your team’s work lives across multiple tools - Google Drive, SharePoint, Slack, you name it. Connectors pull data from third-party tools into Rovo, but not everything should be pulled in. As your source connections grow, so does the risk of confidential data slipping through. It’s nearly impossible for teams to manually sift through years of accumulated content. So we’re introducing controls that work across your connected sources - not just Atlassian apps.
Guard Premium's connector data security will scan content in real-time to block data ingestion based on your content detection rules. For example, you can set a rule to block financial data or customer PII from being pulled in.
Rovo Chat security and security posture insights
Controlling what gets ingested is just one piece of the puzzle — what about what happens in the conversation itself? Your users chat with Rovo throughout the day, and they may accidentally share things they shouldn't - maybe they are troubleshooting an issue and accidentally paste AWS credentials into a prompt, or they ask a question that includes a customer's personal information.
Rovo Chat security applies the same policies you set at the connector level to the conversation itself, preventing users from adding or accessing confidential data in Rovo. For example, if a user accidentally pastes AWS credentials into a prompt, Rovo Chat security catches it before it's sent to an LLM. You can choose to block the message entirely or redact the confidential sections, so the conversation can continue without exposing that data.
How can you be sure your controls are working? That’s where security posture insights come in. You can pair it right alongside Rovo Chat security to help you track the impact and effectiveness of your AI controls, such as redactions made and content blocked. These insights show you what’s being caught and how often, so you can decide if your current rules need loosening or tightening, and make adjustments.
Enhanced audit logs
As an admin, it’s critical to know who's using AI, what data it's touching, and how effective your security policies are, especially if you have to prove compliance to auditors and regulators.
Enhanced audit logs will track Rovo interactions, such as who invoked it, what it did, which controls and policies were triggered, and what content was blocked or redacted, so you have a clear trail for compliance reporting and can fine-tune your controls.
🔍 Data loss prevention (DLP) at scale
Here's a scenario that might sound familiar: your org has been on Confluence for years. Thousands of pages. And somewhere in there, someone put customer PII or API keys that should've been restricted. You know it's there. You just don't know where. The more your organization grows and the more your data sprawls, the harder it is to keep track of it all. Security isn't just about what's happening today; it’s also about protecting the years (or decades) of data you already have.
Guard Premium already provides data loss prevention capabilities, including content scanning and data classification. Now we’re adding new tools to help you find risky behavior and content across your site and classify data consistently: full-site scanning, 60 new out-of-the-box detections, automatic data classification, and reclassification controls. These features will help you get a clearer picture of what data lives where, so you can classify it properly and confidently turn on AI.
To learn more about these DLP capabilities, check out this Community post.
|
⭐These DLP features will be generally available later in June, so keep an eye out! |
| 🎥 Check out our security breakout session: Calling all security admins: Your AI is only as safe as your data with Head of Information Security, Rob Bissett, and Sudhanva Ramesh from Synchrony Financial, as they discuss why data security is the fastest path to safer AI. |
☁️ Infrastructure to meet your strictest requirement
Whether you're navigating FedRAMP, GDPR, or your own internal security requirements, the need is the same: a secure, reliable foundation that adapts to change. Here's what we've been working on:
Security and compliance
Regulatory and security requirements keep evolving, and so does Atlassian cloud. For our German customers, we’re deepening regional compliance with upcoming data residency support for Loom and C5 Type 2 attestation, which is now available. For teams looking for options to further mitigate risk, Customer-managed keys puts control of your data encryption in your hands (paid add-on). For organizations with single-tenant requirements, Atlassian Isolated Cloud will be generally available later this month, providing a dedicated environment with isolated compute and networking for ultimate control.
Scale
We've raised the ceiling: Confluence now supports 250,000 users, and Jira Service Management scales to 100,000 agents per site. This means more of your teams can work in one place, sharing the same knowledge base and portal, and Rovo gets richer context across all of it.
Reliability
We’ve introduced capabilities to protect your data and to keep your sites available. Atlassian Backup and Restore (paid add-on) runs scheduled and ad hoc backups of your Atlassian apps with full audit visibility so you can recover from data loss in hours, not days. System Health gives org admins personalized, site-level visibility into the operational status of your cloud apps, so you hear about incidents faster, spend less time monitoring, and have the data when someone asks: 'How reliable is this?'
🛠️ Tools to help admins manage users and data
One thing we definitely did not hear at Team is that admins don't have enough work to do! As your organization scales, so does your to-do list. Here's what we're introducing to lighten the load:
User provisioning improvements
More teams on cloud means more permissions to manage, and if you’re doing this manually, it can be unsustainable and error-prone. Guard will soon offer user provisioning enhancements to allow you to control which users sync to which sites or Units based on IdP groups and attributes, and onboard and manage external collaborators via SCIM.
Read our administration recap to learn more about the new admin tools, workflows, and controls that are helping admins steer their organization in the AI era.
Bottom line: the foundation is getting stronger and more resilient, so you can scale your teams and AI confidently.
💬 We want to hear from you
That wraps our Team ‘26 US recap series. From enriching your AI with connected work context, to turning data into trustworthy insights, we’ve shown that you don't have to choose between adopting AI and keeping your data locked down — you can do both with the Atlassian Platform.
Now we want to hear from you. We know cloud security, especially with AI in the mix, can feel like a moving target:
-
What security or admin capabilities would give you the most confidence to scale AI in your organization?
-
Has your team started using AI tools like Rovo yet? If so, what was the first security control you put in place before turning it on?
-
How are you currently tracking confidential data across your Atlassian sites — manual audits, scripts, or something else? Would these new DLP features, like full-site scanning, change your workflow?
Drop an answer to one (or all!) of the questions in the comments below.
We’ll be picking our final two lucky commenters to win a free Atlassian certification voucher, so you can earn a credential and level up your career. The winners will be randomly chosen and mentioned in the comments in 2 weeks.
Was this helpful?
Thanks!
Natalie Koh
0 comments