Using AI to improve public services

By Shannon Haintz I APS Academy

When I think about how I use AI in my work, I think of the simpler things. Like generating an image in Google Whisk or asking CoPilot to help me with ideas and information. On a small scale, these tools are helpful for me. But there are APS employees who are using AI to do some large-scale innovative projects. Their projects improve productivity and service delivery. They solve problems at scale. They’re improving public services by thinking of problems they need solved and using AI to help solve them. They’re sharing their projects so others can learn from them.

You can find examples of these projects in the APS AI Use Case Library. It’s an online library of AI projects on the GovTeams platform that are grouped into categories.

You can also find insights and resources about how to plan, implement and manage AI responsibly on the GovAI Blog.

Improving public services is action 6 in the National AI Plan. The projects highlighted in this article are making public services more efficient, more effective and more accessible.

Claims processing and search tool at the Department of Veterans’ Affairs

The Department of Veterans’ Affairs (DVA) publishes their claims processing statistics online. In February 2026, the DVA received 8,967 compensation claims. And there are 65,661 claims with an officer for processing. The Royal Commission into Defence and Veteran Suicide recommends urgent action to speed up claims processing. That’s where AI comes in. DVA has developed a proof-of-concept AI tool to understand whether AI can help speed up processing claims. If it is implemented, it could reduce manual review time and improve accuracy for DVA claims.

The DVA also launched a test (beta) version of their AI-enhanced search tool. Veterans and their families can now ask questions and receive clear responses drawn from DVA’s websites including:

Open Arms
the Anzac Portal
Veteran Employment Program
Minister for Veterans’ Affairs.

This AI tool saves people time and improves their experience. They can get the information they need without scrolling through multiple pages and websites.

Bowerbird audio and video transcription engine

The National Film and Sound Archive of Australia (NFSA) has created Bowerbird, a machine learning-enabled mass audio and video transcription engine. Bowerbird transcribes and documents hundreds of thousands of hours of audio and video material. It allows NFSA staff, researchers and media industry professionals to find previously hidden stories from the NFSA collection. The engine is trained on Australian speech, Australian-accented English and is being trained to recognise Australian places, people and things that other engines might miss.

Indigenous ranger coastal clean-up project

Traditional Owners and rangers are combining Indigenous knowledge with drones, state-of-the-art mapping and AI to find ghost nets across the Gulf of Carpentaria. Ghost nets are abandoned or lost fishing nets that threaten marine life and pollute ecosystems. Rangers are trained in operating drones, capturing data and mapping. This is tailored to the needs and landscapes in their local area. AI processes the data they collect to calculate where the debris is. To date, 900 ghost nets have been removed and 190,000 kilograms of marine debris has been cleared. The method saves time and can cover a broad range of the area.

Share your projects

How is your agency improving public services? You can share what you’re doing with the wider APS. If you have an AI project you’re working on that you’d like to share with other APS staff, fill out the expression of interest form.