ÎÚÑ»´«Ã½

Opportunities, considerations and next steps for Australian public sector agencies to harness the agentic advantage

As government agencies across Australia continue their digital transformations, a new paradigm has emerged – one defined by agentic platforms: AI-powered systems capable of reasoning, planning and acting independently to support mission-critical outcomes.

This paradigm is already changing how we design, build, deliver and operate software and technology platforms – and it will continue to shape and underpin the delivery of citizen-based and internal public sector services into the future.

This evolution mirrors the cloud transformation of the mid-2000s. The key difference: it’s moving much faster. More than simply a trend or an option for agencies to consider, agentic is fast becoming foundational to future-ready public service delivery.

With 2025 named the year of the agent, and research indicating that 82% of organisations plan to integrate AI agents within 1-3 years (yet more than half of these feeling unprepared for the change ahead), the question is no longer if agentic capabilities will be adopted in the Australian Public Service, but how fast, how well and how wisely.

Humans + agents = hyperscaled capability and output

Explained in simple terms, agents work alongside – and are overseen by – an organisation’s skilled humans. This hybrid human + agentic squad allows an organisation to vastly hyperscale its capability and output.

Agentic platforms extend beyond traditional AI and automation. These systems don’t wait for prompts or scripts. Instead, they identify tasks, determine optimal solutions and act – often coordinating with other systems or agents – to execute business goals. This evolution from reactive to proactive technology opens doors to improved efficiency, scalability and citizen service.

What does this look like?

As shown , agents don’t simply answer questions: they ask them. A virtual team of autonomous (but supervised) agents can take care of mundane tasks, proactively gather data and share data-driven insights – and they can support employees and citizens alike. The potential is considerable, and the possible applications are many – already, we’ve identified hundreds of agentic use cases within a public sector context, with many of these being implemented within Australian public sector agencies today.

The Australian public sector state of play: from exploration to enterprise readiness

Australian government agencies are increasingly recognising the potential of agentic AI, and many have progressed rapidly from proof-of-concept explorations in 2024 to real-world, enterprise-scale implementations and platform deployment in 2025.

Conversations have moved quickly from, “What is this, and what should I be doing with it?” to, “How can this help me solve the problems I face today?”

Whether it’s automating policy research, streamlining grants management, augmenting capability or creating significant cost efficiencies in the software delivery lifecycle (SDLC), agentic systems are already starting to demonstrate tangible benefits for agencies.

With initiatives such as the now firmly in place to help ensure the safe and responsible integration of AI technologies within public services, agencies can look to maximise the benefits of agentic AI while addressing associated risks and challenges in parallel.

If the agentic programs that are in flight today continue at their current pace, we can expect to see endorsed and fully operational agentic platforms in the Australian public sector before the end of 2025.

Why agents – and why now? Real-world benefits for the Australian public sector

The Australian public sector faces numerous challenges to its ability to deliver optimum services to citizens – from the limitations imposed by legacy systems, siloed data and manual processes to the reality of under-pressure teams increasingly expected to do more with less while also continuing to keep pace with constant change.

Agentic AI enables agencies to shift from fragmented, reactive delivery to continuous, contextual and anticipatory services – streamlining operations, optimising costs, augmenting capability and addressing needs before they arise.

Agentic platforms provide an immediate opportunity for organisations to deliver cost efficiencies at scale, improve decision-making with real-time, data-driven insights, accelerate program delivery and responsiveness and personalise citizen services and improve trust.

To give just one example, a large Australian agency has already seen a 75% reduction in the time taken to process specific requests using agentic methods.

Key agentic use cases and benefits include:

• Alleviation of cost and efficiency pressures:
  For example, reducing overhead in software development and delivery cycles to free up investment in innovation.
• Service scalability:
  Addressing resourcing constraints and expanding 24/7 support.
• Control:
  Capturing and systematising institutional knowledge and IP, creating internal enablement and giving agencies greater control of their information, processes and relationships.

Unique considerations and challenges for public sector agentic adoption

Despite momentum, there are specific considerations and challenges the public sector needs to consider as it integrates agentic AI – in particular to ensure a constant focus on:

• Sovereignty:
  Ensuring that it is in-country and locally controlled.
• Security:
  Ensuring it meets the public sector’s strict security requirements.
• Consistency and compliance:
  Ensuring consistency and compliance of information across the public service.
• System modernisation:
  Unlocking budget tied up in outdated technology to enable future-ready and adaptive solutions.
• Change governance:
  Ensuring consistency and traceability as agents evolve.

A case in point: Salesforce’s Agentforce

In recent years, there has been a proliferation of agentic platforms from various vendors, with Salesforce’s emerging as one of the standout options. Agentforce is a logical choice for public sector organisations who are existing Salesforce users.

A proactive, autonomous AI application that answers questions, takes actions and improves productivity, Agentforce can be equipped with necessary business knowledge to execute tasks according to its designated role, supporting employees and customers 24/7.

In the public sector, Agentforce enables governments to deploy autonomous AI agents that automate routine tasks, offer intelligent recommendations and take action when appropriate. These agents can also seamlessly escalate complex cases to human representatives, creating human-agent collaboration which, in turn, improves efficiency and service delivery.

Public sector Salesforce customers can also use Agentforce to migrate off their legacy systems quickly and at lower cost, leveraging their existing Salesforce implementations to include agents.

How to get there:

1. Start with value, not technology:
   Anchor decisions in how the technology will help you solve complex problems, create efficiencies and deliver value to the citizen.
2. Pilot for progress:
   Identify real business pain points, then trial agentic responses. Start small and iterate, using agents to help deliver incremental change along the way.
3. Ensure the right levels of governance and human oversight:
   Establish clear guidelines and ethical frameworks, ensuring transparency, accountability, fairness and human oversight in all AI-driven decisions.
4. Prioritise sovereignty:
   Ensure data and decision-making stays within your control.
5. Support your people:
   Upskill existing teams and attract new digital talent.

Final thought

The landscape of AI agents is rapidly evolving, and agentic platforms are redefining what’s possible in the Australian public service – at speed. As these tools grow smarter and more autonomous, the greatest risk isn’t in adopting them; it’s in waiting too long to start.

Of course, alongside this potential for tremendous upside, there is a need for caution, contextual awareness and oversight. Addressing the public sector’s unique challenges and ensuring trust in these technologies will be critical for the successful integration of agentic AI in the public sector.