AI and Productivity: Uncomfortable Allies?
Artificial intelligence. Like a toddler with a question, it’s everywhere you turn. It’s usually sold as a quick fix to faster work, lower costs, and smarter decisions. Despite the hype, the reality is proving more nuanced.
We’re not here to argue that productivity and AI aren’t linked. That the technology can’t make tasks faster. But it would seem the relationship is more awkward than some would have you think.
That awkward relationship played out publicly for the Commonwealth Bank. In July, CBA announced 45 call-centre roles were ‘no longer required’. A new AI voice-bot would take over, so staff were told their jobs were redundant. Within weeks, the decision unravelled.
Call volumes were up. Customers, frustrated with the bot, demanded to speak with an actual person. Team leaders had to pick up the phones themselves, taking them away from their work. Far from improving productivity, the new system created extra work. CBA backtracked, apologised, and reinstated roles (read about it here).
It’s a case study in how easy it is to confuse automation with improvement. More technology doesn’t automatically mean more productivity. Sometimes it’s the opposite.
Productivity Isn’t Just Speed
The CBA saga’s a reminder of a bigger question: what do we mean when we talk about productivity?
The ABC’s Future Tense podcast explored the link between AI and productivity. One of the guests, CSIRO’s Jon Whittle, pointed out that productivity isn’t simply about speed. AI can draft a report in seconds, but is that more productive than a human spending hours to write something considered, accurate, and useful?
Productivity is about outcomes. It’s about doing better, not just faster. A chatbot that can handle more customer calls isn’t valuable if customers hang up unsatisfied. An AI that can churn out endless content isn’t useful if none of it connects with readers.
AI has the power to give us scale, but scale alone isn’t the same as progress. Sometimes it’s just more noise.
The Illusion of Efficiency
One of the traps with AI is the illusion of efficiency. It looks slick, it feels quick, but it doesn’t always hold up.
Generative tools are notorious for hallucinations. Confidently walking us in the wrong direction. A task that looks effortless on the surface can create hours of extra checking and correction. What was meant to be time saved turns into time wasted.
There’s also a risk of skill erosion. The more we rely on AI to write, analyse, or problem-solve, the less we practise those skills ourselves. In the short term, it feels efficient. Over the long term, it leaves individuals and organisations vulnerable.
Imagine a lawyer who outsources drafting to AI. It works well until they’re found out, when the large language model hallucinates case law and the judge catches you out (read about it here). Productivity is hollow if it undermines capability and trust.
Local Examples: Telstra, Qantas, and Beyond
Australia has a history of high-profile experiments with digital technology. Not all of them have gone smoothly.
Telstra has pushed hard into automation. Its chatbots were meant to make customer support faster. Instead, they’ve often been the target of frustration. Customers reported endless loops and unresolved queries, forcing them back to call centres. The technology sped up simple tasks but created friction for anything complex. The result? More complaints, not fewer. Productivity on paper, but at the expense of service quality.
Qantas also offers a cautionary tale. In the push to automate and streamline, the airline leaned heavily on self-service systems and digital check-ins. For many customers, it worked. For others, particularly during disruptions, it created confusion and longer waits. Productivity looked good in board reports, but the lived experience told another story. In fact, Qantas’ brand reputation has suffered precisely because efficiency was prioritised over reliability and trust.
The public sector has had similar challenges. MyGov was designed to simplify access to government services. In theory, it centralised information and saved time. In practice, the system often buckled under demand, leaving people stuck on hold or locked out of accounts. It’s an example of how ‘digital productivity’ needs to be balanced with resilience and accessibility. Without that, efficiency for government creates inefficiency for citizens.
These cases show the same pattern as CBA. The benefits of AI and automation only hold if the outcomes (the real test of productivity) improve alongside the metrics.
The Economic Promise and the Human Cost
That doesn’t mean AI has no role in improving productivity. Far from it. Analysts are bullish about the potential gains. Goldman Sachs estimates AI could lift global GDP by 7 per cent by 2030. Consultants promise faster turnaround times. Politicians see AI as the answer to Australia’s flat productivity growth.
The potential is real. Automating repetitive work does save hours. AI can analyse massive datasets in a fraction of the time. But the benefits aren’t evenly shared. If AI cuts costs by reducing headcount, the gains may flow to shareholders rather than workers. If it speeds up processes but erodes service quality, customers may not feel the improvement. Productivity as an economic measure is one thing. Productivity as lived experience is another.
