Services Australia's AI Strategy

Services Australia is embarking on a significant strategic initiative by way of its Automation and Artificial Intelligence (AI) Strategy 2025-27, setting a path to digitalise service delivery via intricate ethical, governance, and trust spaces. The strategy presents substantial learnings for C-suite leaders and senior managers in any field contemplating or growing utilisation of automation and AI. Currently, Services Australia has more than 600 automated processes that deliver to its customers and employees. The processes aim to eliminate and minimize high volumes of repetitive and rules-based work. The scale of the current automation gives the agency a strong platform for its future goals.

Purpose and Goals: Simple, Helpful, Respectful, and Transparent Services

The underlying motivation behind the strategy of Services Australia is to responsibly and safely harness the potential of AI and automation to make service delivery to staff and customers better. The end vision is simple government services so that people can get back to living their lives. Considering the volume of work of the agency, managing about 10 million customer interactions weekly and processing 468.5 million claims in 2023-24, AI and automation are considered to be central to being able to make it possible.

Through automating routine and repetitive work, the agency foresees freeing up staff time to be able to serve people with high needs or who are vulnerable. The strategy foresees AI and automation as empowering better and faster government services, more efficiency, enabling more smart decisions, and made easier in general better citizen experience. There will be anticipated gains in customer experience, staff motivation, cost saved, service integrity, and trust building.

Governance and Frameworks: Anchored in Trust and Accountability

A central pillar of Services Australia's strategy is the commitment to ensuring the use of automation and AI is human-centric, safe, responsible, transparent, fair, ethical, and legal. This approach is explicitly anchored by established principles and policies:

• Experience Design Principles: Guiding decisions to uplift the experience of customers and staff.
• Australia’s AI Ethics Principles: A national framework guiding the ethical design, development, and implementation of AI.
• Commonwealth Ombudsman’s Automated Decision-Making Better Practice Guide: Providing practical guidance to ensure automated systems comply with administrative law principles (legality, fairness, rationality, transparency), privacy, and human rights obligations.
• Policy for the responsible use of AI in government: A whole-of-government policy supporting public service AI adoption while strengthening public trust.
• National framework for the assurance of artificial intelligence in government: Setting a nationally consistent approach to AI assurance based on the AI Ethics Principles.

The strategy emphasizes robust governance, assurance, and decision-making frameworks. This includes assessing each solution individually based on varying levels of risk, predictability, impact, and scale. Safeguards are embedded, such as experimenting in controlled environments, implementing controls before wider use, evaluating against requirements, continuous monitoring with immediate pauses if standards aren't met, and having a human 'in the loop' where appropriate.

Accountability is addressed through the appointment of an AI Accountable Official responsible for implementing the DTA policy, notifying high-risk AI uses, and engaging in whole-of-government coordination. Services Australia is also considering a review of historical automation processes to ensure consistency with current governance standards. The agency acknowledges the legacy of the Robodebt Scheme and its influence on the need for clear review paths for affected individuals and transparency in automated decision-making.

Challenges and Priorities: Overcoming Barriers to Adoption

Services Australia recognizes several barriers to the successful adoption of automation and AI technologies. These include:

• A trust deficit with stakeholders (customers, staff, partners).
• A risk of technology driving transformation rather than being led by human needs.
• Outdated, siloed, or undervalued governance and planning functions not suited for dynamic emerging technologies.
• Legislation and policy that may not enable the safe and responsible use of rapidly evolving technologies.
• Limited workforce capability to safely build and manage automation and AI.
• Limited infrastructure and interoperability, stemming from legacy systems.

To address these challenges, the strategy outlines six coordinated priorities:

1. Build trust: Through transparency, data privacy, robust decisions, and human-led scrutiny.
2. Human-led initiatives: Ensuring solutions are problem-oriented and anchored on genuine customer or staff needs using human-centred design.
3. Mature governance and investment frameworks: Establishing consistent frameworks aligned with whole-of-government approaches to ensure consistency, contestability, and accountability.
4. Contemporary legislation and simplified policy: Working with partners to reform legislation to enable safe, responsible, and efficient use of emerging technology.
5. Uplift workforce capability and capacity: Investing in training, reskilling, and attracting talent to ensure staff are equipped to work with automation and AI safely and effectively.
6. Modular, connected and standardised systems: Reviewing technology infrastructure to ensure it is secure, resilient, and enables scalable, innovative initiatives.

Strategic Partners: An Ecosystem for Maturity

Collaboration with strategic partners is considered core to understanding customer needs, addressing community concerns, and maturing the agency's automation and AI capability. These partners include Advocacy Groups, unions (like the CPSU), federal and state governments, academia, and industry. They provide valuable input on customer needs, help operationalize policy and legislation, enable legislative reform, and contribute to building a robust, evidenced-based decision-making process.

Types of Automation: From Rules to Intelligence

Services Australia categorizes its automation solutions into three groups: rules-based, adaptive, and intelligent.

