The pace of artificial intelligence adoption has been relentless. Across industries, organisations have raced to
embed AI into customer service, operations and back-office functions, often under pressure to demonstrate innovation and cost control. Headlines have focused on efficiency gains and reduced headcount, with automation positioned as a strategic necessity rather than an optional enhancement. Yet as early implementations mature, a more complex reality is beginning to surface.
For some employers, the promise of AI-driven savings is being offset by unintended consequences. Customer experience gaps, cultural disruption and weakened talent pipelines are prompting a reassessment of earlier decisions. What began as a decisive shift towards automation is, in some cases, evolving into a recalibration of the human-AI balance.
As organisations move from experimentation to execution, a new pattern is emerging. Some of the very roles that were cut in the name of innovation are now being quietly reinstated. The conversation is shifting from how quickly AI can replace people to how effectively it can work alongside them.
“I think part of what we’re experiencing is thinking of AI as a tool. In some cases, it is a tool, but it’s fundamentally different to any kind of other technological tool that we’ve ever used before.”
On a recent AU Market Update, Ben Wheeler, Queensland Managing Director, people2people, was joined by Guest Joshua Price, Co-CEO, Symmetra, to unpack what many are calling the AI-driven redundancy cycle. Referencing reporting that suggests up to half of employers who reduced customer service headcount due to AI may rehire into similar roles by 2027, the discussion explored why organisations are reassessing their strategies and what this means for workforce planning.
Price suggested that much of the early AI adoption was driven by urgency — “a big FOMO situation where everybody is feeling under pressure to adopt, to implement and to make significant changes”. With substantial investment flowing into AI development, organisations often made decisions based on “the potential of AI, what we believe it can do versus what it can actually do”. Once implemented in real workplace settings, those assumptions have been tested.
A central distinction emerged between automating tasks and replacing roles. While AI can follow structured logic and process high volumes of information efficiently, it struggles with nuance. As Price explained, it may be “very good at solving a particular task, but it has a very limited ability to sometimes extrapolate that to something else and incorporate new pieces of information”. In customer service environments especially, it cannot recognise when “human error may be coming into play” or understand emotional context. These are, as he noted, “things that only people can do when interacting with other people”.
This realisation is prompting a strategic pivot. Rather than removing entire job families, organisations are beginning to ask, in Price’s words, “how do we assign particular tasks to AI rather than trying to replace entire roles, which it’s probably not ready to do yet”. Wheeler reinforced that many businesses are now in the phase of asking to see the measurable return on investment, and for some, those returns have not matched expectations.
The discussion also highlighted a longer-term risk that is only now gaining attention. Entry-level roles are often targeted as “low-hanging fruit” for automation. However, Price warned that organisations may be “destroying their pipeline”. Without those foundational roles, how do people learn judgement, build communication skills, and develop the expertise required to step into senior leadership positions? As Wheeler reflected after hearing commentary from the legal sector suggesting there may one day be no junior lawyers, the logical question becomes: how do we create future senior lawyers without early-career development?
Despite the cautionary tone, both agreed that AI has clear potential to enhance performance when applied thoughtfully. Price shared that he personally uses AI in certain aspects of his work, particularly where it assists in “clarifying my own thinking and decision making”. Research increasingly shows middle and senior managers acting as orchestrators of human-AI collaboration, extracting value from the technology while retaining human oversight.
The real challenge lies in integration. AI, as Price described it, operates like a “semi-autonomous person slash agent slash team, which plays by different rules”. It is not transparent in how it reaches conclusions, and it is not present in team discussions in the way a human colleague would be. This creates new coordination challenges. Leaders must ensure teams understand how decisions are made, where accountability lies, and what it means when AI contributes to errors.
Ultimately, as Price emphasised, these are “fundamentally human coordination and collaboration problems”. Successful AI implementation requires leaders who can build trust, psychological safety and inclusion, ensuring employees feel they “have a future, that gives them a sense of inclusion, a sense of voice”. Without that foundation, even the most advanced AI systems risk undermining team cohesion and performance.
As AI adoption continues to evolve, the early redundancy wave may give way to more sophisticated workforce design. The organisations that succeed will not be those that replace people the fastest, but those that combine technological capability with long-term talent strategy. The focus is shifting from automation for its own sake to building resilient, human-AI teams that deliver sustainable value.
How to balance AI adoption with sustainable workforce development?
- Identify specific tasks suitable for automation rather than eliminating entire roles
- Map long-term workforce pipelines to ensure entry-level development pathways remain intact
- Invest in leadership capability to manage human-AI collaboration effectively
- Establish clear accountability frameworks for AI-supported decision-making
- Create psychologically safe environments where employees can question and challenge AI outputs
- Measure AI return on investment against both productivity and people outcomes
