53% of organisations across Australia and New Zealand now have AI implemented in workflows or embedded in strategy. Yet only 8% are measuring its impact through clear metrics or KPIs.
Source: Talent, 2026 AI in the Workplace Report – 1,505 business leaders and technology professionals across Australia and New Zealand.
AI has moved very quickly from experimentation to everyday business use.
For founders, the attraction is obvious.
If AI can draft faster, analyse information, automate administration, improve customer response times and take repetitive work away from employees, it should create greater capacity without headcount increasing at the same rate.
And for a growing business, that is incredibly valuable.
But there is a question founders should be asking before assuming AI automatically equals productivity:
What work has AI actually removed from the business?
Because adding AI to an existing job is not the same as redesigning the job.
Increasingly, the data suggests there is a significant gap between AI adoption and organisations understanding whether it is actually changing performance.
AI Adoption Is Accelerating. Measurement Isn’t.
Talent’s 2026 research across Australia and New Zealand found that 53% of organisations now have AI implemented within workflows or embedded in strategy, compared with just 13% a year earlier.
Almost one-quarter — 24% — have now fully embedded AI into their strategy and operating model.
Yet only 8% are measuring AI’s impact through clear metrics or KPIs.
That is a significant disconnect.
We know AI is being used.
What many businesses don’t yet know is whether it is actually creating more organisational capacity.
The Australian Government’s National AI Centre reports a similar acceleration among SMEs.
Across December 2025 to February 2026, 43% of Australian SMEs reported some level of AI adoption, with adoption reaching 44% in February alone.
Among businesses already using AI, broader integration across the organisation is also increasing.
So the conversation is changing.
It is no longer simply:
Should we use AI?
It is becoming:
How should work change because we are using AI?
We Are Very Good at Adding. We Are Not Always as Good at Taking Away.
This isn’t really an AI problem.
It is an organisational design problem that existed long before AI arrived.
Growing businesses accumulate work.
Another process.
Another report.
Another approval.
Another system.
Another meeting.
Another spreadsheet.
Another task that somebody started doing three years ago and nobody has ever stopped to ask whether it is still necessary.
Then AI arrives.
We add that too.
The employee still attends the same meetings.
They still prepare the same reports.
They still navigate the same approvals.
They still work through the same processes.
They are simply now expected to use AI somewhere along the way.
That may make an individual task faster.
It doesn’t necessarily make the organisation more productive.
There Is Now a Term for Some of This Hidden AI Work: “Botsitting”
Recent research reported by the Australian HR Institute provides an interesting counterpoint to the productivity discussion.
A study of 6,000 knowledge workers across Australia, the United States and the United Kingdom found employees were spending an average of 6.4 hours each week supervising and reviewing AI-generated work.
That includes providing context, reviewing outputs, correcting mistakes and managing the technology.
The researchers call it “botsitting”.
It accounted for 37% of employees’ total AI-related time in the study.
That doesn’t mean AI isn’t saving time.
It does mean the productivity equation is more complicated than:
AI completed the task faster = productivity gain.
Human review is going to remain appropriate — and necessary — for plenty of work.
The question is whether the total workflow is now better.
Founders Should Look at the Whole Job, Not the AI Task
This is where the HR conversation around AI needs to mature.
Instead of looking only at which individual tasks AI can perform, founders should look at the whole role and the workflow around it.
For each role, ask:
- What can AI genuinely automate?
- What can AI accelerate?
- What requires human judgement?
- Where is human review still necessary?
- What work could disappear completely?
- What processes no longer make sense?
- What capability will employees need as their work changes?
- Where should the capacity created by AI be reinvested?
That last question matters enormously.
Imagine AI genuinely gives someone five hours back each week.
What happens to those five hours?
Do they spend more time with customers?
Do managers finally have time to manage?
Does a salesperson spend more time selling?
Can an employee take on higher-value work?
Can the business absorb growth without immediately recruiting another person?
