The AI gap in local government and what councils are leaving on the table

In January this year, the NSW Audit Office released its annual Local Government report. Buried in the sections on financial sustainability, infrastructure backlogs, and cyber security was a chapter on artificial intelligence that should have made a little more noise than it did.

Only 11% of NSW councils have a strategy for AI adoption. Fewer than half have implemented a formal AI policy. Most have yet to integrate AI-specific risks into their existing governance frameworks. And the Office of Local Government hasn't established a mandatory framework or minimum requirements for councils around responsible AI use.

In NSW, the Government does have a mandatory AI Assessment Framework, managed by Digital NSW, that applies to state government agencies. It's actually one of the more mature frameworks in the country. But councils aren't state government agencies. They operate as separate entities under the Local Government Act, which means the framework doesn't extend to them, meaning until something happens, councils are navigating this on their own, or deferring to other state's local government frameworks to support.

This isn't a criticism of councils. If anything, it's a reflection of just how much local government has been dealing with. Financial sustainability challenges, aging infrastructure, workforce shortages, disaster recovery, and an unrelenting housing crisis that has put councils at the centre of one of the most politically charged policy debates in a long time. When 17 out of 128 NSW councils are reporting operating losses and there's a $1 billion water infrastructure backlog across regional and rural communities, AI strategy is understandably not the first thing on the agenda.

But the thing I keep coming back to is that while councils have been focused on keeping the lights on and the backlog of works, the technology has moved rapidly. And the gap between what's possible with AI and what councils are currently doing is now large enough that it carries its own risks.

‍

What councils are leaving on the table

We recently conducted some economic modelling for the Australian local government sector, examining the productivity benefits that AI adoption could deliver across council operations. Across the national local government sector, the modelling projects over $4 billion in economic value over five years, with nearly 8,300 full-time equivalent roles worth of capacity freed up by Year 5. When we proportion those figures to NSW, using ABS public sector employment data, the picture becomes even more tangible. NSW councils could be looking at approximately $1.3 billion in economic value over five years, with around 2,500 FTE worth of capacity freed up annually by Year 5. That capacity could go towards processing the development applications sitting in the pipeline, delivering the community services that get pushed back when resources are stretched, or supporting the kind of long-term strategic planning that always seems to get crowded out by the operational.

Those are sector-wide figures, and I want to make sure they don't feel abstract for smaller councils. To bring it closer to home: a regional council with around 200 office-based staff could be looking at somewhere in the range of 15 to 20 FTE worth of capacity freed up by Year 5 under the model's assumptions. That might mean the equivalent of a small team's worth of hours redirected from drafting reports, chasing compliance documentation, and manual data entry towards strategic planning, community engagement, or proactive asset management.

To be clear, these are modelled projections rather than guarantees. They assume a realistic adoption curve that accounts for the fact that local government will lag the private sector by two to three years, which is consistent with historical technology adoption patterns. They also assume that the majority of time saved through AI is reinvested into higher-value work rather than being used to reduce headcount. This is an important distinction, and one that reflects how we believe councils should be thinking about AI.

‍

What's already happening in local government

It's worth acknowledging that AI in local government didn't start with generative AI. Several Australian councils have been using computer vision and machine learning for years in areas like road asset inspection, where cameras mounted on council vehicles, including garbage trucks and ranger vehicles, capture footage that's analysed by ML algorithms to detect pavement defects, potholes, and surface deterioration. The NSW AssetAI program, involving Canterbury Bankstown, Georges River, Blayney Shire, and Central Coast councils, is a good example of this kind of mature, operational AI that has been quietly improving how councils manage infrastructure. These aren't experiments. They're embedded in council operations.

On the generative AI side, some councils are moving with real intent. The City of Stirling in Western Australia was one of the first councils in the country to adopt a formal Generative AI Policy, and has trialled an AI-powered community engagement platform to bring more voices into infrastructure planning. Mornington Peninsula Shire is experimenting with Claude Code in a local government context, building agentic chatbots and automating testing workflows using isolated environments and public data.

Beyond Australia, the momentum is even more visible. In Canada, cities like Vancouver, Winnipeg, and Windsor have deployed AI-powered chatbots on their websites to handle resident enquiries around the clock, in multiple languages, reducing call volumes and freeing up front-line staff for more complex interactions. South Australia has launched an automated decision-making pilot for certain dwelling applications. In Queensland, Sunshine Coast Council runs an AI development assistant tool that gives applicants preliminary planning guidance 24/7. And in Honolulu, an AI-powered permit pre-screening tool cut wait times to reach a reviewer from six months to two to three days.

‍

The housing crisis as a case study

If there's one area where the AI opportunity is already becoming concrete for councils, it's development applications.

NSW councils are responsible for assessing approximately 85% of all residential development applications. The volume of work is enormous, and the pressure is only growing. The state government has committed to delivering around 377,000 homes, planning approvals need to be faster than they were in 2023, and the backlog of applications is a constant source of frustration for councils, developers, and communities alike.

The NSW Government has recognised this and adapted accordingly. Over $2.7 million was awarded to 16 councils to trial AI solutions in their planning systems through the Early Adopter Grant Program. An AI Solutions Panel was established, giving councils access to vetted AI tools that can be integrated into their pre-lodgement workflows. The technology can help with things like checking applications against planning scheme requirements, flagging non-compliant information early, and reducing the back-and-forth that currently adds weeks or months to the process.

