The AI skills story isn’t what NZ employers think

In boardrooms across New Zealand, artificial intelligence is starting to feature in strategy decks and CEO speeches. Policies are being drafted, licences signed, training sessions run. On paper, many organisations can now say they “have AI”.

But having AI and getting value from it are not the same thing.

A recent report based on a survey of 5,000 knowledge workers in large companies across the US, UK and Canada offers a sobering reality check – one that should resonate strongly with New Zealand HR leaders facing similar pressures around productivity, talent shortages and digital transformation.

The core message: employees are indeed using AI. They just aren’t using it in ways that meaningfully change how work gets done.

For HR, this is not a technology issue. It is a skills, work-design and management problem – and it’s quickly becoming one of the most important capability gaps in the modern workplace.

The proficiency bar has moved – and most people haven’t

A year ago, “AI proficiency” meant something relatively simple: understanding what these tools are, knowing the basic risks, and being able to write a decent prompt. Organisations invested heavily in this kind of training. As a result, most knowledge workers now know how to ask an AI to summarise an email, rewrite a message or provide quick information. They understand the guardrails on privacy and bias.

In 2026, that definition is no longer enough.

The bar has shifted from “can you prompt?” to “can you build AI into meaningful, value-adding tasks every week?” Proficiency now means using AI as a standard part of real work – hiring, customer service, analysis, operations – not as an occasional helper at the edges.

By that standard, the report’s data is stark:

• 97% of the workforce are using AI poorly or not at all.
• Only 2.7% qualify as “AI practitioners” – people who have embedded AI into their workflows and are seeing significant productivity gains.
• A vanishing 0.08% are considered true “AI experts”.

The majority sit in two buckets:

• 69% are “AI experimenters” – they use AI for simple tasks like summarising notes, rewriting emails and getting quick answers.
• 28% are “AI novices” – they rarely or never use AI, or tried it briefly and stopped.

When you look at time saved, the gap between expectation and reality becomes even clearer:

• 24% of workers say they save no time with AI.
• Another 44% save less than four hours per week.
• Only a small minority hit what leaders would call transformative numbers: just 6% report saving 12+ hours a week.

For NZ organisations quietly hoping AI might claw back the equivalent of a day a week per employee – to counteract labour shortages and rising costs – those numbers should be a serious warning.

The “use case desert”: people don’t know what to use AI for

It’s tempting to blame all this on a lack of training. The report suggests something more specific: employees are not stuck because they can’t prompt; they’re stuck because they don’t know what to use AI for in the context of their own job.

The authors call this the “use case desert”.

Across thousands of workers:

• 26% say they do not have a single work-related AI use case.
• 60% say their use cases are beginner-level.
• When researchers analysed 4,500 reported work-related use cases, they judged only 15% likely to generate ROI for the business.

The sentiment is telling:

• 85% of knowledge workers have beginner or no AI use cases.
• 25% say they never use AI for work.
• 40% say they’d be fine never using AI again.

People understand that AI can summarise a document or fix the tone of an email. What they don’t see clearly is how it can change a recruitment funnel, a customer complaint workflow, a monthly reporting cycle or a backlog of contracts – the core processes where real time and money are tied up.

For HR, this should ring familiar bells. We’ve seen other technology – HRIS, collaboration tools, analytics platforms – fail to deliver on their promise when the focus stayed on features rather than redesigned ways of working.

AI is following the same pattern, only faster and with higher stakes.

What employees are actually doing with AI

The report’s breakdown of “most valuable” use cases makes the gap between promise and reality painfully clear.

Top individual use cases (% of knowledge workers who say they use AI this way):

1. Google search replacement – 14.1%
2. Draft generation – 9.6%
3. Grammar and tone editing – 5.7%
4. Basic data analysis – 3.8%
5. Code generation – 3.3%
6. Ideation and brainstorming – 3.2%
7. Meeting support (e.g. notes) – 2.7%
8. Document summarisation – 2.0%
9. Learning and skill development – 1.6%
10. Task and process automation – 1.6%

When grouped into broader categories, writing (18.1%) and research (19.6%) dominate. But both are being used at a beginner level – one-off copy tweaks and simple information gathering.

By contrast, categories where AI could reshape how work is done – data analysis (6.6%), code (6.2%), customer service (2%), operations and “task efficiency” (4.9%) – are comparatively underused.

