Key Takeaways

  • Startups try to hire AI-native PMs, but Atlassian trains their existing staff to become AI-native builders across products like Jira and Confluence.
  • Over 1,000 Atlassian PMs and designers completed quarterly Builder Weeks, creating more than 120 active internal workflows.
  • PMs are evaluated on a 1-to-5 skill scale across six capabilities, aiming for Level 3 ('Capable') across the board and Level 5 ('Pioneering') only where their immediate team needs it.
  • Product teams pause routine sprints for one full week every quarter to build actual AI tooling alongside external experts and peer coaches.
  • Atlassian operationalizes this shift using Atlassian's AI Fluency Index & AI Builder Week Upskilling Model.

The Atlassian's AI Fluency Index & AI Builder Week Upskilling Model

Atlassian Chief Product Officer Tamar Yehoshua rejected the common startup playbook of replacing staff to find AI talent. “When I talk to people at startups, they tell me hire people that are AI native,” Yehoshua explained. “Well, we have a lot of PMs already who are amazing PMs at Atlassian. So, we want them to become AI native.”

To build that muscle, Atlassian runs a structured internal framework:

  • AI Fluency Index Capabilities: Assess product team members across six core AI proficiencies: tool adoption, writing and conducting evals, automating data insights, rapid prototyping, AI agent configuration, and baseline technical/coding literacy.
  • Five-Level Competency Scale: Measure progress on a 1 to 5 scale: Level 1 ('Curious'), Level 3 ('Capable'), and Level 5 ('Pioneering'). Use this as a continuous personal development North Star rather than a rigid promotional ladder.
  • Fluency Targets: Target achieving Level 3 ('Capable') across all capabilities across the product organization over time, while empowering individuals to reach Level 5 in the specific capability most crucial to their immediate team.
  • Quarterly AI Builder Week: Pause routine product delivery for one full week per quarter. Structure the week with external guest experts, peer-led internal training by advanced PMs, and hands-on project work dedicated to building practical AI workflows.
  • Themed Curriculum Rotation: Focus each builder week on a distinct practical competency (e.g., Week 1: Prototyping, Week 2: Evals, Week 3: Building Agents, Week 4: Checking in Code).

Yehoshua noted: “And then we have an index from one to five. One is curious, five is pioneering, three is capable, and we do not expect every PM to be a five at everything. And this is not a ladder. This is for development.”

When This Works (and When It Doesn't)

This framework works in established tech organizations with strong product managers who need technical upgrades without disrupting sprint cadences. Halting regular product cycles once a quarter creates psychological safety for non-technical PMs to experiment. “Our AI builder week is once a quarter. We take a week where you're not doing your regular work and we train PMs and designers,” Yehoshua said. Over 1,000 employees participated, producing 120 new workflows.

The model fails if leadership treats the 1-to-5 index as a performance management tool. If PMs believe a Level 2 rating damages their compensation or promotion prospects, self-reporting breaks. They hide knowledge gaps instead of learning. It also fails in early-stage startups that lack the luxury to freeze sprints for 25% of a month. In a seed-stage team, PMs must already row by writing code and testing evals daily.

What to Do With This

If you run a product team, audit your PMs against the six capabilities this Friday. Pick one area where your team is weakest, such as writing evals or agent configuration. Block off next Thursday afternoon as a mini Builder Day. Cancel all sprint rituals, bring in an engineer to run a live demo, and require every PM to ship one working automated prompt or evaluation script before 5 PM.