Turning skills frameworks into workforce capability across central and state government in Aotearoa New Zealand and Australia
Public services across Aotearoa New Zealand and Australia are being asked to do two demanding things at once: deliver more from a workforce that is flat or shrinking and under tight fiscal control, and deepen the capability they hold in-house rather than rely on external labour. Neither can be met without knowing, in detail, what skills the workforce holds, where the gaps are, and how to close them. Most agencies have already invested in capability frameworks, however, a framework and the assessment it enables is necessary but not sufficient.
For agency and central-services workforce, people and digital leaders · Aotearoa New Zealand & Australia · 2026
THE SHORT VERSION
Public services in Aotearoa New Zealand and Australia are being asked to do two demanding things at once. They must deliver more from a workforce that is flat or shrinking and under tight fiscal control, and at the same time they must deepen the capability they hold in-house rather than rely on external labour. In Australia the second demand is now explicit policy, through the APS Strategic Commissioning Framework. In New Zealand it is the practical consequence of sustained savings targets alongside a stated commitment to lift system-wide capability. Neither demand can be met without knowing, in detail, what skills the workforce holds, where the gaps are, and how to close them.
Most agencies have already invested in capability frameworks to answer the first part of that question. The Skills Framework for the Information Age, the Data Capability Framework, the Policy Skills Framework, the Māori Capability Framework, Leadership Success Profiles and the Australian Public Service work level standards all describe what good looks like. However, a framework, and the assessment it enables, is necessary but not sufficient. The value is realised only when assessment feeds a wider chain: standardising roles, tracking capability, developing people, and planning the workforce, at both agency and system scale. This paper sets out that capability value chain, shows how this work serves two distinct readers, the individual agency and the centre, and describes how to begin without requiring major procurement.
The argument is platform-agnostic in principle, and many tools support parts of the chain. We illustrate it with TalentJam because our platform was built for the public-sector from the ground up, rather than retrofitted from a generic human-resources product, but the model matters more than the tool, and a reader should test any platform, including ours, against it.
PART ONE · THE PROBLEM
The context across both jurisdictions has converged on the same point, from different directions.
The Australian Public Service (APS) Strategic Commissioning Framework, in force since 2023, sets the expectation that core work is delivered by public servants and not inappropriately outsourced. Agencies identify their core tasks, functions and roles, set annual targets to bring that work in-house, and report progress in their corporate plans and annual reports, in the capability section. The stated intent is to deepen capability over time and reduce risks to expertise, integrity and public trust. The first reporting cycles showed strong appetite to insource digital, data and cyber work in particular, and equal candour that it will be hard, because those skills are scarce, the work comes in peaks and troughs, and internal talent is difficult to mobilise. The APS Workforce Strategy 2025 and the Commission's Workforce Planning Centre of Excellence point the same way: people and the capabilities they bring are the key asset, and workforce planning is the discipline that makes the most of them.
There is a hard dependency hidden in all of this. An agency cannot decide what to insource, sequence the rebuild, or evidence its progress, unless it can see its in-house capability at the level of roles and skills. Strategic commissioning is, underneath, a skills-data problem.
New Zealand has arrived at the same place by a fiscal route. Agencies have been required to find savings of around 7.5 per cent, Public Service staffing has been held flat to falling, with roughly 62,700 full-time-equivalent staff across the Public Service in mid-2025. The wider public sector employs close to 477,000 people across central and local government, the health and education sectors and Crown entities. Against that backdrop, Te Kawa Mataaho Public Service Commission (PSC) has set out a clear ambition for a unified, trusted and high-performing service that is future-focused, digitally enabled and fiscally responsible, and that lifts system-wide capability. The Public Service Amendment Act, passed in 2026, sharpens the system's focus on performance and stewardship, and the Commission's Future of Work programme looks ahead to the capabilities a 2035 workforce will need. The centre is also consolidating: from 1 April 2026 the Government Digital Delivery Agency, which carries the former Government Chief Digital Officer functions, sits within the Public Service Commission, bringing system digital, data and capability leadership alongside the workforce-system steward.
When the workforce cannot simply grow, capability has to be found, moved and grown inside the system. That requires the same thing strategic commissioning requires across the Tasman: a clear, current and comparable picture of the skills the workforce holds.
Layered over both is the arrival of artificial intelligence as a workforce question, as well as a technology one. New Zealand's Public Service AI Framework, issued in 2025 by the Government Chief Digital Officer, which sits within the National AI Strategy, makes capability one of its six pillars, building the AI knowledge and skills of public servants, alongside governance, guardrails, innovation, social licence and a global voice. Australia's reform agenda treats data, technology and flexible workforce models as a central focus. In both, the practical question for a workforce leader is the same: which roles and skills are exposed to change, on what timeline, and what does the reskilling path look like. That, too, is answered from skills data, not from intuition.
