Australian organisations increased their AI spending by 115 per cent year on year - five percentage points above the global average - yet fewer than one in four can properly test and audit the outputs their AI systems generate, according to research from ServiceNow.
The company's Enterprise AI Maturity Index found that only 23 per cent of Australian organisations have AI testing or auditing processes in place. Just 17 per cent have a system to manage and track governance and compliance. The findings paint a picture of a market that has committed the budget but has not yet built the structures needed to make AI reliable at scale.
The governance gap takes on added urgency as new Australian obligations take effect on 10 December, requiring organisations to disclose when automated or substantially AI-assisted decision-making is used in decisions that affect people.
Trust as the defining tension
The index found Australia is the country most concerned about AI accuracy, with trust in AI-generated outputs and accountability emerging as the central tension for both leaders and the employees working with AI tools.
The top challenges cited by Australian executives were inadequate data accuracy, access and management at 72 per cent, lack of transparency and potential for misinformation from AI at 62 per cent and data privacy and security concerns at 54 per cent.
Group Vice President for Australia and New Zealand at ServiceNow, Pete Andrew, said: "Australia's AI challenge is a trust challenge."
"Organisations have made the investment, but they will not unlock the productivity benefits if leaders cannot trust AI outputs and employees don't have the confidence to use AI safely. The next phase of AI maturity is about building the governance, transparency and connected workflows that make AI reliable and scalable."
These challenges are central to AI for Executives & Strategy, where governance and trust remain top concerns for leaders scaling AI across their organisations.
Fragmented systems and undertrained workers
The research shows most organisations are still treating AI as a collection of individual tools rather than a connected system that can redesign work processes. Only 18 per cent of Australian organisations have integrated AI across business functions, and just 8 per cent have moved beyond individual tools to autonomous workflows. The result is that many organisations are paying for capabilities that are not being fully implemented.
For workers, the fragmented approach is fuelling distrust. Forty per cent of employees across the APAC region reported they are not receiving enough AI training, while 48 per cent cited a lack of AI strategy from leadership. The research noted that while organisations have improved on AI leadership and strategy at the executive level, that direction is not yet reaching the workforce.
Executives looking to close this gap can explore an AI Learning Path for CEOs, which addresses strategy, governance and the practical steps needed to turn investment into productivity.
Salvation Army case study
The Salvation Army Australia took a different approach. Rather than adding AI on top of existing fragmented systems, it designed governance and guardrails into Goldie - an AI-powered employee experience tool built on ServiceNow's EmployeeWorks platform - from the outset.
The tool is accessible through Microsoft Teams and allows employees to ask questions in plain language and receive answers drawn from approved internal systems based on existing permissions. Governance was built in through a three-tiered model spanning organisational AI strategy, responsible usage guardrails and tool-specific controls.
Since implementation, 90 per cent of queries have been resolved without escalation and nearly 2,700 hours have been returned to frontline staff.
CIO of The Salvation Army Australia, Milad Kruze, said: "Our people were overwhelmed by fragmented systems, duplicated processes and the cognitive burden of navigating more than 20 applications just to do their jobs."
"That load was taking them away from what matters most - empathy, compassion and connecting with the people we serve. By giving our team a single, conversational front door to get things done, we've removed that administrative friction and returned time, confidence and purpose back to the frontline."
Progress but still below par
Australia's overall AI maturity score sits at 51 out of 100, a 15-point increase on the previous year that brings it closer to the global average. Across the APAC region, 23 per cent of organisations have reached what ServiceNow calls "Pacesetter" status - the top tier of AI maturity. Those organisations are recording six times higher productivity than others in the index.
Andrew added: "Implementing AI can help Australian organisations reverse two decades of productivity stagnation. But only if they move beyond pilots and point solutions, connecting AI across workflows, embedding governance from the start and training employees to use AI with confidence."
The findings suggest an emerging window for Australian organisations to convert their investment surge into measurable productivity gains, but only if governance, data quality, testing and employee training are treated as prerequisites rather than afterthoughts.
Why this matters for executives and strategy
The spending surge shows commitment, but the low testing and governance scores reveal a critical risk: without verification, AI outputs cannot be trusted for decisions that affect customers, employees or compliance. Executives must treat governance, training and integration as core to the AI business case, not as follow-on projects. The organisations that embed these from the start - as the Salvation Army did - are the ones that will turn budget into measurable productivity, while others risk spending heavily on tools that underdeliver and create new liabilities.
Your membership also unlocks:
