Atlassian Put AI Under the CPO. That Matters More Than the Layoffs.
On 11 March 2026, Atlassian made global headlines for the wrong reasons.
Six weeks later, they made one of the most important — and quietest — moves in enterprise AI strategy this decade.
The story has barely been told. So let me tell it.
What actually happened
On 28 April, Atlassian’s Chief People Officer, Avani Prabhakar, was given a new title: Chief People and AI Enablement Officer.
Her team grew overnight from roughly 700 HR professionals to around 3,500 people — including engineers, data scientists, and the entire internal customer engineering function.
Her stated rationale was disarmingly simple: the company had no shortage of AI technology, but teams were struggling to turn that technology into outcomes. Her positioning was even more pointed — AI is a cultural transformation first, and a technology shift second.
If you have followed my work for any length of time, you will recognise that sentence. It is the thesis I have been arguing from stages and in boardrooms for the last three years.
Why this is the right call
Most enterprises facing AI transformation install a Chief AI Officer somewhere in the technology stack, or hand the brief to a CTO who already has too much on their plate.
That structure assumes the bottleneck is technical.
It is not.
The bottleneck is human. It is comprehension. It is willingness. It is whether the people you have asked to “use AI” actually understand how it changes their work, their value, and their future.
Atlassian’s own data confirms it. Their State of Teams 2026 research shows only about 14% of teams globally are at what they call the strategic collaborator stage — the point where AI delivers roughly four times the productivity of basic use.
That gap is not a tools problem. It is a capability and culture problem.
Putting that work under the person who already owns culture, capability and change is the most logical move I have seen any large enterprise make on AI in the past eighteen months.
It is also exactly what I teach. The frame I use is comprehension over compliance. Mandated adoption produces performative use. Genuine adoption requires people to understand what AI does, what it cannot do, and where their judgement still creates the value the organisation pays them for.
You do not get to comprehension through a download mandate or a vendor demo. You get there through capability building, work redesign, and trust.
That is HR work. Or, at the very least, it is now.
What this means for every other organisation
If Prabhakar’s experiment works, expect this org-design pattern to replicate across the Fortune 500 in 2027 and 2028. The CHRO becoming the workforce architect for the AI era is not a quirky Atlassian decision. It is the leading edge of a structural shift.
Boards should be asking the question right now: who in your organisation has the authority, the credibility, and the remit to make AI transformation actually land in workflows?
If the honest answer is “no one yet,” that is not an HR problem.
That is a strategy problem.
But the harder conversation has not happened
Here is where I have to be direct.
A genuinely thoughtful AI strategy does not erase a deeply unthoughtful redundancy process.
The 1,600 people made redundant on 11 March received the news by email, after a pre-recorded video from the CEO. They were given limited time on Slack to say goodbye before their access was cut.
Atlassian’s first published value is “Open company, no bullshit.”
Trust — and every other currency I track in my 9 Currencies of Choice® framework, developed from more than 5,000 exit interviews — is built in the easy moments and tested in the hard ones.
What that morning tested, it largely failed.
It is possible to make difficult decisions. It is possible to right-size a workforce. It is possible to redirect headcount toward AI capability and enterprise sales. None of that is the problem.
The problem is the gap between the values printed on the careers page and the values experienced by the people walking out the door.
The sequencing was backwards
Here is what most observers have missed.
The CPO expansion happened in late April. The layoffs happened on 11 March. The order of those two events tells you almost everything about how the process was managed.
Imagine the alternative.
Imagine Atlassian had elevated Prabhakar’s role twelve months earlier. Imagine the company had spent a year publicly building out an AI capability function under HR — naming it, resourcing it, telling the workforce what it was for. Imagine leadership had spent that year having an honest, ongoing conversation with employees about what AI was likely to change, where the company was investing, where the skills mix would shift, and what reskilling pathways existed for people whose roles were exposed.
Imagine, in other words, that the strategic narrative had been built before the difficult decisions, not assembled afterwards in response to them.
The outcome may well have been the same. AI is genuinely reshaping the economics of software companies. Some redundancies were probably unavoidable.
But the process would have been profoundly different. People would have seen it coming. They would have had time to make choices, take training, look internally, or look externally on their own terms. The people who stayed would have understood why. The people who left would have left with a story that made sense, rather than one they had to construct from a pre-recorded video.
That is the difference between a workforce that has been led through change and a workforce that has been on the receiving end of it.
It is also, almost always, the difference between an organisation that emerges from disruption with its trust intact and one that has to rebuild from zero.
What better would have looked like
Almost none of this would have cost more money. Most of it costs leadership time and discomfort. Both are usually in short supply during a market correction.
The CPO/AI architecture in place first, not last. Build the capability function, name the leader, fund the work, and tell the story — before the workforce decisions, not six weeks after them.
A transparent ongoing narrative about AI and the workforce. Not a single town hall. A consistent, public conversation throughout 2025 about where the company was investing, what was changing, and what it meant for skills and roles. People can handle hard truths. What erodes trust is being told nothing, and then told everything at once.
Live communication from the CEO, not a pre-recorded video. People making the hardest transition of their careers deserve to hear it from a human being, not a file.
Direct manager conversations before the company-wide email. Affected employees should not be learning their fate from a forwarded link.
An honest acknowledgement of the October 2025 statement. Mike Cannon-Brookes told a podcast audience that Atlassian would have more engineers in five years, not fewer. Five months later, more than 900 engineering roles were cut. The contradiction has not been addressed publicly. Silence is not strategy. It is the absence of one.
A meaningful runway on system access. Long enough to say goodbye properly, exchange contact details, and exit with dignity intact.
A clear public articulation of the capability investment for the people remaining. Not “we are funding AI.” The specific reskilling, redeployment, and growth commitments for the workforce that has been asked to carry the company through the transition.
Here is the long-term cost of skipping that work: the people who stay are watching just as carefully as the people who left. So are your customers, your investors, and the talent you hope to hire next. Atlassian’s Glassdoor score has already dropped 12% year on year. The NLRB case against the company, over the firing of an engineer who criticised the CEO at an internal AMA, remains unresolved. The reputational tax compounds quietly.
The leadership challenge, named
Atlassian’s strategic direction is, in my view, largely right. The CPO move, the Browser Company acquisition, the agent governance investments, the bet on workflow data as a defensible moat — these are coherent, courageous calls.
But strategic clarity and human leadership are not interchangeable. You need both. And you need them in the right order.
The companies that will look back on this decade and feel proud will not be the ones that adopted AI fastest.
They will be the ones that built the capability and the conversation before the cuts — not after — and adopted AI without spending down the trust they will need in order to actually use it.
That is the leadership challenge. It is finally getting the structural attention it deserves — even if the conversation about how we got here is still missing.
For C-suite leaders and boards: If you are sitting with the question of who in your organisation should own AI enablement — or whether your AI strategy is built on the capability and trust it will need to actually land — that is a conversation worth having before the difficult decisions, not after. I work with executive teams and boards to build the workforce architecture, communication strategy, and leadership capability the next phase of AI adoption demands. Reach out at kim@kimseelingsmith.com.
For conference and event organisers: If your 2026 or 2027 program is grappling with AI, the future of work, or leadership through disruption, the Atlassian case is the most teachable real-time example I have seen of all three colliding at once. I am currently booking keynotes built around this case study and the broader frameworks behind it — Future-Proof Your Business, Everyday AI, and 9 Keys to Keep Your Best People from Walking Out the Door. Get in touch at kim@kimseelingsmith.com to discuss your event.