In short: both systems now need to see, develop, move and plan their workforce by skill, at agency and at system level. That is precisely what the capability value chain delivers. The historical isolated assessment approach will not meet the needs of the future workforce.
Capability frameworks do real work. They give a shared language for skills, a defensible basis for assessment, and a standard against which development can be measured. But two limits are worth naming plainly, because they explain why framework investment so often fails to deliver the expected value.
First, every framework has gaps, horizontally and vertically. No single framework covers every role an agency runs, and even within its span it tends to under-describe three things: the new (emerging practice, new tools and technology), the local (specific legislative, regulatory or cultural requirements) and the old (legacy integration and the real processes that hold an organisation together). A framework is a map, and the territory always exceeds it. Agencies therefore run several frameworks at once, and need the gaps filled with local and custom content.
Second, assessment is only one link in a longer chain. Assessing a person against a framework produces a data point. Value comes from what that data point feeds: standardised roles, a live capability picture, targeted development, mobility, and workforce and training-demand planning, at multiple scales. An organisation that assesses but does not connect the rest of the chain has paid for the framework and left the return on the table.
The naming those steps in the capability value chain makes those links explicit and deliberate.
| Stage | What it does | Who leads |
|---|---|---|
| Stewardship of frameworks | Maintain, curate and keep current the frameworks in use, and the local and custom content that fills their gaps | Framework steward / centre |
| Standardised role descriptions | Define roles and their required skills consistently, within and between agencies | Heads of profession, with the centre and agencies |
| Assessment against frameworks | Evaluate people's skills against the agreed standard | Agency |
| Aggregating skills data | Build a current capability picture at agency and system scale | Agency and centre |
| Aggregating workforce-planning data | Connect capability to demand, budget and headcount | Agency and centre |
| Forecasting long-term skills demand | Plan recruitment, development, mobility, training provision and AI transition | Centre, with the training system |
Two design principles run through the whole chain. It should complement, not replace, the systems an agency already runs, leveraging existing human-resources and financial-management information systems rather than displacing them, in the spirit of small, interoperable parts doing one thing well. And it should aim for the most incremental benefit for the least effort, sequencing by value rather than attempting a single transformation.
The value chain serves two distinct user groups, and any paper or platform that conflates them will satisfy neither. The first is the agencies: a department, agency, or a Crown entity, managing its own capability. The second is central services, which in this paper means primarily the system stewards (PSC/GDDA in New Zealand and the APS in Australia), but also the shared corporate-services functions that deliver common back-office capability across agencies. Their questions differ, and so do the outputs they need.
| Dimension | The agency | Central services (stewards and shared services) |
|---|---|---|
| Primary question | What capability do we hold, and where are the gaps against the work we must deliver? | Where does capability sit across the system, and how do we match it to demand and prioities? |
| Core uses | Role design, assessment, development, in-house delivery, workforce planning, regulatory evidence | Standardisation, cross-agency benchmarking, mobility, system workforce and training-demand planning, stewardship |
| What they must evidence | Capability and its progress, to the centre and to regulators | System capability and value, and Treaty and Indigenous-data obligations |
| The risk if absent | Cutting or insourcing blind, and losing critical capability unseen | No system view, duplicated effort, and capability mismatched to demand. Resource allocation based on local but not system level priorities |
The two scales are not separate systems; they are the same capability data, read at different altitudes. An agency's standardised roles and skills assessments, aggregated and made comparable, become the centre's view of where capability sits and where it is scarce. That is what allows pockets of capability to be identified and matched to demand across agencies, mobility to be planned rather than improvised, and training to be commissioned against real, system-wide need. The shared corporate-services dimension matters here too: where back-office and corporate roles are described through a common process and role taxonomy, capability can be described, reported, and planned consistently across every agency that shares those services.
PART TWO · THE SOLUTION
A credible capability model does not ask an agency to abandon its existing frameworks. It holds them, cross-references them, and fills their gaps with local and custom content. A single role frequently spans more than one framework, a technology role assessed against SFIA may also carry data capabilities described by the DCF, and a policy role may need both the Policy Skills Framework and Māori capability, so the model must produce role profiles that reference several frameworks at once and analyse gaps across all of them.
| Framework | Domain it describes | Typical steward |
|---|---|---|
| SFIA 9 | Digital, data and technology skills, across all levels | SFIA Foundation; used widely in NZ and AU government |
| Data Capability Framework (DCF) | Data and analytics capability | Adopted across NZ and AU government |
| Policy Skills Framework (PSF) | Policy practice and craft | DPMC (New Zealand) |
| Māori Crown Relations Capability Framework (MCR) | Māori capability across the public service | PSC/Te Puni Kōkiri |
| Leadership profiles (for example LSP) | Leadership capability and succession | PSC / APS |
| Common Process Model (CPM) | Corporate and back-office service roles | PSC/GDDA |
| APS work level standards and craft frameworks | APS roles, classifications and professions | APS |
| DDaT and jurisdiction-specific frameworks | Digital, data and technology in some jurisdictions | Various; several derive from SFIA |
| Agency-specific and custom frameworks | The new, the local and the old that standard frameworks miss | The agency, curated centrally |
State public services in Australia, such as those of New South Wales and Victoria, maintain their own capability frameworks and commissions; the model holds for them in the same way it does for the APS, with the relevant framework substituted in. The point is not which frameworks an agency uses, but that the capability model can hold all of them in one integrated picture, rather than forcing a choice between them or leaving the gaps between them unmanaged.
