When Atlassian announced in March this year that it would cut 1,600 jobs – or roughly 10% of its global workforce – the company was careful about the language it used. Why? Because the Australian-founded software giant was posting $1.4 billion in quarterly revenue and 23% year-on-year growth. 

Restructuring was framed as a forward-looking strategic pivot: the company needed to “self-fund” its investment in artificial intelligence and sharpen its focus on enterprise customers. The bottom line was simple: AI was arriving, and human headcount had to make way.

It’s a narrative that has been repeated ad nauseam across the technology sector. According to consulting firm Challenger, Gray and Christmas, more than 54,000 AI-cited layoffs were announced in the United States alone in 2025. In most of these cases, the argument for the impact on the workforce is that AI is transforming how work gets done, and organisations are simply adapting.

But a closer look at the underlying evidence raises an uncomfortable question: if AI is truly capable of replacing ten, twenty, or forty per cent of a skilled workforce, why do independent researchers consistently find that most AI projects fail to deliver meaningful returns?

A striking contradiction

A 2025 report by MIT's NANDA initiative found that 95% of enterprise generative AI pilots failed to deliver measurable impact on profit and loss, a finding widely reported and echoed by Forbes analysis of enterprise AI programmes in the same year. Only around 5% achieved rapid revenue acceleration.

Separately, a 2025 Deloitte survey of 1,854 executives across Europe and the Middle East found that most organisations reported a payback period of two to four years for a typical AI use case, significantly longer than the seven to twelve months historically expected from technology investments. Only 6% reported returns within the first year.

Businesses are simultaneously declaring that AI is productive enough to justify eliminating substantial portions of their workforce and conceding they cannot yet demonstrate reliable financial returns from it. The gap between those two positions deserves more scrutiny than it typically receives.

The reality is that many corporations are using AI as a justification for large-scale layoffs when the real motivation is often broader cost-cutting and restructuring. Some organisations may initially purge roles under the banner of AI transformation, only to quietly rebuild teams offshore once public scrutiny has eased. The AI narrative may not be the whole story.

The strategic logic behind the narrative

Companies facing investor pressure to demonstrate efficiency have found that positioning restructures around AI transformation tends to be better received by markets than admitting to straightforward cost-cutting. In Atlassian's case, the stock climbed after the announcement, demonstrating that this is a more compelling story to tell shareholders.

That does not mean AI is irrelevant to these decisions. Technology genuinely is reshaping certain categories of white-collar work, particularly in software development, administrative processing, and structured data roles. The Atlassian restructure disproportionately affected software research and development positions, the very roles closest to what AI tools can now credibly assist with. 

What gets lost when people leave

Much of what keeps a business functioning is not written in policy documents. It exists in the judgments people make, the relationships they have built, and the accumulated understanding of why things work the way they do. This institutional knowledge is the kind of contextual intelligence that AI systems struggle to replicate. The true operational cost of losing it may not become apparent until well after the restructure is complete.

Australian business culture has long placed significant value on trust and human relationships as the foundation of commercial activity. Customers make decisions based on who they know and who knows them, and organisations that strip back their workforces too aggressively may find themselves poorly equipped for a market that still runs on personal credibility and relational trust.

The impact of restructures does not stop at the door of those who are let go. Research on organisational behaviour consistently shows that layoffs have lasting effects on those who remain. ‘Survivor's guilt’ can take hold, compounded by the disorienting experience of receiving pay rises or share allocations while former colleagues are pushed out the door. Trust, once damaged, is expensive to rebuild. Employer reputation suffers, and future recruitment becomes harder.

There are also emerging concerns about the longer-term effects of deep AI integration in workplaces. Research from the University of Sydney Executive Centre suggests that prolonged use of AI tools may reduce cognitive engagement by as much as 55%, with effects that linger even after people stop using them.

Younger workers may be especially exposed, not only because they default more readily to AI tools, but because they have less accumulated experience to draw on when assessing whether AI-generated outputs are accurate or reliable.

There are also concerns about relying on AI for emotional processing and decision-making, with potential impacts on emotional development, workplace relationships, and intergenerational dynamics over the next two decades. While these concerns are only emerging, they paint a broader picture of where AI adoption is heading.

What responsible leadership looks like

AI technology is genuinely capable of augmenting human performance and reducing the burden of repetitive cognitive work. The question is not whether to adopt it, but how.

Responsible leadership means being honest about what is actually driving workforce decisions rather than reaching for a technological narrative by default.

It means consulting employees meaningfully before restructures, not after and thinking carefully about the institutional knowledge that will walk out the door and how to preserve it.

Perhaps most importantly, it means designing AI adoption around the goal of enhancing human capability, rather than simply reducing the number of humans on the payroll.

About Karlie Cremin, CEO of DLPA and Crestcom ANZ:

Karlie Cremin is the CEO of DLPA and Crestcom Australia, organisations dedicated to helping businesses solve complex people challenges with practical, real-world solutions. With a career spanning construction, government and not-for-profits, Karlie brings a wealth of experience in crafting versatile strategies and business models that deliver exceptional results.

Karlie herself has also been recognised with several high-profile honours, including the Bronze Stevie® for Thought Leader of the Year (2025), the Bronze Stevie® for Woman of the Year (2025), the Silver Stevie® APAC for Most Innovative Entrepreneur of the Year (2025), the Silver Stevie® APAC for Innovative Achievement in Thought Leadership (2025).

Her focus remains on building sustainable, profitable businesses by equipping leaders and teams with the skills and tools they need to succeed.