The problem with digital in Australian government isn't ambition, it's the conditions to deliver
I have worked inside both the UK and Australian government digital systems. Making meaningful change to complex policy, service delivery and digital infrastructure is hard wherever you are.
The UK experience isn't a blueprint, but it is a useful reference point: one that shows what the conditions for change look like, what it takes to build them, and how long that actually takes.
What makes transformation possible isn't a single programme or a strategy document. It requires coherent intent, genuine investment in capability, and above all, time — time to build, learn, and set realistic expectations about what change involves and what it costs. Digital delivery only becomes sustainable when it's treated as an ongoing part of how an organisation operates, not as a project with an end date, measured by outputs rather than outcomes.
The Australian government has been investing in frameworks with the intent to make digital change. That investment is real, and in pockets and at moments, the results have shown what is possible. But frameworks and the conditions to deliver on them are different things. Without those conditions, digital change doesn't fail dramatically. It quietly languishes.
The importance of continuity in the digital centre
The most significant difference in how digital capability has been built and maintained comes down to one thing: continuity at the centre of government. In the UK, that centre has been the Government Digital Service (GDS). Founded in 2012, GDS began with a clear mandate — redesign how the government works for citizens, starting with GOV.UK and 25 exemplar services. It has survived restructuring, political pivots and budget pressure across consecutive governments. It has not solved every problem, but through it all, digital has remained centralised within an expert division with a mandate to set standards, develop capability and drive change. That continuity matters. The current UK picture is not without its own pressures and unevenness, but the structural continuity has held in a way that Australia's hasn't.
The culture that grew around GDS matters as much as the institution. Standardisation, modularity, human-centred design, and evidence-led delivery have spread across departments, agencies and local government. It didn't happen by accident. It happened because there was a centre with the authority and the staying power to make it happen. While the work has evolved, the purpose remains: “every corner of the state is using technology to make government work for the citizens it serves” (GDS Blog, 2026).
In Australia, the Digital Transformation Office, and later the Agency, was conceived with a similar mission: “strategic and policy leadership, expert investment advice and oversight to drive digital transformation that delivers benefits to all Australians” (DTA, Corporate Plan, 2025). The DTA has never achieved the same continuity or cross-government momentum that allowed GDS to progressively raise the standard across the whole system.
The ambition has existed, but the conditions to sustain it at scale have not.
Fragmentation as the norm, not the exception
Australia's federal structure means national digital responsibility is divided across four distinct levels:
Federal policy and services, which must integrate with state implementations
Eight states and territories, each with their own political agenda, funding levels and digital maturity
Departments, agencies and bodies responsible for digital within those structures, subject to different policies at both federal and state levels
The services themselves, delivered within programmes and projects shaped by whatever capability and strategic direction exists within their organisation.
The DTA plays a role broadly similar to GDS at a federal level, but does not hold responsibility for all digital policy and does not sit above a system with shared standards. The further down the system you go, the more that variation compounds.
The result is a digital landscape where coherence is the exception, not the norm. Fragmentation is not simply an accident of federal structure. Departments actively protect their autonomy, and the system is, in many respects, designed to produce exactly this outcome.
Fragmentation increases the inconsistency of digital quality
Without a shared baseline for standards and capability, quality becomes uneven in ways that are hard to see until something goes wrong.
One team can be delivering work of genuine excellence while another, in the same organisation, makes risky decisions with little understanding of what's at stake. The conditions that made the first team successful are not visible, replicable or sustainable across the whole. Where those conditions do exist, the work shows what is possible. Service NSW at the state level and the Australian Tax Office (ATO) at the federal level have maintained practical digital delivery. But pockets of excellence don't automatically raise the floor elsewhere.
The Australian National Audit Office (ANAO) found that 45 per cent of its 2024-25 findings related to IT controls, with persistent weaknesses in governance across the sector. What those audits measure is adherence to frameworks and controls. The question of whether services are actually working for the people who use them sits outside that frame. That gap in how digital quality is defined and assessed is part of the same problem: process and compliance are measured, outcomes less so.
When risks materialise in government, the consequences fall on citizens. Robodebt is the most visible example. The Royal Commission found there was ‘remarkably little interest’ in the scheme's legality, that implementation was rushed, and that almost no thought was given to how it would affect the people it was meant to help. Recommendation 23.3 states that all new programmes be developed in a ‘customer-centric way with recipients at the forefront’. Describing what should have been standard practice, not a reform to be mandated after the fact. A failure of capability, conditions and governance that should have caught harmful design before it reached anyone.
The Commission also documented what exclusion of digital expertise looks like in practice. When the DTA was asked to review the scheme for compliance, it was given four days to respond. It stated it could not work within that timeframe and that its ‘involvement should not be read as an endorsement’. It was not consulted again.
The Bureau of Meteorology (BOM) redesign is a different kind of failure, but the same underlying pattern. The $96.5m project became a political and reputational problem when changes to the public-facing service made it more difficult to access important weather data. Leisa Reichelt's analysis, ‘What went wrong with the BoM Redesign?’ investigates the practice failures that led to the design.
Her analysis points to something structural at the heart of the failure. When a commissioning agency lacks the internal expertise to define what ‘good’ looks like, it delegates that judgment to its delivery partner. But the partner is measured within constraints set by that same commissioner. Neither party ends up truly accountable for the quality of what gets built. Governance determines what is ‘gated’, prioritised and ‘protected’. Everything else falls by the wayside.
Getting digital services wrong costs money. It costs the public the service they need, and it costs the government the credibility it needs to keep investing.
