Atlassian puts HR at the centre of AI transformation

Most enterprise AI rollouts follow a predictable choreography: the CTO picks the platform, engineering builds the proof of concept, and HR arrives months later to manage the change fatigue. Atlassian decided that order was backwards. When the company began its AI journey three years ago, it seated the Chief Information Security Officer, Chief Technology Officer, Chief Information Officer and Chief People Officer at the same table from day one. Eighteen months into that cross-functional effort, the group reached a shared conclusion: the transformation would succeed or fail on the people side.

That conviction led Atlassian to expand Avani Prabhakar's remit. As Chief People and AI Enablement Officer, she now owns AI adoption as a company-wide redesign of how work happens. The title reflects a belief the founders hold: AI is fundamentally a people-led transformation. On Workday's Future of Work podcast, Prabhakar told Workday Chief People Officer Ashley Goldsmith that the shift changed what both HR and technical teams do. HR evolved from writing acceptable-use policies to designing human-AI collaboration across the entire employee lifecycle. Technology teams moved from deploying models to redesigning workflows and building the context layers that make AI genuinely useful.

What changed when HR led

Prabhakar described the early phase as security-first: the cross-functional group tackled guidelines, then workflow questions. The pivot came when the team recognised that adoption, capability-building and work redesign were inseparable. Putting HR in charge of enablement meant the company could address three problems at once: which roles need AI fluency, how to measure productivity when agents handle routine tasks, and how to redesign jobs so people spend time on work that requires judgement.

The structural change also forced technology and people functions to share accountability. If an AI agent fails because employees lack the context to prompt it well, that is both a training gap and a design flaw. If a team ignores a new tool because it does not fit their workflow, that is both a change-management failure and a product-integration miss. Atlassian's model makes those tensions visible early, rather than discovering them after a failed pilot.

Bottom-up builders, not top-down mandates

Rather than issuing AI quotas or usage targets, Atlassian ran internal "AI Builder Weeks" and cultivated super-users who experimented with custom agents. The approach produced tools that employees actually wanted, because the people closest to the work designed them. Prabhakar's comments suggest the company values organic adoption over compliance metrics, a stance that aligns with its broader culture of autonomous teams.

The builder weeks also surfaced a capability gap. Employees who understood their domain but lacked prompt-engineering fluency built agents that worked in demos but broke in production. That gap clarified HR's role: help people articulate their work clearly enough that an AI can assist with it. The skill is less about code and more about decomposing tasks, naming edge cases and specifying what good output looks like.

What this means for L&D and transformation leaders

Atlassian's blueprint offers three lessons for organisations that want AI adoption to stick. First, seat HR and technology leadership together from the start, with shared goals and joint accountability. Second, design enablement around workflows. Training that teaches people to use an AI assistant in isolation will fail if the assistant does not integrate with how the team already works. Third, measure capability. Track whether people can redesign a process with AI.

The model also clarifies what "AI literacy" means in practice. It is the ability to recognise which parts of your work an agent can handle, how to give that agent enough context to succeed, and when to override it. Those are organisational design skills, and they belong in HR's remit as much as in IT's. Prabhakar's expanded title recognises that the hardest part of AI transformation is helping people do different work.

---

Sources:
Atlassian blog