An agency head who signals a product operating model transition on Tuesday will receive questions about workforce impact by Thursday. The questions are legitimate: the transition changes what accountabilities exist, what skills they require, and what happens to employees whose current roles the new structure does not retain. The answers are determined by the legislative and industrial framework governing APS employment — and by how early the workforce planning conversation started. An agency that begins that conversation at the same time as the governance redesign has the full sequence available. An agency that treats workforce as a downstream consequence will discover, mid-programme, that the legal framework sets the pace, not the strategic intent.
This page extends Appendix D of Evolve: The Operating Model AI Demands (Hodgson, 2026), which introduces the three workforce populations, the duty reassignment mechanism, and the capability pipeline risk. The implementation detail below — legislative instruments, enterprise agreement procedures, classification mechanics, and case material — is the companion content Appendix D refers readers to.
Every product operating model transition produces three structurally distinct workforce populations. They are not differentiated by seniority or function. They are differentiated by their relationship to the new model.
Three workforce populations, three pathways
The first population — typically 10–15% of the affected workforce — was already operating with the accountability orientation the product model demands before the transition began. They were asking the structural questions the project model provided no formal channel for. This population does not require formal workforce intervention. The governance redesign confirms what they had already identified. The priority for this group is to move them into the product leadership accountabilities quickly — Product Manager, Product COO, Chief Product Architect, Product Owner — before the transition’s uncertainty gives them a reason to leave. An agency that loses this population before the model is operational loses the institutional knowledge the model depends on to function. Product teams require contextual understanding of user needs, legacy system constraints, and ministerial accountability; that understanding cannot be regenerated from scratch once it leaves.
The second population can hold product accountability in routine decisions and escalate high-stakes ones at the right level. They are currently performing functional management work that the product structure displaces — but the displacement is a role redefinition, not a capability failure. The legislative mechanism for this population is duty reassignment under section 25 of the Public Service Act 1999 (Cth), not reclassification. The distinction matters: duty reassignment does not require a new engagement, a reclassification event, or a competitive merit process. It requires the Agency Head to assign the employee duties consistent with their existing classification level.
The third population consists of employees whose functional management work disappears in the new structure and who cannot transition to the product accountability orientation within the transition timeframe. This population triggers the excess employee pathway. The determination is structural — a finding about what the organisation requires, not a performance finding about the individual. Conflating the two creates significant legal risk and workforce relations damage that the transition does not require.
Duty reassignment is the bridge mechanism
Section 25 of the Public Service Act 1999 (Cth) provides that an Agency Head may from time to time determine the duties of an APS employee in the Agency, and the place or places at which the duties are to be performed (Commonwealth of Australia, 1999, s. 25). This is the operative mechanism for moving the second population into product team structures — not the classification process, which comes later.
At the Australian Taxation Office, EL2 officers formed the product leadership layer — Product Manager, Product COO, and Chief Product Architect — under the same classification they had held in the project model. The titles in the HR system did not change immediately. The structure of the work did. The ATO’s approach reflects the sequencing logic section 25 enables: move the structure first, allow the accountabilities to stabilise, then formalise duty statements and run merit selection once those statements can reflect what the roles actually require.
The most common workforce planning failure in operating model transitions is doing this in reverse — writing formal duty statements before the model has operated long enough to reveal what the accountabilities demand. A Product Manager duty statement written before the first Sprint Review describes what the transition plan imagines the work will require. A duty statement written after twelve months of operating the model describes what it actually does. Section 25 creates the bridge between those two moments.
The formal sequence follows under section 22 of the Public Service Act 1999 (Cth), which governs the engagement of APS employees on ongoing and non-ongoing terms (Commonwealth of Australia, 1999, s. 22). Section 23 establishes the Classification Rules that govern work value assessment and formal reclassification decisions (Commonwealth of Australia, 1999, s. 23). Classification decisions are subject to the procedural requirements in the Australian Public Service Commissioner’s Directions 2022, including the documentation, consultation, and authority-level requirements that govern restructuring decisions. An agency that attempts to formalise product team classifications before the duty reassignment period has produced stable, verified role descriptions will produce classifications that do not fit the operational reality of the model it is building — and the cost of correcting misaligned classifications, given the consultation and documentation obligations that revisiting those decisions requires, is substantially higher than the initial role definition work.
When the excess pathway applies
Where the transition produces positions genuinely surplus to agency requirements, section 23(4) of the Public Service Act 1999 (Cth) permits an Agency Head to reduce an employee’s classification without consent on the ground that the employee is excess to the requirements of the agency at the higher classification (Commonwealth of Australia, 1999, s. 23(4)). This provision operates alongside the excess employee procedures in the agency’s applicable enterprise agreement. The Department of the Prime Minister and Cabinet Enterprise Agreement 2024–2027, for example, sets out the notification obligations, retention periods, and redeployment requirements that govern excess procedures in that agency (Department of the Prime Minister and Cabinet, 2024). Most APS enterprise agreements contain equivalent provisions, and the procedural obligations they create run parallel to the structural transition itself — they cannot be collapsed or accelerated to fit a programme timeline.