• Rules-Based Automation: This forms the vast majority (approximately 95%) of current automations. It relies on predefined rules to complete tasks and includes:
• Straight Through Processing (STP) and End to End Automation: Automating a process or claim entirely from start to finish based on business rules.
• Process Step Automation (PSA) and Partial Claim Automation (PCA): Automating specific tasks within a process, often working alongside manual assessments by staff before proceeding to an automated outcome.
• Digitally Enabled Processing (DEP): Technology that mimics human interaction with systems to automate repetitive, high-volume tasks by logging in, navigating applications, and inputting/gathering data.
• Intelligent Automation: These solutions use technology to complete tasks, incorporating elements like Optical Character Recognition (OCR) to extract data from images/forms and Intelligent Voice Response (IVR) services to route calls more effectively using AI.
• Adaptive Automation: The agency is experimenting with and expanding into this space, which includes technologies like chatbots, support with error codes, and leveraging Large Language Models (LLMs).

This layered approach demonstrates a clear progression from established rules-based automation to exploring and integrating more complex, data-driven capabilities.

Implications and Advice for C-suite and Senior Executives

Services Australia's comprehensive strategy provides a blueprint and valuable lessons for C-suite executives and senior managers assessing or implementing AI and automation within their own organizations. Here’s how you can benefit from this government strategy:

1. Embrace the Human-Centric Imperative: The strategy repeatedly emphasizes that automation and AI must be human-led and beneficial for staff and customers. Executives should internalize this principle. Prioritize identifying genuine human problems before applying technology. Successful transformation is "human-led transformation aided by technology". This counteracts the risk of deploying solutions that are technically sound but fail to deliver real value or worse, cause harm.
2. Proactively Build and Maintain Trust: Services Australia explicitly tackles the "trust deficit" barrier by focusing on transparency, data protection, and involving diverse stakeholders. For executives, this means trust isn't a byproduct but a strategic outcome to be actively pursued. Be transparent about where and how AI is used, protect personal information rigorously, and engage with your employees, customers, and external groups to understand their concerns and build confidence in your systems.
3. Establish Robust Governance, Not Just Guidelines: The strategy highlights the need for mature governance and assurance frameworks tailored for dynamic emerging technologies, moving beyond traditional IT governance. Learn from their structured approach involving checkpoints, risk assessment, and engagement with internal/external bodies. Identify accountable individuals for AI deployments. Consider reviewing existing processes through a contemporary AI/automation lens to ensure compliance and alignment with organizational values.
4. Invest Heavily in Workforce Capability: Recognizing limited people capability as a key barrier, Services Australia plans significant investment in training, upskilling, and reskilling staff. Executives should understand that technology adoption is limited by human readiness. Budget for comprehensive training programs on AI fundamentals for all staff, and specialized training for those involved in developing or managing AI systems. Ensure change management is a core part of your strategy, not an afterthought.
5. Assess and Modernize Your Foundational Infrastructure and Data Practices: Services Australia acknowledges that legacy infrastructure and data silos can limit the scalability and effectiveness of automation and AI. Executives must honestly evaluate their current technology stack and data management practices. Investing in modular, connected, and standardized systems and strengthening data governance are prerequisites for successful, scalable AI deployment.
6. Cultivate Strategic Partnerships: Services Australia leverages an ecosystem of partners (government, academia, industry, advocates) to inform strategy, co-design solutions, and build capability. Executives can apply this by collaborating with technology vendors, academic institutions, and relevant industry or community groups. These partnerships can provide external expertise, diverse perspectives, and accelerate maturity.

Warnings and Considerations for Executives:

The most critical warning comes from the context of the Robodebt Royal Commission, which highlighted the severe consequences of poorly governed automated decision-making. Executives must be acutely aware of:

• Automated Decision-Making Risks: Implementing AI for decisions, particularly those with significant impact on individuals (like payments or eligibility), carries high risk. Ensure clear accountability, transparency, and human oversight where appropriate. Provide clear avenues for review and contestability.
• Transparency is Non-Negotiable: Customers and staff need to understand how and why decisions are reached, especially when automation or AI is involved. Be prepared to be transparent about the use of these technologies.
• Legislation and Policy Lag: Be aware that legal and policy frameworks may not keep pace with technological advancement. Engage with policy makers where possible and ensure your legal and compliance teams are deeply involved from the outset in designing and implementing solutions.
• The 'Build vs. Buy' Decision: Carefully weigh the benefits and drawbacks of developing solutions in-house versus buying commercial products. Consider factors like relevance to local context, intellectual property, maintenance, and access to specialized expertise.
• Change Management is Complex: Even small changes can have significant impact. Implement changes within a robust control framework to manage impact effectively.

By studying Services Australia's strategic approach – acknowledging past challenges while setting a clear, principle-driven path forward – C-suite executives and senior managers can gain practical insights into deploying automation and AI responsibly, effectively, and in a way that truly serves their organization's purpose and stakeholders.