Microsoft ANZ Chief Technology Officer Sarah Carney did not mince words when she told RNZ that older workers are “really good at AI” because they understand where value sits in their jobs. They know what to offload, what to question, and when the machine has given them rubbish. Meanwhile, their younger colleagues copy-paste the first response and call it done.
The finding cuts against the lazy assumption that digital fluency equals AI competency. It does not. And the gap between those two things is where New Zealand employers are losing money they do not even know they are leaving on the table.
Experience is the real prompt engineering
University of Sydney research through the Skills Horizon project confirms what Carney described anecdotally. Professor Kai Riemer found that experienced workers possess crucial advantages in AI deployment, particularly in judging output quality and prompting effectively. A South American creative agency CEO observed that senior colleagues use multiple AI tools, re-prompt when answers fall short, and iterate until they get something usable. Juniors, by contrast, “don’t yet know what they are looking for.”
The mechanism is straightforward. Years of managing teams, clients, and complex projects develop exactly the skills that make AI useful: context, clear expression of requirements, and the ability to spot a wrong answer. Comfort with technology is not the same as competence with it.
The demand signal is not slowing down
LinkedIn has recorded a 300% increase in job ads emphasising AI skills. SEEK data shows demand for AI-related skills in job ads has more than quadrupled over the past decade. PwC’s Global AI Jobs Barometer, covering close to a billion job postings from 2012 to 2024, found AI-related postings held firm even as the broader NZ job market weakened. Financial services overtook education as the sector with the highest share of AI skills requirements at 9.3%.
Crucially, PwC found that occupations with higher AI exposure show a 27% higher rate of net skill change. Augmentation-exposed jobs, where AI enhances workers, have grown faster across almost all sectors than automation-exposed jobs. The story is transformation, not replacement.
Yet 52% of organisations cannot hire the data and AI skills they need, according to Emergn research. Manufacturing and utilities reported the highest difficulty at 67%. CEO Alex Adamopoulos put it bluntly: “The real constraint is access to the skills that make it useful.”
Workers are running, employers are watching
A Cultivate and NewZealand.AI survey of 829 white-collar workers in May 2025 found 44% already use AI tools daily. Only 13% have received any company-led training. Over half have had no training at all.
Cisco’s AI Readiness Index paints a similar picture: just 23% of NZ organisations say their talent is ready to leverage AI, and only 35% are investing in upskilling existing staff. The preferred response is hiring contractors or new talent, not developing the people already on the payroll.
Globally, Forrester found only 16% of employees achieved a high AIQ in 2025, despite 68% of organisations running generative AI in production. Just 23% offer prompt engineering training. Forrester VP JP Gownder called the training gap “a clear bottleneck to productivity and ROI.”
Efficiency gains are hiding a revenue problem
The AI Forum NZ’s August 2025 report found 91% of NZ businesses report efficiency improvements from AI, with three-quarters reporting setup costs under $5,000. Sounds like a success story.
But Deloitte’s 2026 State of AI survey of 3,235 business leaders found only 20% demonstrate measurable revenue impact. And 84% of organisations globally have not redesigned jobs around AI. Most firms are stuck at the efficiency stage, using AI to do old tasks slightly faster rather than rethinking how work gets done.
A survey of 5,000 knowledge workers found only 2.7% qualify as genuine AI practitioners. The most common use case? Replacing Google search. Only 15% of reported use cases were judged likely to generate ROI. The research identified a “use case desert” where employees simply do not know what to use AI for in their own jobs.
There is also a dangerous perception gap. 48% of executives believe widespread AI adoption is happening across their organisations. Only 8% of individual contributors agree. C-suite leaders have clear access to AI tools in 80% of cases; frontline workers, who do the most repetitive and AI-amenable work, have access in just 32%.
Stop blaming workers for a management failure
Sixty percent of NZ hiring leaders say they would pay a premium for demonstrated AI capability, and 42% have already adjusted candidate requirements. Employers want AI-ready workers but are not building them. That is not a skills shortage. That is an investment failure dressed up as a hiring problem.
The evidence is clear: experienced workers bring context, judgement, and the ability to direct AI tools purposefully. Younger workers bring comfort with interfaces. Both need structured training, role redesign, and defined use cases. None of that happens by accident, and right now 84% of organisations have not even started the job redesign work that turns AI from a novelty into a revenue driver. The bottleneck is not age. It is not technology. It is management.
Sources
- RNZ: Demand for AI-related skills has grown and older workers are acing the pivot (2025-08-08)
- Evening Report: Are you in a mid-career to senior job? Don’t fear AI (2025-08-08)
- PwC: 2025 Global AI Jobs Barometer – New Zealand Analysis (2025)
- IT Brief: Organisations face widening skills gap for AI roles (2025)
- Cultivate: AI adoption surges in NZ workplaces but training fails to keep up (2025-05)
- Cisco AI Readiness Index 2024 – New Zealand (2024)
- NZ City: Forrester finds AI training gap stalls workplace gains (2025)
- AI Forum NZ: AI in Action – Key Findings from New Zealand’s Third AI Productivity Report (2025-08)
- Deloitte: The State of AI in the Enterprise 2026 (2025)
- HRM Online: The AI proficiency gap is now an HR problem (2025)