The redeployment obligation extends across the APS, not only within the home agency. An employee declared excess cannot be separated before the redeployment process has been genuinely exhausted across the service. That process takes time — typically three to twelve months, depending on the enterprise agreement and the employee’s classification level. This is a sequencing constraint. An executive who initiates the excess pathway after the structural change has been announced is running the clock from a position the framework does not accommodate.
Under section 389 of the Fair Work Act 2009 (Cth), redundancy is genuine only when three conditions are all satisfied: the employer no longer requires the job to be performed by anyone; the employer has met any modern award or enterprise agreement consultation obligation; and it was not reasonable in all the circumstances to redeploy the employee (Commonwealth of Australia, 2009, s. 389). The condition most frequently contested in APS restructuring is the third — the redeployment test. This is an objective legislative standard. It asks whether reasonable redeployment options existed and were genuinely available — not whether the agency made an effort. An agency that has not structured a genuine reskilling and redeployment pathway before declaring positions excess will find the third condition difficult to satisfy.
The right sequence is to identify the reskillable cohort, invest in development pathways for that cohort, and bring the excess pathway in only for those the reskilling process cannot accommodate within the transition timeframe. The sequence matters legally, not only operationally: an agency that has not structured a genuine reskilling pathway before declaring positions excess will find the redeployment test under s. 389 difficult to satisfy.
Capability assessment: SFIA and the workforce plan
The Skills Framework for the Information Age version 8 (SFIA Foundation, 2021) provides the capability taxonomy for assessing both the second and third populations. In the APS context, SFIA serves two functions. The first is role definition: the SFIA skill descriptors at each responsibility level describe what practitioners are expected to do with precision sufficient to inform duty statements that are legible both operationally and within the Classification Rules framework. This is the starting point for the formal duty statements that follow the section 25 duty reassignment period.
The second function is population assessment. SFIA enables a comparison of an employee’s current capability profile against target role requirements at a level of specificity that generic position description matching cannot achieve. That comparison determines whether an employee belongs in the second population — able to be developed into product accountability within the transition timeframe — or the third.
SFIA v8 was developed for the commercial technology sector. Government assumptions about team structure, authority, and ministerial accountability are not built in. Applying SFIA without contextual adaptation to APS delivery produces role frameworks that are SFIA-compliant and operationally unworkable. The translation requires both SFIA expertise and direct experience of APS delivery. Engaging external advisers for the SFIA component without internal capacity to validate the APS translation is a known failure mode that has produced frameworks sound on paper and unworkable in practice.
The APS Data, Digital and Cyber Workforce Plan 2025–30 situates capability development within the broader APS labour market (Digital Transformation Agency, 2025). Its workforce pipeline projections — including the risk that nearly 20% of digital professionals will reach average retirement age by 2030 — are decision inputs for the transition, not background reading. An agency that designs a product workforce framework dependent on capabilities the plan identifies as structurally constrained will discover those constraints through failed recruitment and extended vacancies rather than through planning.
The capability AI removes without replacing
The APS Data, Digital and Cyber Workforce Plan 2025–30 addresses the retirement pipeline risk. It does not address what happens when AI removes the routine work through which digital capability was previously built. The executive who automates compliance assessment, standard correspondence, and submission processing has improved throughput and removed the apprenticeship pathway through which junior practitioners developed into senior ones.
SFIA v8 can classify the capabilities the organisation needs. It cannot replace the work through which those capabilities were built. That incidental pathway — the accumulation of experience through routine exposure — was never a designed mechanism. It was a by-product of operational volume. When automation removes the volume, the by-product disappears with it.
The decision about what replaces that incidental pathway is a governance design decision. It must be made before the pathway is removed. An agency that automates first and designs the replacement mechanism second will discover the capability gap when the next generation of practitioners is not there. That is not a human resources problem. It is the consequence of an operating model decision that was not fully scoped at the time it was made (Hodgson, 2026).
The sequencing imperative
The workforce transition is not a downstream consequence of the operating model change. It is a design variable within it. The executive who sequences the workforce conversation after the structural change has already designed those two things separately: the structure through which work flows, and the people who will operate within it. Both are governance design decisions. Both must be made with the legislative and industrial framework in view, because that framework determines the pace at which the transition can proceed, regardless of how urgently the structural case demands it.
References
- Commonwealth of Australia. (1999). Public Service Act 1999 (Cth). https://www.legislation.gov.au/Series/C2004A00538
- Commonwealth of Australia. (2009). Fair Work Act 2009 (Cth). https://www.legislation.gov.au/Series/C2009A00028
- Department of the Prime Minister and Cabinet. (2024). Enterprise Agreement 2024–2027: Section 10—Management of excess employees. Australian Government. https://www.pmc.gov.au/resources/pmc-enterprise-agreement/section-10-management-excess-employees
- Digital Transformation Agency. (2025). APS data, digital and cyber workforce plan 2025–30. Australian Government. https://www.dataanddigital.gov.au/actions-underway/workforce
- Hodgson, M. (2026). Evolve: The operating model AI demands. Zen Ex Machina.
- SFIA Foundation. (2021). Skills Framework for the Information Age (SFIA 8). SFIA Foundation. https://sfia-online.org
