Return on AI investment has become one of the sharpest pressure points in the Australian boardroom. Writing for Technology Decisions, Apptio’s APAC field CTO Matt Pinter argues that as agentic AI drives up ‘token costs’, the question for leaders is shifting from whether to adopt AI to whether they can see, and control, what it costs.
The numbers behind that shift are pointed. Apptio’s 2026 Technology Investment Management Report found more than two-thirds of ANZ organisations are now funding AI by reallocating money from existing technology budgets, while 88 per cent of Australia’s senior decision-makers say they lack confidence in the value a tech investment will deliver.
The rise of ‘tokenomics’
Pinter’s core argument is that agentic AI has changed the cost equation. Where a chatbot answered one question at a time, an agent can chain dozens of model calls to finish a single task, and each call consumes tokens that carry a price. He frames the discipline of managing that consumption across teams and tools as ‘tokenomics’.
The risk is that these costs accrue invisibly. Spending spreads across departments, cloud platforms and data environments, so no single owner can see the full bill or tie it back to a business outcome. That is how AI budgets overrun without anyone noticing until the invoice lands.
When the bill outruns the benefit
The report points to a real consequence: some organisations have wound back their use of AI tools after costs exceeded budgets. That is the opposite of the productivity story most businesses set out to write, and it usually reflects a governance gap rather than a failure of the technology itself.
The 88 per cent confidence figure is the tell. When most senior leaders cannot say what a tech investment is worth, the problem is rarely the tool; it is the absence of a way to connect what AI costs to what it returns.
Why Australian businesses should care
For Australian mid-sized firms, the lesson is to treat AI cost governance as a first-class part of any rollout, not an afterthought. That starts with knowing which tools are actually in use, what data they touch, how they are billed, and who owns the outcome each one is meant to deliver.
It also means resisting the urge to bolt AI onto every process at once. A smaller number of well-measured use cases, each with a clear owner and a visible cost, will beat a sprawl of pilots that no one can account for. Building that discipline early is the core of a sound AI strategy.

From adoption to accountability
The first phase of AI in business was about adoption: getting tools into hands and seeing what stuck. Pinter’s argument is that the next phase is about accountability, connecting financial data to operational outcomes so leaders can prove value rather than assume it.
The firms that come through this phase in good shape will not be the ones that spent the most on AI. They will be the ones that could always say, to the dollar, what their AI was for and what it returned.
