Behind the scenes of Atlassian’s AI Builders Week

At Atlassian, we are often asked how we are adapting internally to AI and transforming our R&D organization to be AI native. One lever we have used successfully is an event we call AI Builders Week. Each quarter, we bring together hundreds of product managers and designers to learn new AI skills and land tangible changes within a few days. We use this dedicated time to experiment together and embrace the change that AI presents, so our customers can thrive in this new era. When we build for ourselves, we’re also building the playbook for every team.

Our first AI Builders Week focused on establishing the foundations. We grounded teams in responsible AI use, built early prototyping skills, and introduced basic agent workflows and evaluation practices. The goal was to help PM and Design crafts gain familiarity with AI, understand its potential, and learn the basics.

Our most recent week was about taking it a layer deeper, with the goal of driving measurable impact. 77% of our Product Management org and 80% of our Design org attended a highly technical curriculum designed to help them use AI to make day-to-day work easier. We focused on advanced agentic workflows, production-ready prototyping, replacing time-intensive tasks with agents, and improving both evals and the golden data sets that power them.

One thing that made this week especially impactful was connecting the dots from ideation to production through real, working prototypes. With support from our Design Technologists and Engineering leaders, we helped Product Managers and Designers get closer to the environments engineers use every day.

As AI collapses the boundaries between crafts, there are general skills that all builders should have. We are actively enabling our teams to succeed in that future with teams across crafts learning how to commit production code and turn ideas into experiences customers can actually use.

How did it go?

• 298 Agents were built. We saw a jump in AI intensity with super users increasing usage by 147% and average users by 175%. Plus a 15-20x spike in daily usage for Rovo Dev, Atlassian’s coding agent.
• 96% of attendees gave positive ratings for AI Builders Week.
• 95% said they left the week with concrete changes they can implement right away.

Read on for more details on how we design and execute our AI Builders Week.

Overarching AI Builders Week Design Principles

Running a global, hands-on program like this takes close coordination across program managers, craft experts, mentors, customers, external AI leaders, and IT to design content and help attendees get the most out of their experience.

We structure every event with a few principles in mind:

Learn through hands-on building
To ship real changes, attendees form teams to create tangible artifacts, including prototypes, workflows, templates, processes, and demos. Virtual mentoring and technical support are available so teams can quickly get guidance as they build. The Build Day gives visibility into what’s possible while showcasing real solutions that can immediately be remixed or adopted by anyone.

Be customer zero for Atlassian AI and dev tools
We make our team the best possible “first customer” of Atlassian AI and developer tools — dogfooding them deeply to improve both our products and how we work, while adopting the latest AI capabilities ourselves.

Build practical workflows to shape the future of teamwork
The goal is to move beyond basic or theoretical AI use cases to build end-to-end agentic workflows that automate real work and enable new ways of collaboration. Much of our curriculum is advanced and technical, and we recruit internal volunteers to help everyone push through challenges.

Experiment boldly and learn fast
AI Builders Week is designed to help teams move from experimentation to real, measurable impact. We encourage teams to openly share what they learn – successes and failures alike. Our capstone Build Day showcases everything from breakthrough ideas to failed experiments that generate valuable lessons.

Kicking off the week with new perspective

We start each week hearing from an array of AI experts on everything from where the industry is going to what tools work and the best hacks. The goal is to foster a sense of excitement and visionary thinking that feeds into the rest of the week. In our most recent event, we brought in groups across:

External thought leaders
A huge thanks to Ravi Mehta and Anil Sabharwal for sharing their time and wisdom with Atlassians about the AI era. Ravi shared a great metaphor:

Anil challenged Atlassians to embrace our uniquely human role: defining what “good” looks like. In a future where humans and AI are teammates, human judgment, empathy, taste, and creativity turn powerful pattern recognition into experiences people trust and love.

Internal AI early adopters
We asked many of our early adopters to demonstrate what is possible and already happening within various teams across the company. The beauty of celebrating our internal experts is that they make complex AI concepts accessible to their teammates AND their tips are immediately applicable to our environment.

Our customers
We invited several super-user customers to share the powerful agents and Rovo workflows they’ve built. Additionally, a number of customers partnered with our Product Managers and Designers to build, test, and iterate on ideas in real time. One of the most popular sessions came from teams prototyping and quickly gathering live customer feedback. A huge thank you to the customers who partnered with us to make this possible.

Sample AI Builders Week Agenda

Here’s a look at our actual agenda and some details on the curriculum. I hope this content helps inspire or gives you a starting point to remix for your own internal AI enablement.

More Detail on the Curriculum

We designed AI Builders Week to be a technically challenging week, balanced with sessions that include hands-on mentoring and shipping real AI naive workflows. Here’s a bit more detail on what we captured in our advanced skills & the real-life application that resonated most with our attendees.

Prototyping → Production

Teams explored how rapid prototyping environments – such as Replit, DevBox, and Cursor – can accelerate concept-to-code workflows. These environments were paired with our design systems to ensure prototypes remain consistent with Atlassian production standards.

Sessions also introduced the fundamentals of modern developer tooling, including IDEs, development environments, and how tools like Bitbucket support collaboration and deploying prototypes.

Teams learned how to:

• Set up a remote development environment without needing to install code locally
• Run a development server to quickly test and iterate on ideas
• Use Rovo Dev to explore, modify code, and ask questions about Atlassian’s front-end monorepo

For example, a Loom growth designer used Cursor and Rovo Dev with a Loom Blueprint Bitbucket repo to build a live, code-based prototype for Loom’s “Make Edits” panel. The prototype allows them to fine-tune hover animations in real time using on-screen sliders (e.g. duration, opacity), instantly see changes in the browser, and push updates directly to Bitbucket.

This shifts how teams work: instead of back-and-forth with engineers over small interaction details (like “80ms vs 120ms”), product and design can own and refine these decisions themselves moving closer to shipping production-ready interaction tweaks directly from prototypes.

Continuous evaluation of AI quality

Teams used Rovo Studio to run no‑code evaluations on their agents: importing real user questions, setting up automatic pass/fail criteria, and identifying where agents hallucinated, missed context, or ignored instructions.

Automating workflows with MCP integrations

MCP-enabled sources unlock so much opportunity to automate everyday workflows. In this hands-on session, we helped teams automate common workflows including:

• A feedback agent that aggregates and shares customer insights
• A decision update agent that keeps stakeholders aligned
• Weekly digest agents connected to third-party sources
• PRD from a JPD backlog (in under 10 minutes!)
  

Using Forge & App Builder to build powerful customizations

Participants got hands-on with Forge and the new App Builder (currently part of our Studio beta) to extend Jira, Confluence, and JSM – building use cases like custom fields, macros, panels, full-page apps, and automation skills. As customer zero, Atlassians had the chance to shape how these tools evolve while exploring new ways to build on our platform.

AI Builders Week: Living and Breathing the AI Future

At Atlassian we believe you have to be the change you seek. If we want to build the future of teamwork for our customers, we need to live it ourselves first. AI Builders Week created the space for teams to experiment with agents, evaluations, rapid prototyping and getting closer to code so we can bring the best patterns directly into how we build and ship for customers. I hope sharing a bit more about how we bring this to life for our teams can help others on their AI transformation journey.