The 30-minute AI practice rule that could help close Australia’s skills divide

With 25 per cent of workers lacking opportunity to use AI, Australia’s workplace AI gap is often an equity issue disguised as a skills shortage. Here is how HR can help workers build confidence and spot errors on real workplace tasks without having to spend hours in training sessions.

Australia’s workplace AI gap may be less about a lack of motivation from employees and more about who gets permission to practice.

PwC’s September 2026 survey of 1,250 Australian workers found that 91 per cent of senior executives had used AI at work, compared with just 39 per cent of non-managers. Only seven per cent of non-managers said using AI well was rewarded.

Employers can’t close that gap with an optional webinar and a reminder to experiment after hours. HR teams can coach managers to give employees a small amount of paid, supervised time to practice on real tasks. 

Then they should check what people can do independently, how they recognise an AI mistake and whether the work improves. A 30-minute weekly exercise for four weeks offers organisations a manageable starting point.

1. Pick one task, not an AI tool

Start with work that employees already understand.

In a contact center, the task might be drafting a response to a complicated customer complaint. In a distribution business, it could be checking an inventory exception note. For a payroll team, it might be producing a plain-English explanation of a confusing policy, using approved internal materials.

Ask each employee and manager to identify one recurring task with a clear definition of a good result. Remove personal and confidential data from practice examples. 

Next, write down the baseline: how long the task ordinarily takes, which facts must be correct and which mistakes create downstream work. HR can supply a one-page worksheet rather than a new training platform.

The point is to make the work visible. Employees who rarely use AI should get the same opportunity to try it as managers who already have licenses, time and encouragement.

“Employers can’t close that gap with an optional webinar and a reminder to experiment after hours.”

2. Make the practice supervised and safe to get wrong

Reserve 30 minutes during the paid workday. Give the employee a realistic example and an approved AI tool. Have a supervisor or experienced colleague ask the employee to produce an answer and explain how they would check it.

The supervisor should plant one detectable flaw in the source material, or provide an AI response that omits an important condition. 

A customer service trainee might receive a plausible draft that promises a refund without checking policy. A payroll colleague might see an answer that confuses eligibility with a deadline. The employee’s job is to find the problem, consult the actual source and then correct the recommendation.

Do not grade people on how fluent their prompts sound. Grade whether they notice the unsupported claim, can explain why it matters and know when to escalate. This turns AI literacy into a work skill rather than a test of enthusiasm.

3. Measure judgement before speed

A useful four-week pilot can be simple. In week one, each participant performs the selected task without AI so the manager can record a baseline. 

In weeks two and three, the employee uses the approved assistant and explains how they checked the result. In week four, the employee faces a fresh task with a deliberately flawed AI suggestion and must decide which parts to accept, revise or reject.

Use a short rubric: factual accuracy, recognition of the planted error, correct use of the company policy and an intelligible explanation of the final decision. 

Record time as well, but only after accuracy. A faster wrong answer is usually more expensive than a slightly slower correct one.

Research gives HR a reason to distinguish assistance from independent ability. A field study of customer support workers reported in the NBER paper Generative AI at Work found substantial productivity gains from AI support, with larger benefits among less experienced employees. 

That makes access to useful assistance valuable. It doesn’t establish that those workers automatically acquire judgment they can exercise without it.

4. Give managers a role and employees a reward

HR should ask managers to schedule the practice, choose examples and coach error detection. 

A supervisor can acknowledge when a worker catches a mistake even if the exercise takes longer. This matters when workers fear negative consequences from trying a new tool. 

In PwC’s survey, 17 per cent identified concern about such consequences as a barrier, while 25 per cent cited insufficient opportunity to use AI.

After the pilot, compare results by role and team. 

• Did non-managers get access to practice? 
• Could they spot the deliberately flawed answer? 
• Did accuracy improve on a comparable new task? 
• Were the actual job steps redesigned to save time without shifting correction work elsewhere?

If the answer is no, revise the exercise before purchasing another training course.

AHRI’s 2026 research places reskilling alongside recruitment, retention, reorganisation and redundancies as interconnected workforce challenges. A fair AI program should reflect that connection. Employees need time to build an ability that organisations increasingly expect them to use.

Thirty minutes a week will not turn novices into experts, but it can give HR an inexpensive, observable way to find out whose skills are developing and where opportunities have been missed,

🧰 HR’s career resource toolkit

• Course: AI & HR: What You Need to Know to Stay Compliant Course – Part 1
• Course: AI & HR: Embedding Responsible AI – Shaping Workforce & Culture Course – Part 2
• Article: 10 strategic prompts to help your business assess the downstream risks of AI
• Toolkit: AHRI members can access AHRI’s AI toolkit, which includes a range of resources and content to help your organisation make strategic workforce decisions about how to deploy AI.
• Article: Why the majority of your workforce isn’t getting into a committed relationship with AI

Gleb Tsipursky, PhD is a behavioral scientist, CEO of Disaster Avoidance Experts, and author of The Psychology of AI Adoption at Work: From Resistance to Results (Georgetown University Press, 2026).