Avani Prabhakar

Everyone’s adopted AI but now they’re asking about the ROI. At @Atlassian, we found our fastest AI teams and studied them. They were the teams increasing speed anywhere between 2-15x. We’ve called them our frontier teams and here’s what we learned 🧵

This is the Teamwork Graph (and doesn’t this experience look great). All your context in one place, so AI can actually help. We’re just getting started 🚀

The lesson for any team: the question isn’t how do we use AI more? It’s how do we give AI the context it needs to actually help us? Setup is the hard part, but it takes less time than you think - a sprint, a shared commitment, a few people willing to lead. 🔗atlassian.com/blog/how-we-bu…

The payoff: 91% reported a concrete change to how they work — not just awareness but actual workflow change 🔥🔥

Our People Insights team grew AI usage 562x in a single week by giving it something most teams skip: persistent memory of how the team actually works. Here’s what changed...

The teams pulling ahead right now are changing how they work and experimenting relentlessly. @Atlassian’s Rovo Growth team is a great example. In just six months, they ran 40+ product experiments and shipped 80% of them (more than double the ~30% industry average). That came from rewiring the way they work: tools like Cursor for rapid prototyping and AI bridging the gap between product and engineering so ideas move from concept to live in a fraction of the time. But speed didn’t replace rigor. The team only shipped experiments that produced a statistically significant lift on pre-set metrics for our customers. The takeaway: when experimentation is built into how you work, this is the kind of impact that follows.

Rovo MCP server just hit 1M monthly users- 5M+ tool calls every working day! The real story is who is using it though. Nearly half aren't software teams. C-suite leaders and PMs use it just as intensely as engineers. What started as a dev power-user tool is now how any knowledge worker connects their context and works at a whole new level. 👇 Watch ChatGPT review a Jira epic, spot missing steps and create the tasks in seconds.

We don’t have all the answers. These are working hypotheses - five 12-month bets our team is testing together. We’ll share what we learn along the way. Full thinking here: atlassian.com/blog/how-we-bu…

Belief #5 There’s a growing narrative that AI will replace large parts of HR. We believe something different. As AI takes on the admin, what remains is the work that runs on relationships. People don’t navigate uncertainty with a decision tree. They navigate it with the trust a human HR partner builds.

85% of knowledge workers are using AI. Only 29% have changed how they work. That gap is one of the biggest opportunities HR has ever been handed. We published five beliefs we’re testing over the next 12 months. Here’s what we think 🧵

Here’s a new capability that helps teams redesign how work happens You can mention @Jira in any Slack channel to create a Jira work item that reads the context of your thread to creates a structured work item ready to go

Love this thinking from @sherifmansour It shows how AI doesn’t replace the craft, it raises the ceiling for everyone. The best teams aren’t collapsing roles, they’re building a shared brain. sherifmansour.medium.com/direct-the-shi…