AI is changing work – now, work has to change

Artificial intelligence has reached a turning point.

For the past two years, much of the conversation has centred on experimentation. Companies rushed to purchase licences for generative AI tools, employees experimented with new ways of working, and executives wondered whether they were moving quickly enough to keep pace with competitors.

Now, the focus is shifting.

At a recent industry lunch hosted by UNSW Employability, Business Sydney and UNSW Business School, business leaders from AMP, Westpac and BDO Australia argued that AI has entered a more mature phase – one where competitive advantage will come from redesigning organisations around new ways of working, rather than simply adopting the latest AI tool.

Photo gallery: The transformative potential of AI industry event

As Professor Paul Andon, Dean of UNSW Business School and facilitator of the panel, observed, most businesses now occupy an uncomfortable middle ground: experimenting with AI, achieving isolated productivity gains, but still searching for a clear pathway to enterprise-wide transformation.

“We’ve implemented in individual ways, but there is more to come,” Prof. Andon said, arguing that the ad hoc nature of early experimentation has limited AI’s business potential. “What we’re hearing more recently from the big tech companies and senior consulting firms is that this is creating a gap between the potential of what organisations can achieve with AI and what they’re actually achieving right now.”

That gap increasingly defines the next phase of AI adoption: the conversation is moving away from technology itself and towards the organisational capabilities needed to realise AI’s potential – from leadership and governance to workforce capability and education.

From individual productivity to organisational redesign

The first wave of AI adoption has largely been about individual use. Employees are using AI assistants to summarise meetings, draft reports, and accelerate research, freeing up hours each week for higher-value work.

The next phase is less about personal productivity than organisational transformation – rethinking how work flows across teams, functions and entire businesses. Prof. Andon frames this transformation through three interconnected ideas: workflows, workforce and work-life.

Learn more: How AI is changing work and boosting economic productivity

Rather than asking where AI can be inserted into existing processes, companies are increasingly asking a more fundamental question: if we were building this business today, what would our workflows look like?

They must also equip people with the skills to thrive in new ways of working, while recognising that AI’s implications extend beyond productivity to broader questions about society and the future of work. “AI is not just about how we can improve national productivity and improve the value generation of organisations; there are significant societal implications,” Prof. Andon said.

“That’s where universities have a huge role to play – not just in getting our graduates ready and capable, but, with the up to 1 million lifelong learners that we want to serve at UNSW, how do we help industry upskill their workforces to be ready for what’s to come and what’s already happening now? If you change your workflows, how is your workforce going to be ready for that change?”

Prof. Andon noted that, alongside new undergraduate offerings – including an AI in Business and Society major – UNSW Business School has expanded its executive education and lifelong learning programs to help experienced professionals develop strategic AI capability rather than simply technical proficiency.

The goal, he said, is “not just thinking about how to tool up with AI, but how to develop the capability to ask the right questions, to be curious, to be able to think”.

Why AI strategy has become a leadership issue

Many leaders have already learned that enterprise AI adoption cannot be reduced to software procurement. The real work begins once organisations decide how AI fits into their broader business strategy – a shift that is changing what successful deployment looks like, according to Fahim Khondaker, Partner, Data & AI at BDO Australia, who also serves as a Professor of Practice for the School of Accounting, Auditing and Taxation at UNSW Business School.

“It’s a combination of planning, implementation and governance, and getting the balance right between those three,” said Prof. Khondaker. “Planning in the sense of having a purpose as to why you’re doing it, with end-to-end workflow becoming quite big. People are no longer looking to deploy AI just for the sake of AI, so it’s going back to the drawing board a little bit about planning.”

Many businesses assumed enterprise AI adoption would naturally follow once employees were given access to tools such as Microsoft Copilot. But widespread access rarely translates into meaningful organisational value.

"For those who haven’t started or are a little bit trepidatious about what it means for your businesses: start now"

RENAY RINGMA

Attention is now returning to more fundamental strategic questions. What business problem needs to be solved? Which workflows should be redesigned? How does AI support broader organisational objectives?

