What the research found

CEDA's research found three areas where the opportunity for AI to improve the approvals system is greatest.

Better information access would improve decisions from the start

The earliest stages of the approvals process are where the most time is lost. Proponents struggle to access clear environmental guidance, which leads to incomplete applications.

Agencies then spend significant time requesting further information back, creating cascading delays that compound across the life of a project.

The report finds that AI tools can address this directly:

• In Queensland, a proof-of-concept used large language models to let users ask plain-language questions across thousands of complex geological reports, surfacing insights that keyword searches alone could not provide.
• In the Pilbara, initiatives to freely share ecological data are estimated to unlock $1.4 billion in economic value, driven in part by a 15% reduction in survey costs.

When the right information is easier to find, applications are better quality, assessment officers spend less time chasing details, and decisions get made faster.

AI is already delivering results for assessment teams overseas

There’s compelling evidence coming in from countries facing similar pressures to Australia, who’ve begun deploying AI across their approvals systems.

Take Denmark, for example, where an AI-assisted platform called ‘EA-Hub’ is centralising environmental assessment reports and is projected to save up to $80 million annually in assessment time.

In the United Kingdom, a pilot across three councils found that AI-assisted analysis of public submissions delivered a 66% improvement in processing time, saving assessment officers up to 10 minutes per response.

Australia is well-placed to follow. We already rank second in the world on the OECD Digital Government Index. Indeed, the NSW Government has already launched a tender for an AI pre-screening system that would review development applications before lodgement and flag likely compliance issues, reducing the back-and-forth between applicants and assessment teams.

With a strong digital foundation already in place, the real question is how quickly we move.

The foundations have to come first

The international evidence is encouraging, but the CEDA research includes an important warning: AI layered onto broken systems will not deliver results.

Three things need to be in place before technology can do its job.

First, environmental data needs to be in formats that AI can actually use. Much of what government agencies hold was never designed to be machine-readable.

Second, digital infrastructure needs to be secure and sovereign. AI models trained on foreign regulatory regimes will not reflect the complexity of Australian planning schemes, environmental standards, and jurisdictional overlaps.

Third, the workforce needs the capability to use these tools confidently. Currently, only 16% of government workers feel equipped to use AI in their work, and as mentioned earlier, only 12% actually do so. These numbers need to be much higher.

Getting the foundations right is a reason to start right now.

How TechnologyOne is already helping streamline approvals

For TechnologyOne, this is not a theoretical conversation. We've been building software for Australian governments and councils for nearly four decades, and the challenge of slow, complex approvals is one we've been working to address directly.

In 2026, we applied AI to one of the most time-consuming elements of local government service delivery: development application approvals. Working with councils, we used AI to cut approval times by around 40 per cent without compromising the integrity of the process.

The Policy and Technical Assessment Accelerator, a feature of our Property & Rating product, uses AI to interpret and apply complex planning schemes in plain language. A resident can ask their council online "Can I build a granny flat?" and the AI instantly searches the relevant planning rules, filters out 85% of irrelevant content, and presents clear guidance.

It then generates a pre-filled development application and prepares a summary ready for planner review. The result is a reduction in approval times of 10 to 21 working days on average.

This directly addresses what the CEDA research identifies as the biggest source of delay: the earliest stages of the process, where incomplete applications and information gaps create cascading delays that compound across the life of a project.

Our approach to AI: intelligence with responsibility

Trusted by more than 230 government departments and agencies across Australia and New Zealand, TechnologyOne has spent nearly four decades building software for Australian governments, under Australian law, for Australian communities.

That experience shapes everything we build.

Our AI capabilities follow the same principle. We have embedded AI across our products guided by a human-centred approach: AI provides the intelligence; people provide the ethics and the empathy.

This is verified through our certification under ISO 42001, the global standard for AI management systems. TechnologyOne's agentic AI product, Plus, and Guide, which extends that capability directly to the communities our customers serve, take this further still.

Australia has the foundations, the momentum, and now the research to act with confidence. The question is simply whether we move fast enough.