The report’s verdict:

• 59% of reported use cases are basic task assistance.
• Over 25% had no relevant use in larger processes or workflows.
• Only 2% were judged “advanced” use cases.

In short, AI in many workplaces has become a convenience layer – a slightly smarter spellcheck and search engine – rather than a genuine productivity engine.

For NZ employers wrestling with productivity gaps and tight labour markets, that is a missed opportunity hiding in plain sight.

The executive optimism problem

Perhaps the most politically sensitive part of the report is the gap between how leaders think AI is going – and how everyone else experiences it.

C-suite respondents are overwhelmingly positive:

• 75% say they are excited about AI’s implications for their work.
• 94% say they trust AI’s contributions.
• 57% use AI for work daily; only 2% don’t use it for work at all.

They also report that their organisations have “done the right things”:

• 81% say there is a clear, actionable AI policy that effectively guides use.
• 80% say tools exist with a clear access process.
• 71% say there is a formal AI strategy.
• 66% feel encouraged to experiment and build their own AI solutions.
• 48% believe there is widespread adoption with open sharing of use cases and best practices.

Now compare that with individual contributors – the people without direct reports, doing much of the day-to-day work:

• Only 28% agree there is a clear, actionable AI policy (a 53-point gap from the C‑suite).
• 39% say tools exist with a clear access process (vs 80% of executives).
• 32% say there is a formal AI strategy (vs 71%).
• 25% feel encouraged to experiment and build solutions (vs 66%).
• Just 8% think there is widespread adoption with open sharing of use cases (vs 48%).

This is more than a perception glitch. It means many leaders are looking at adoption dashboards and policy documents and concluding “we’re in good shape” while frontline employees feel unclear, under-equipped and unconvinced.

For HR in New Zealand, which often sits between those two worlds, this gap is dangerous. It breeds complacency at the top and cynicism at the bottom – a recipe for stalled transformation.

Individual contributors: least supported, most affected

The report is equally clear about who is missing out on support.

Individual contributors (ICs) are the least likely to benefit from their company’s AI resources, despite often doing the most repetitive, automatable work.

Access to tools:

• 32% of ICs say they have clear access to AI tools.
• That jumps to 57% for managers, 72% for directors/VPs, and 80% for C-suite leaders.

Training:

• Only 27% of ICs have received company AI training.
• For managers it’s 48%; for directors 68%; for VPs 70%; and for the C‑suite, 81%.

Reimbursement for tools:

• Just 7% of ICs are reimbursed for AI tools.
• Managers: 26%.
• Directors: 44%.
• VPs: 33%.
• C‑suite: 63%.

Emotionally, the pattern shows up as well:

• ICs are more likely to feel anxious or overwhelmed by AI, less likely to trust it, and less likely to say it’s having a transformative impact on their work.
• Only 7% of ICs say their managers expect daily AI use, and only about one-third receive real encouragement to use it.
• Manager support for IC AI use has declined 11% since May 2025.

For NZ employers – particularly in sectors like healthcare, education, government services and retail that rely heavily on frontline knowledge workers – this is a backwards allocation of support. The people whose workflows could yield the biggest productivity gains are receiving the least investment.

Where industries and functions stand – and why it matters in NZ

Although the survey covers North America and the UK, the patterns are recognisable in the New Zealand context.

By industry, AI proficiency (out of 100) looks like this:

• Technology – 42
• Finance – 36
• Consulting – 35
• Manufacturing – 34
• Media – 33
• Real estate – 32
• Food & beverage – 29
• Education – 29
• Healthcare – 28
• Retail – 27

Leading sectors are more likely to have a formal AI strategy, clear policies and accessible tools. Lagging sectors – highly relevant to NZ, like healthcare, education and retail – are more likely to lack strategies and to have unclear or non-existent policies.

By function, proficiency shows a similar skew:

• Engineering/Tech – 41
• Strategy – 39
• Business development/Sales – 37
• Human resources – 37
• Marketing – 36
• Finance/Legal – 35
• Product – 34
• Operations – 32
• Customer service/Support – 27

The usage details are startling:

• 54% of engineers don’t use AI for writing or debugging code, scripts or formulas – the most obvious application for their role.
• 56% of marketers don’t use AI for creating first drafts of content.
• 87% of product managers don’t use AI for creating prototypes.