Strategic intent alone is not enough; implementation becomes the next challenge. Prof. Khondaker observed that organisations seeing the greatest value are moving beyond straightforward efficiency gains and focusing instead on how AI can optimise end-to-end processes and create new revenue streams, rather than simply reducing costs.

Governance, meanwhile, has become increasingly important as AI adoption matures. Leadership alignment is now essential, with coordinated decision-making required across finance, operations, technology, risk and people functions.

At the same time, IT teams are beginning to rein in what has become known as “shadow AI” – employees independently adopting external AI tools outside formal organisational oversight – as enterprise investment grows and stronger controls become necessary.

“We’re seeing a significant shift,” Prof. Khondaker said. “Suddenly, it feels like AI’s grown up a bit from a toy to experiment with to now being an enterprise solution that’s here to stay, and now we need to change our thinking around how we adopt it.”

Building capability before chasing technology

If technology alone will not transform organisations, what will? For Renay Ringma, Head of Artificial Intelligence at AMP, the answer begins with people. Her advice to leaders still hesitant about AI adoption was simple: don’t wait until you feel ready.

“For those who haven’t started or are a little bit trepidatious about what it means for your businesses: start now,” she said. “You will never reach the optimal state of readiness, particularly in this space; readiness is always a state that is coming.”

At AMP, that has meant establishing strong foundations before scaling solutions across the business. One of the earliest steps was developing a responsible AI framework to guide investment decisions, governance and employee education – an initiative she described as fundamental to building trust both inside and outside the organisation.

Equally important for readiness is developing organisational capability and “thinking about your people”, Ms Ringma said.

“Thinking about the culture as well as the capability is key,” she said. “It’s not just around AI literacy; it’s also about attitudinal behavioural approaches: openness to experiment, and for the organisations to allow that.”

Ms Ringma also argued that AI strategies need a clear sense of purpose and value. “Long gone are the days of ‘let a thousand flowers bloom, and something good will show up,’” she said. Instead, it’s important to be “laser-focused on what you are trying to achieve for your business” and use AI to support those strategic objectives, rather than adopting new tools simply because competitors are doing the same.

Prof. Andon reinforced this point, observing that companies can become preoccupied with technological advancements before fully appreciating their broader implications. “There’s a significant human element in the transformation,” he said. “We often get consumed by the tool and by the models, but a lot of the transformation is educational, it’s cultural, it’s attitudinal.”

Innovation needs guardrails – not handbrakes

Encouraging innovation while maintaining trust presents an additional challenge, particularly in highly regulated industries.

Minnie Singh-Murphy, Executive Manager – AI Enablement and Learning at Westpac, described AI readiness through the pillars of technology, people and process.

"We’re very, very conscious that we’re a financial services organisation, so we’ve got to make sure that we’ve got guardrails in place"

MINNIE SINGH-MURPHY

Understanding the technology itself is only the starting point, she argued. Employees also need role-specific education to use AI responsibly, supported by organisational guardrails that encourage experimentation without compromising customer trust.

“It’s making sure our people know responsible use,” Ms Singh-Murphy said. “If we’re giving them tools, we’re very, very conscious that we’re a financial services organisation, so we’ve got to make sure that we’ve got guardrails in place.”

One of the most important capabilities leaders now need to cultivate, she suggested, is judgment. At Westpac, education increasingly focuses on helping employees navigate what Ms Singh-Murphy described as the “jagged frontier” between accepting AI-generated outputs and critically evaluating them before acting.

That idea – that AI should augment rather than replace human decision-making – surfaced repeatedly throughout the discussion. While generative AI can dramatically accelerate analysis and routine work, competitive advantage ultimately comes from combining technological capability with domain expertise, critical thinking and sound judgment.

Scaling AI means solving real business problems

Businesses are already moving AI applications beyond isolated pilots into enterprise-scale deployments. Ringma shared several examples from AMP’s transformation journey in which early proof-of-concept projects have evolved into business-wide capabilities.

Within the company’s contact centres, AI-powered systems now achieve high levels of accuracy in supporting customer interactions while improving both customer and employee experience. Automated transcription and summarisation capabilities are also reducing administrative workloads for financial advisers, allowing them to spend more time with clients and less time on meeting documentation.

