How do you talk about layoffs when AI is part of the story?

Atlassian’s restructuring has exposed a communications problem that is spreading across the technology sector faster than many boards seem willing to admit. Companies are under pressure to show they are adapting to artificial intelligence, yet the language of that adaptation rarely sits comfortably with a workforce that is already alert to euphemism.

The difficulty is not simply that jobs are being cut. It is that management increasingly wants to present those cuts as part of a strategic shift into an AI era, while staff hear a more direct message about shrinking opportunity, reordered priorities, and a changing view of what the company values.

In a note to staff, Mike Cannon-Brookes said Atlassian had decided to reduce the size of its team by about 10%, or roughly 1,600 employees, in order to self-fund further investment in AI and enterprise sales, strengthen its financial profile, and reorganise around its “System of Work”. He added: “Our approach is not ‘AI replaces people’. But it would be disingenuous to pretend AI doesn’t change the mix of skills we need or the number of roles required in certain areas. It does.” The company also laid out a minimum 16-week separation package, prorated bonuses, healthcare extensions, and mobility support. In the market, the message landed well enough for the shares to rise in extended trading.

That response captures why this kind of announcement is so difficult to write. Investors hear urgency, cost discipline, and a leadership team that is not waiting for the economics of software to worsen before acting. Employees hear something else. They hear that a new technology wave has changed the value of different kinds of work, and that the company is moving quickly to realign itself around that fact. Once AI is included in the explanation, every phrase becomes charged. Staff are listening not only for what is said, but for what is being softened, avoided, or folded into vague language about agility and future readiness.

This is where a great deal of corporate language fails. The instinct to tidy up bad news with abstract terminology is understandable, but it usually backfires. People do not experience a restructuring as “portfolio optimisation” or “resource reallocation”. They experience it as a sudden change in security, status, workload, and trust.

The more management describes the mechanics of a decision without describing its underlying logic, the more employees assume the harder truth is being hidden. In an AI-linked restructuring, that suspicion is sharper still because staff know the company is also trying to tell a forward-looking story to investors and customers at the same time.

Clearer communication does not remove the pain, but it does reduce the trust deficit. The strongest messages tend to do four things well. They state plainly what has changed in the economics of the business. They distinguish between efficiency gains and genuine capability shifts. They explain why redeployment or retraining could not absorb more of the impact, where that is the case. And they make the support package concrete rather than ceremonial. Cannon-Brookes’ note came closer to that standard than most, precisely because it acknowledged a hard truth that many executives still try to blur: AI does not simply sit alongside the organisation as a helpful tool. It changes the skill mix, and eventually it changes headcount.

The deeper problem is that AI has changed the emotional texture of corporate announcements. Traditional restructurings could be framed around duplication, soft demand, or macroeconomic pressure. AI introduces a different kind of unease because it suggests the terms of contribution themselves are shifting.

That places a higher burden on leadership language. Care is not enough. Precision is required. Employees can absorb difficult news when they believe it is being delivered straight. What they struggle to absorb is a mismatch between the gravity of the decision and the softness of the vocabulary used to explain it. That is now one of the defining communications tests of the AI era.