Workforce capability data is sensitive, and in government it is doubly so. A capability model that is not trustworthy by design will not, and should not, be adopted. Three commitments are non-negotiable for this audience.
In Aotearoa, Māori data is a taonga that requires culturally grounded protection and care. The Māori Data Governance Model developed by Te Kāhui Raraunga, designed by Māori data experts for use across the public service and consistent with the Crown's obligations under te Tiriti o Waitangi, is the reference point, and agencies are expected to draw on it when updating how they protect and share Māori data. For a capability platform that may hold data about Māori public servants, or capability described through the MCR, this means values-led governance, consented use, the ability to keep data within appropriate boundaries, and rohe-aware reporting, following the model rather than retrofitting it. In Australia, the parallel obligation is Indigenous data governance, and the CARE Principles for Indigenous Data Governance, developed through the Global Indigenous Data Alliance, provide an internationally recognised basis. A platform built for this context treats these as architectural requirements, not optional add-ons.
Capability data must be held under the security expectations agencies already operate to: New Zealand's Protective Security Requirements and Australia's Protective Security Policy Framework, security-by-design, integration with agency identity and access management, and transparency about cloud jurisdiction and data residency. Where an agency requires data to remain onshore, or within a particular jurisdiction, that must be a configurable option rather than a negotiation.
Because a modern capability platform uses inference, and because it speaks directly to AI-transition planning, it must align with the AI expectations now set for the public service. New Zealand's Public Service AI Framework sets out governance, guardrails, capability, innovation, social licence and a global voice as its pillars, and agencies are encouraged to assess AI initiatives against its values and rules. The same discipline applies to any AI used within a workforce platform: transparency, human accountability, and use that the public, and public servants, can have confidence in.
Capability data earns its keep when it informs decisions. The later stages of the value chain are where that happens, and they produce different, complementary value for the agency and for the centre.
The thread through all of these is that the record an agency builds to develop and deploy its people is the same record it needs to plan, to evidence, and to steward. Build it once, and use it at every scale.
The barrier to capability work is rarely conviction; it is the size of the first step. Two things make a sensible start achievable inside ordinary agency authorities.
A low-risk first engagement. A fixed-price, defined-scope skills baseline, covering chosen business units and role families and producing a concrete capability picture and gap analysis, fits within the low-value, low-risk procurement rules most agencies operate under. It is a tangible deliverable that tests the approach before any larger commitment, and it gives an executive sponsor something real to judge.
A greatest-benefit-first roadmap. From that baseline, the broader chain is sequenced by value: stand up the capability picture, standardise the roles that matter most, connect the frameworks already in use, then add system-level aggregation, mobility and forecasting as the data and the appetite grow. Much of this can be achieved in months rather than years, because it complements existing systems rather than replacing them.
A pragmatic sequence, drawn from system-level design work, might be as follows: define the data and information flows and the products they serve; baseline current capability and mobility maturity; stand up a skills-based talent picture alongside existing systems; set an implementation roadmap, greatest benefit first; connect the common role and process taxonomy; roll out to agencies in priority order; have heads of profession confirm standardised roles; deliver system-level planning and analytics; define inter-agency mobility; and connect to training providers. Each step is useful on its own, which is what makes the whole achievable.
TalentJam is a skills intelligence platform built for the public-sector context rather than adapted from a generic human-resources product. It holds the frameworks an agency runs, SFIA, the DCF, the PSF, the MCR and others, together with custom frameworks and roles, in a single model, and produces role profiles and gap analysis that cross-reference them. Its Strategic Workforce Intelligence service is designed around public-sector obligations: cross-agency benchmarking, capability reporting aligned to the relevant regulators, demand-management that connects skills to budget and headcount, and AI-transition planning. It is designed to work alongside existing human-resources and financial-management systems, with data-sovereignty and security commitments treated as architecture, and a fixed-price skills baseline as a low-risk way to begin.
None of that displaces the argument of this paper. Many tools support parts of the capability value chain, and an agency should hold any platform, TalentJam included, against the model: does it complement rather than replace, hold your frameworks rather than impose one, serve both the agency and the centre, honour data sovereignty and security, and turn assessment into development, mobility and planning. The model is the point. Any platform choice is how an agency delivers it.
This paper draws on publicly available material current as at mid-2026. Figures and policy settings change, and readers should confirm the latest position with the relevant agency.