In roles where high-impact use cases are obvious, most people are still not using AI for them. Multiply that reality across NZ’s engineering teams, marketing departments, product groups and contact centres, and the unrealised potential is clear.

Training and investment: helping, but not nearly enough

The report does acknowledge progress. Since March 2025:

• Access to a formal AI policy is up 17%.
• Clear guidelines for AI usage are up 16%.
• Investment in AI tools and platforms is up 2%.

Those efforts do make a difference:

• Employees in companies with a company AI strategy are 1.6x more proficient than those without one.
• Employees with access to AI tools are 1.5x more proficient than those with no access.
• Employees who have been trained on AI are 1.5x more proficient than those who have not.
• Employees whose managers expect AI usage are 2.6x more proficient than those whose managers discourage it.

But even with these advantages, most workers remain stuck at an “experimenter” level.

On average, employees who have undergone AI training score 40 out of 100 in proficiency.

The most likely explanation is that training is still focused on the wrong things: how to access tools, how to stay safe, how to write a basic prompt. Those are necessary foundations, but they don’t close the gap between “I know what AI is” and “I can use AI to redesign how my team gets work done.”

What NZ HR leaders need to do next

For HR leaders in New Zealand, the implications are clear – and they go well beyond running another AI 101 course.

1. Change what you measure

If your AI success story is built on access and adoption – number of licences, logins, course completions – it’s almost certainly overstated. Shift the focus to:

• Time saved per role.
• Quality and maturity of use cases by function.
• The percentage of core processes in each team that have been redesigned with AI.
• Clear business outcomes (cycle time, quality, error rates, customer satisfaction).

2. Make use case development a core management responsibility

The report is clear: leaving “use case discovery” to individual curiosity doesn’t work. HR can:

• Require every people manager to identify and document at least a handful of AI use cases for each role.
• Build function-specific playbooks and libraries (e.g. for recruiters, contact centre staff, analysts, payroll teams).
• Recognise and reward managers who systematically redesign work using AI, not just those who attend training.

3. Bridge the individual contributor gap

ICs in NZ organisations are often the ones dealing with the highest volume of repeatable work – exactly where AI can help most.

• Standardise access to approved AI tools across levels, not just at the top.
• Create fair reimbursement policies that don’t leave ICs paying personally for productivity tools.
• Set clear expectations that managers will support, not silently discourage, AI experimentation in their teams.

4. Redesign training around workflows, not tools

A workforce scoring 40/100 after training doesn’t need more slide decks about prompt structures.

• Teach staff to map their own workflows, spot bottlenecks and test AI in those specific places.
• Emphasise evaluation skills: how to check AI outputs, handle errors and refine processes over time.
• Move from one-off events to continuous learning – internal clinics, peer groups, role-based certifications from basic to advanced.

5. Close the executive awareness gap

In many NZ organisations, leaders are more bullish on AI than their people – and less aware of the day-to-day barriers.

• Build regular skip-level forums focused specifically on AI challenges and use cases.
• Ask executives to shadow employees as they try to use AI in real work, not demos.
• Ensure reporting surfaces uncomfortable truths – where AI is underused, misused, or not used at all – alongside the success stories.

6. Accept the bar will keep rising

​​​​​​​As AI capabilities advance, the distance between “experimenters” and “practitioners” will only grow.

• Treat AI proficiency as a moving target, not a box to tick.
• Embed continuous learning into your L&D strategy and performance conversations.
• Make AI capability part of role profiles and progression pathways across functions, not just in tech.

The real question for NZ employers

The report closes with a simple but uncomfortable observation: AI in the workplace is not primarily a technology rollout; it is a redesign of how work is done.

New Zealand HR leaders, as stewards of skills, culture and work design, are at the centre of that shift whether they like it or not.

The global adoption numbers can be misleadingly impressive. ChatGPT reports nearly 900 million monthly users; 56% of Americans say they use AI. Yet, underneath, 85% of the workforce still lacks a value-driving AI use case, and 25% don’t use AI for work at all.

The real question for NZ organisations in 2026 is not “Do our people have access to AI?” or even “Have they done the training?”

It is this: Have we helped our people turn AI from a clever assistant at the margins of their day into a core part of how value is created – and are we measuring the answer honestly?

That is no longer a question for IT alone. It is a question for HR. And the window to get ahead of it is closing fast.