Australia enters the agentic AI era

Sponsored by HPOctober 22, 2026 — 10:43amAustralian businesses are moving beyond the first wave of workplace artificial intelligence, as employees adopt AI agents and companies begin to work out what happens when the technology takes on tasks rather than simply helping people complete them.HP’s 2026 Work Relationship Index found 56 per cent of Australian knowledge workers – employees whose jobs primarily involve working with information rather than manual tasks – now use work-provided AI daily or weekly, up from 38 per cent a year earlier. Some 45 per cent use AI agents, putting Australia ahead of the developed-market average.Australian businesses are moving beyond AI experimentation as workers adopt AI agents and companies rethink workflows, security and oversight.iStockThe figures challenge the idea that Australian workers have been reluctant to embrace AI. But they also point to a harder phase of adoption, where businesses must decide which tasks can be delegated to AI, who remains accountable for the result and how autonomous systems should interact with people and existing workflows.Brad Pulford, HP vice president and managing director for Australia and New Zealand, says the change is already visible in the way employees use the technology.“There’s a degree of scepticism in the public conversation about AI at the moment, but look at how people are actually using it,” says Pulford.“The job isn’t done. For many businesses, awareness isn’t the issue, it’s working out how to give their teams context and a clear road map, making sure people know why AI is being introduced, what it means for their role, and how it helps them do their best work.”KPMG Australia has also found businesses pushing further into AI. Its July 2026 AI Pulse found 89 per cent of Australian organisations expected AI to remain a top investment priority, while 83 per cent reported active chief executive ownership of AI initiatives. The report found agentic AI was shifting the focus from experimentation towards execution.Pulford says the consequences extend beyond individual productivity to the way work itself is organised.“When someone uses AI to draft a document, that’s their own workflow. When agents handle multi-step tasks, it changes how work moves between teams, who signs off on what, and what a role is for.”From pilots to workflowsHP’s research also found adoption is uneven within organisations, with Australian knowledge workers trailing IT decision-makers by 19 percentage points and business leaders by 16 percentage points in their use of agentic AI.Pulford says companies need to treat the arrival of agents as workflow redesign rather than another software rollout.“Leaders should be clear where accountability sits, be deliberate about what runs where, and involve employees in shaping the new way of working,” Pulford says.Human oversight becomes more important as AI takes on more complex work. HP’s global research found 78 per cent of proficient AI users were confident they understood where AI ended and their own judgment began, compared with 52 per cent of basic users.“Human judgment is the part that gets least attention. Leaders should be clear about what data can go where, define the decisions that always need a human, and give people the training to interrogate an AI output rather than accept it,” Pulford says.Deciding where AI runsThe spread of workplace AI is also forcing businesses to decide which workloads should run locally and which belong in the cloud.Generative AI services have largely relied on cloud-based models, but newer business devices can perform some AI processing locally, an approach often described as “AI at the edge”.Processing AI at the edge can bring computing closer to the employee and the task. Companies can choose between on-device and cloud processing according to the task, the data involved and the computing power required.Pulford says security and cost are significant factors in that decision. More than half of enterprise leaders surveyed for HP’s Work Relationship Index cited security as a major barrier to AI investment, while cost was another concern across businesses of different sizes.“Start with the data and the task, not the technology,” Pulford says.“Workloads involving sensitive information, or needing speed and offline reliability, often suit on-device processing. Large-scale models and collaborative workflows may belong in the cloud. Most organisations will need both.”Running suitable workloads locally can also avoid sending every request to a cloud service and the associated per-query costs. Pulford says businesses should assess workloads according to sensitivity, performance and cost rather than assume every AI task belongs in the same computing environment.Those decisions are also changing what businesses require from workplace computers. HP’s Work Relationship Index found 51 per cent of Australian knowledge workers nominated faster, more reliable devices as something that would help them do their best work. Yet only about one-third said they had been given an upgraded computer to run AI, compared with about half of business leaders and IT decision-makers.“The expectation of the workplace PC is fundamentally changing. It is no longer simply a device for accessing applications,” Pulford says.“Increasingly, it is becoming an intelligent platform where AI, security and collaboration come together.”HP’s Intel Powered EliteBook X G2i is one example, combining dedicated on-device AI processing with enterprise security and AI-enabled collaboration tools. Businesses can determine which workloads operate on the device and which remain in the cloud according to responsiveness, connectivity and security requirements.The human boundaryDeciding where AI runs is only part of the equation. As agents take on sequences of tasks across business systems, companies also have to decide where their authority ends.An employee using generative AI to help write an email can check the output before sending it. An agent carrying out several tasks across different systems raises more complicated questions about permissions, data access, mistakes and responsibility.Pulford says guardrails need to be established before agents begin moving information through company systems rather than added later.“Security, governance and human judgment aren’t brakes on AI, they’re what lets you scale it with confidence,” he says. “Guardrails and good judgment are what turn adoption into something you can trust.”The next phase of workplace AI will be less about whether Australian businesses use the technology than what they allow it to do.As agents move from assisting individual employees to carrying out work across teams and systems, businesses will have to make explicit choices about where machines act, where people decide and who remains responsible when the two work together.As AI reshapes how Australians work, having the right technology underpin that shift matters more than ever. The Intel® powered HP EliteBook X G2i is built for this new era of work, combining on device AI capability, enterprise grade security and seamless collaboration tools in a lightweight, adaptable design, so professionals can work confidently across hybrid environments. HP Inc. (NYSE:HPQ) is a global technology leader redefining the Future of Work across more than 180 countries. Discover the HP EliteBook X G2iHP EliteBook X Series | HP® Australia.From our partners