That’s why the CBA case landed so heavily. To management, AI promised cost savings. To staff, it looked like their jobs were being undervalued. To customers, it meant longer waits. Productivity on paper doesn’t always match productivity in practice.
Productivity vs Output
Part of the confusion comes from equating productivity with output. More words written, more calls answered, more data processed.
But true productivity is about outcomes and impact. An article churned out by AI in seconds may never be read. A call answered quickly by a bot may not resolve the problem. What matters is whether the work creates value.
Work futurist Dominic Price makes the point that organisations should stop measuring output and start focusing on outcomes (read more about it here). Productivity isn’t about how much we do. It’s about what we achieve.
The question for leaders shouldn’t be “how can AI help us do more?” It should be “how can AI help us do better?”
The Uneasy Alliance
AI and productivity are uneasy allies because they force us to rethink what work means. They challenge the way we measure success. They unsettle the relationship between employers and employees.
When AI is introduced badly, as it was at CBA, trust breaks down. Staff feel threatened. Customers feel let down. Leaders look reactive rather than strategic. Productivity falls.
When it’s introduced with care, AI can help. It can strip away low-value tasks, free people to focus on higher-value work, and expand capacity without sacrificing quality. The alliance becomes less uncomfortable when the role of technology is clear: to enhance, not replace.
The Cultural Question
There’s also a cultural dimension. AI doesn’t just change how we work. It changes how we feel about work.
If emails are written by bots, do we lose authenticity? If our schedules are managed by algorithms, do we lose control? If our news is curated by AI, do we lose curiosity and end up in an echo chamber, unable to have discussions with people whose world views don’t align with our own? Some might argue that we’re already there on that one…
Culture drives productivity. Teams that feel empowered and trusted perform better. If AI is seen as a tool of cost-cutting, it will damage culture. If it’s seen as a partner, it can build confidence. Technology is never neutral. It either strengthens or weakens the way people connect with their work.
Making the Alliance Work
If you want AI to lift productivity, you need to start with clarity, not code. Organisations that succeed with their AI ambitions don’t begin with technology. They start by defining value, understanding where AI can enhance delivery, and redesigning activities around that clarity.
Here’s what to consider when looking at AI for productivity gains:
Start by defining what success looks like. What it means to be effective, not just efficient, and how you’ll measure your success. This includes setting the right productivity metrics, like outcomes, quality, and customer value, not ‘just’ speed.
Don’t think of AI as all or nothing. It can automate tasks, not necessarily entire roles. Understanding where AI works requires breaking your work into component parts. Using decision-tree logic, you can then map which tasks are transactional and therefore ripe for AI-automation, which are advisory and better handled by a human, and where your competitive advantage sits. The goal is to make deliberate, informed decisions about where AI enhances outcomes, not just removes effort.
Don’t ignore your operating model. AI will change how your work gets done, but that won’t always mean removing entire roles. In most cases, AI will require a redesign of your work, realigning human effort, and creating structures where people and AI compliment each other.
Support people through the change. Even the smartest AI strategy will fail if your people aren’t on board and supported to adopt the new technology. Treat AI like you’re implementing any other technology and help people understand the why, what, and how of the shift. That includes building trust and capability in the solution, addressing concerns early, and supporting your leaders to bring their teams on the change journey.
Looking Ahead
AI isn’t going away. Its adoption will accelerate across industries. Banking, education, healthcare, retail, government. We’re all exploring how to use it. The productivity question will remain.
Australia’s challenge is sharper than most. Our economy relies heavily on services. Many of these are people-based and trust-based. If AI is used simply to cut costs, it will backfire. If it’s used to support people and improve outcomes, it will make a difference.
The next decade will be a test of balance. Organisations that rush in may repeat CBA’s mistake. Those that move with care may unlock genuine improvements. The uncomfortable alliance between AI and productivity can work, but only if guided by fairness, transparency, and a focus on outcomes.
Final Thought
AI was meant to be the hero of productivity. At CBA, it almost became the villain. The Future Tense podcast reminds us why: productivity is about more than speed. It’s about meaning, outcomes, and human capability.
AI and productivity will never be easy partners. But with care, they can be useful ones. The real challenge is to resist the temptation of quick wins and focus instead on what productivity should always have been about: making work more valuable, not just faster.
Reach out to hear about how we can help you integrate AI into your operating model, based on the value your business delivers.
RAPHAEL MAY
Associate Director, Melbourne