Fragmentation runs through the frameworks too
The fragmentation isn't only structural. It runs through the policy and strategy frameworks intended to guide digital delivery.
At the state level, each jurisdiction develops its own digital strategy largely independently of the federal government and of each other. I observed this while researching Australian government digital strategies as part of developing a Digital Enabling Strategy for the Federal Department of Agriculture, Fisheries and Forestry (DAFF). Similar work was being done in isolation across multiple organisations, each producing its own strategy in its own language, with no common frameworks and no mechanism to learn from each other's experience.
At the federal level, the DTA sets digital experience standards. The Australian Public Service Commission sets workforce capability frameworks. Data, digital and cyber are split across departments: the DTA oversees digital, Finance data policy and Home Affairs cyber. In the UK these responsibilities sit closer together. In Australia, each organisation is governed separately, producing its own guidance and frame of reference. When working with the DTA on digital maturity at DAFF, the separation between data and digital created friction, particularly in how frameworks like these are applied to drive organisational change rather than just measure it.
The practical consequence shows up in gaps. The APSC’s Data, Digital and Cyber Workforce Plan 2025-30 describes the skills needed for digital in the Australian Public Service. It omits user-centred design and digital product from its critical and in-demand list — service designers, user researchers, interaction designers and product managers. These are the roles the DTA's own Digital Experience Policy depends on. One organisation sets the practice standard, and another sets the people strategy. They don’t join up.
This is not an isolated example. The DTA’s Digital and ICT Investment Oversight Framework is considered a genuine success. The subject of a dedicated OECD review as a leading example of digital investment governance. But strong individual frameworks don't automatically add up to a coherent whole. Each is working from its own picture of what good looks like. The confusion falls on the people trying to deliver services. And eventually it shows up in what gets built.
Having standards is not the same as being able to apply them
The DTA's Digital Experience Policy sets a credible standard for federal digital delivery. But without meaningful enforcement mechanisms or the enabling tools that make the standard easier to follow, compliance depends on the goodwill and capability of individual teams to make the case internally. In practice, the standard is met where the conditions already exist and missed where they don't, which is exactly where it is most needed.
What enabling conditions look like in practice is visible in the UK experience. GDS had the authority to ‘gate’ digital funding for programmes that didn't meet its standards, an approach with some parallels to what Australia's Investment Oversight Framework sets out to achieve. It also invested in shared guidance and tools that made good practice the easier choice. The GOV.UK design system is the clearest example of this, used across the UK government by millions of users each day, reducing duplicated effort and setting a consistent public-facing benchmark. As digital maturity has grown, that approach has spread across government. The Ministry of Justice (MOJ) Design System and NHS Design System are good examples created for their own distinct service needs from the same underlying principles. Rob McCarthy at MOJ describes these as necessary ‘building blocks’ for reusable, accessible and consistent administrative and operational product design.
Australia decommissioned its equivalent in 2021. It has not been replaced. The result is a gap where the central enabling tools should be. Shared tools are not a luxury. They are how standards become practice at scale.
Australia's global ranking is real. So is the gap beneath it
None of this is to say Australia isn't making real progress. In 2025, Australia ranked second in the OECD Digital Government Index (an internationally recognised measure of digital government performance), achieving the top result in 'digital by design'. That ranking reflects investment in: the Data and Digital Strategy, the Investment Oversight Framework, and the Digital Experience Policy. That result is worth acknowledging.
But the index measures what it can measure: frameworks and policies, not whether they are consistently reaching the teams where services are actually built across government. Nor can it assess how digital policy practically integrates across state and federal boundaries.
The federal Data and Digital Strategy is candid about this. It acknowledges that nearly a decade of ‘underinvestment’ and ‘years of outsourcing’ has left internal digital capability below where it should be, making consistent digital practice across government harder.
Robodebt happened while Australia was performing well on these measures. It demonstrates that rankings and reality do not always align.
But why does all of this matter more now?
The conditions described above (fragmentation, policy incoherence and uneven capability) are not problems on the horizon. They are the landscape AI is arriving in now.
The 2025 DTA Policy for responsible use of AI sets mandatory requirements for strategy, oversight and accountability. It is a notable step, but the framework pilot found that agencies are already integrating AI “without the benefit of established expertise or time-tested frameworks”. Each state government and department determines its own approach in practice, duplicating effort at every level.
The pattern is the same: a framework at the centre, uneven capability to apply it across the system. AI doesn't change that equation. It accelerates it.
More than good intent
Australia has the frameworks and, in places, the ambition to match them. It has demonstrated it can move fast when it needs to. What it has not yet built, consistently and at scale, is the connected capability to make those real at the level where services are actually delivered. Building that capability takes sustained investment, coherence and time, more than any single programme, strategy, or political cycle allows.
AI is arriving into the conditions that exist today, amplifying what is there — the good and the bad alike. The opportunity is real, but it depends on closing the gap between intent and delivery capability before AI widens it.
I have spent time inside both systems. The lesson isn't that Australia should replicate the UK model. The structural differences of the Australian federal system make that neither possible nor appropriate. It is that the conditions for digital transformation can be built, and Australia has the foundations to do it. That requires investment in the practical capability that makes frameworks real, and a clear-eyed view of where digital maturity actually is, not just where the rankings suggest.
—-
This connects to a broader pattern I've written about elsewhere. Check out, Working services that aren't — and why they keep failing the people who use and deliver them, How to understand ‘services’ as a tool for your whole organisation, How to define ‘service’ and Design bias in service blueprints.