Can you improve service without increasing headcount?
Or does that five hours simply get filled with another collection of low-value tasks?
Capacity only creates value when you decide what to do with it.
Be Careful With AI Headcount Maths
There is understandably a lot of discussion about whether AI will reduce headcount.
For some businesses and some roles, it will.
But founders should be cautious about applying a percentage to a role and assuming that percentage translates directly into a reduction in people.
Particularly in SMEs and scale-ups.
Jobs in smaller businesses are rarely as tidy as the position description suggests.
People hold relationships.
They know the history behind decisions.
They solve problems nobody formally assigned to them.
They fill gaps between functions.
They manage exceptions.
They know which customer requires a phone call rather than an automated email.
And they make dozens of judgement calls that often aren’t documented anywhere.
If AI automates 30% of the visible tasks in a role, it doesn’t necessarily mean you have eliminated 30% of the job.
Before changing headcount, you need to understand the actual work.
Otherwise, rather than removing work from the organisation, you simply redistribute it to everyone who remains.
AI Capability Is Becoming Workforce Capability
There is another element founders shouldn’t overlook.
AI literacy is rapidly becoming part of workforce capability.
Gartner’s 2026 Australian research found 38% of Australian employees are now expected to use AI in their role.
Of those employees, 85% have access to enterprise AI tools — but 86% are also using personal AI tools to improve efficiency or complete more complex work.
That creates a different set of people risks.
What information is being entered into those tools?
Do employees understand privacy and confidentiality requirements?
Can they assess whether an AI-generated answer is accurate?
Do managers know how AI is being used within their teams?
Who remains accountable for the output?
Talent’s research found only 16% of organisations were confident their employees knew what data was safe to enter into AI tools.
So AI capability cannot simply mean teaching people how to write better prompts.
It needs to include judgement, governance, privacy, critical thinking and knowing when not to use AI.
What Should Founders Measure?
If you are investing in AI, move away from measuring adoption alone.
Knowing how many employees have an AI licence or how frequently they use a tool tells you very little about whether the business is performing better.
Instead, look at measures such as:
Time
Has the end-to-end process actually become faster?
Capacity
How many hours of employee time have genuinely been released?
Quality
Has the quality or consistency of the output improved?
Customer impact
Has service become faster, easier or better?
Cost
Has AI reduced existing costs or avoided future costs?
Growth
Can the business handle greater volume without headcount growing proportionately?
Risk
Has AI introduced new privacy, accuracy, governance or accountability risks?
And perhaps most importantly:
What work have we stopped doing?
Because that is where you begin to see whether AI has genuinely changed the operating model.
AI Is a Technology Decision. But It Is Also a People Decision.
AI undoubtedly creates a significant opportunity for founders.
Australian SMEs are already using it heavily for content generation and data analytics, with the National AI Centre reporting each as an AI use case for 54% of adopting businesses.
The technology will continue to improve.
The bigger challenge may be whether organisations redesign quickly enough around it.
Because productivity doesn’t come simply from putting an AI tool into somebody’s hands.
It comes from looking at the work around that technology and deciding:
What should we automate?
What should we keep human?
What should we stop doing?
What skills do our people now need?
And where could the capacity we create generate more value?
For founders, that is the real opportunity.
Not simply using AI to do the same work faster.
But designing a better business around what AI now makes possible.
Pupal HR Consulting works with founders, SMEs and growing businesses to design the people structures, roles, capability and workforce strategies needed to scale. If AI is changing how work gets done in your business, your workforce design needs to evolve with it.
Sources
Talent. 2026 AI in the Workplace Report. Australia and New Zealand.
Australian Government National AI Centre. AI Adoption Insights: December 2025 – February 2026.
Australian HR Institute. Are We Developing a Generation of “Botsitters” at Work? 2026.
Gartner. HR Survey Finds AI Is Boosting Productivity but Widening Workforce Divisions in Australia. 2026