Ms Ringma described the AI solutions as a “fundamental game-changer” for AMP, both in terms of outcomes and learning. “It’s all the things we’ve learned through that process, and then also thinking about how the component technology parts can apply to other processes or other parts of our organisation,” she said. “We always think about it in a modular way, so it’s not a one-off build; if we’re going to design and build something, it’s thinking about how that is the stepping-stone to the next thing.”

Subscribe for the latest research, analysis and insights from UNSW Business School

Just as importantly, Ms Ringma argued, business leaders need to keep testing and refining new ideas, even as they focus on scaling proven solutions. “The things we’re delivering now were proof of concepts six months ago,” she said. “Needing to focus on scale – and, increasingly, looking to get value – doesn’t mean that you don’t test, do innovation exercises or do proof of concepts. You’ll throw some things away, but others will be the future.”

Looking beyond the technology

Prof. Khondaker offered a note of caution, however, noting that while AI capabilities are advancing rapidly, the technology itself is often oversold. Many demonstrations marketed as revolutionary AI are little more than sophisticated automation, while others remain closer to aspiration than reality.

“You’ve got to take it with a grain of salt and be mindful of that; there’s no replacement for the hard work of actually thinking through your process and then teaching the AI that knowledge step,” he said.

“AI can’t read your mind yet, nor can it read the operating processes of our companies. There’s so much that’s in people’s heads at the moment that (getting it out of their head) that’s the hard work”, and it’s actually got very little to do with AI,” Prof. Khondaker said. “We do so much on autopilot.”

The organisational challenge may prove greater than the technological one. Employees who are asked to codify years of accumulated expertise naturally wonder whether they are helping build systems that could eventually replace aspects of their own roles.

“What’s going through their minds is, ‘If I help you do that, I become vulnerable; what happens to me once I’ve given you all that knowledge?’” Prof. Khondaker said. “It’s a leadership challenge – how do you give your staff the comfort to feel safe enough to give you that information, so you can build successful AI products?”

Preparing graduates – and organisations – for an AI-enabled future

As AI continues to reshape business, universities face their own transformation challenge. With knowledge becoming increasingly accessible through AI systems, the value graduates bring will depend more heavily on their ability to apply judgment, solve complex problems and integrate technical capability with human insight.

“The value of knowledge is decreasing at a rapid pace because it is so accessible with AI,” Prof. Andon said. “You can access so much knowledge; it’s what you do with it that counts now.”

Learn more: Why AI systems fail without human-machine collaboration

That philosophy is reshaping how UNSW Business School approaches employability, placing greater emphasis on experiential learning, industry partnerships, and opportunities for students to apply their knowledge in real-world business settings.

Employers, the panellists agreed, are increasingly looking for graduates who combine AI literacy with curiosity, adaptability and deep disciplinary expertise. Citing financial advice as a domain where the importance of expertise is clear – “AI is not going to replace someone who deeply understands a retirement product” – Ms Ringma challenged the growing narrative that AI diminishes the value of specialist knowledge.

“In fact, the opposite is true: we need experts now more than ever.”

Adopting AI: 10 recommendations for business professionals

1. Define the business outcome: Leaders should specify the customer, operational, revenue or workforce outcome before approving an AI project.
2. Select an end-to-end workflow: Teams should examine the full process rather than automate one task without considering upstream and downstream effects.
3. Assign accountability: The organisation should identify who owns the outcome, data, model, controls, employee impact and review process.
4. Establish responsible-use controls: Leaders should set approved tools, data rules, human review requirements and escalation processes before deployment.
5. Develop role-based capability: Training should reflect the decisions, data and risks encountered in each employee’s role.
6. Protect employee trust: Leaders should explain how employee knowledge will be used, how roles may change and how staff will participate in implementation.
7. Measure business results: Measures should cover process performance, customer outcomes, employee experience, risk and revenue where relevant.
8. Design for reuse: AI components should support other processes in which the underlying technology, data, or control structure can be applied.
9. Maintain testing alongside scale: Organisations should continue controlled experimentation while expanding solutions that demonstrate value.
10. Retain human judgement: Employees with domain responsibility should evaluate outputs and remain accountable for decisions with material consequences.