To give you a sense of the scale involved. Of nearly 500 applications in the Regional Housing Flying Squad Program, around 30% required additional information, with applicants taking an average of 42 extra days to respond. The combined saving from not having to request that additional information on this relatively small group alone would be equivalent to around 6,300 days, or roughly 17 years. Scale that across the nearly 60,000 applications submitted in NSW each year and the implications are hard to ignore.

‍

The challenges are real, and they matter

None of this is to suggest that AI adoption for councils is simple. The Audit Office report identified several barriers that councils themselves have flagged: funding constraints, the need for governance frameworks, data quality and privacy concerns, and limited AI literacy across the workforce.

These are legitimate challenges, and they echo what we see in our work with organisations across sectors. The gap between AI's theoretical capability and how effectively people and organisations can actually use it is something we've written about extensively. We call it the people gap, and it applies as much in local government as it does anywhere else.

Councils also have unique considerations that the private sector doesn't face. They're custodians of public data, they make decisions that affect people's lives and livelihoods, and they operate under a level of public scrutiny that makes getting this right even more important. Rushing into AI without proper governance isn't just risky from a technology perspective. It risks eroding the community trust that councils depend on.

The Audit Office's recommendation that councils should review and ensure fit-for-purpose governance arrangements are in place before deploying AI is exactly right. Governance first, then strategy, then deployment. This mirrors what we consistently advocate: people first, then process, then systems.

‍

The cost of waiting

What gets less attention, though, is the cost of doing nothing.

As other parts of government and the private sector accelerate their AI adoption, local government risks falling further behind in service delivery, workforce attraction, and operational efficiency. The federal government has committed $22.7 billion over a decade to AI-related initiatives through the National AI Plan. The APS AI Plan is rolling out centrally hosted AI services across Commonwealth agencies. State governments are establishing AI policies, procurement panels, and mandating responsible use frameworks.

Local government sits at the nexus of community service delivery, yet it's the level of government least equipped and least supported to navigate this transition. The 30 June 2026 deadline the Audit Office has set for councils to review their AI readiness and for the Department to establish a mandatory framework is welcome. But the clock is ticking, and councils that wait until they're told what to do will find themselves further behind than those that start building their approach now.

And the reality is that this gap will compound over time. AI adoption follows an S-curve, where early investment in capability creates accelerating returns. Councils that build the foundations now, even modestly, will be positioned to capture significantly more value than those that start late.

‍

Where to start

The starting point doesn't have to be overwhelming.

The most practical first step is exactly what the Audit Office recommends: governance and strategy. Understand what AI is already being used for across the organisation (it may be more than leadership realises), establish clear policies for responsible use, and develop a strategy that connects AI initiatives to the council's broader objectives.

From there, the focus should be on use cases that deliver tangible value quickly and build internal confidence. Development application processing is an obvious candidate for many councils, but there are others including document drafting, data analysis, information retrieval, compliance checking, and community sentiment analysis. The UK Government's trial with 20,000 civil servants found average time savings of 26 minutes per day for document-intensive tasks, which is nearly two weeks per person annually.

Rather than just thinking about deploying AI platforms and tools, it's about building the capability of the people who will use them. Our modelling assumes that only around half of office-based staff will be using AI by Year 5, and even then, the productivity benefits depend on people genuinely understanding what the technology can and can't do. The investment in workforce capability, including AI literacy, change management, and ongoing training, is what turns a software licence into actual value.

For smaller and regional councils, the question of how to even begin looking at AI when resources are already stretched is a legitimate one, and arguably the most important one. I'm not going to pretend that "just start with governance and strategy" is a simple ask when you're carrying three vacancies and running on a skeleton crew. The answer for many smaller councils is almost certainly collaborative. The Audit Office report itself highlights councils that have successfully shared a Chief Information Security Officer as a model for resource-constrained environments. The same approach could work for AI capability: shared procurement, shared training resources, shared governance frameworks, or even shared AI strategy development through joint organisations. The local government sector has a strong history of collaborative delivery, and AI is a natural extension of that. The councils that stand the most to gain from AI are often the ones with the least capacity to pursue it alone, which makes sector-level coordination and support from bodies like the OLG and LGNSW essential.

‍

What comes next

The Audit Office has given councils a clear deadline: review your AI readiness by 30 June 2026. That's not far away, and for many councils the question is where to start.

Wherever your council sits right now, there's a practical next step.

‍

If you're still figuring out where you stand, our free AI Readiness Assessment is designed to give you a structured snapshot of your council's position across strategy, governance, capability, and implementation. It takes about ten minutes and gives you a baseline you can share with your executive team or elected council. If you want to go deeper, we can conduct a comprehensive assessment tailored to your council's specific context, aligned directly to what the Audit Office is asking for. Get in touch to access our free assessment.

‍

If you're building the business case for investment, we're offering councils the opportunity to receive a customised version of our AI productivity model. Tailored to your workforce profile, operational characteristics, and strategic priorities, it gives your leadership team a credible, evidence-based picture of what AI could mean for your organisation.

‍

And if you want to talk through where to start, how to build capability, or where to go next, book a discovery call with our team at thestrategygroup.com.au.

The gap this article describes is real, but it's also closable. It just requires councils to start, and the right support to help get the direction right.