Appendix B of Evolve: The Operating Model AI Demands (Hodgson, 2026) places this page in the investment governance set supporting the third structural commitment in Chapter 15: consumption-speed finance. The set is written for the programme office, the CFO, and the finance director, and assumes the structural decision has already been taken. This page extends it with the accounting policy decisions the finance function must resolve before the first continuous investment cycle, and the step-by-step transition mechanics for moving from annual stage-gate allocation to rolling quarterly review.
Three shifts this transition must address
Three shifts are essential for investment governance of products: from point-in-time approval to continuous portfolio steering; from deliverable-based success criteria to outcome-based success criteria; and from capitalised to consumption-based expenditure governance (Hodgson, 2026, Ch. 6).
The most consequential is the shift from point-in-time approval to continuous portfolio steering — funding that is reassessed against current evidence of value, adjusted on a rolling cadence, and directed by what the evidence shows rather than a commitment made at an annual gate. This is what the investment governance set is designed to produce.
The shift from deliverable-based to outcome-based success criteria is the one finance functions find most difficult to accommodate, because it removes the clean closure event that capital accounting is structured around. The best response is to resolve the accounting problem rather than to let it determine the investment strategy. That is what this page provides: the accounting policy decisions that must precede the governance change, and the mechanics for completing the transition in sequence.
Underlying both is the third shift: from capitalised to consumption-based expenditure governance. Once financial decisions happen at the point of use — inside a Sprint, or at the point where an AI agent executes a task — the delegation to govern that spend must sit with the team making the decision. Persistent product funding is the governance structure that makes that delegation rational. This shift also covers cloud infrastructure and AI token spend, which the companion page for consumption-based expenditure governance addresses using FinOps Foundation standards (FinOps Foundation, 2024).
Investment governance reform is the first step in the structural sequence the book describes (Hodgson, 2026, Ch. 13). Table 4 confirms the order: investment governance and performance measurement precede team design and decision rights, which precede AI augmentation at scale. The order is fixed because the causal logic runs through incentives rather than intentions. A persistent product team funded through stage gates reverts to project behaviour because the funding model makes project behaviour rational. Appendix E, Section E.1 provides investment governance maturity levels for organisations assessing their current state before starting Step 1.
Accounting treatment under persistent product funding
The critical difference between stage-gate and continuous investment from an accounting perspective is not the quantum of spend — it is the capitalisation trigger. Under stage-gate governance, the business case approval at a gate initiates the development phase, and spend against that approved scope is capitalised as an intangible asset in progress. The approval event creates the accounting boundary. Under continuous investment, there is no approval event of that kind. The product team works without project boundaries, which means the capitalisation boundary must be reconstructed from the activity itself.
This is the accounting policy decision most organisations miss when they change their investment governance model. They alter the approval rhythm but leave the capitalisation policy unchanged. The result is a capitalisation approach that still depends on project milestones that no longer exist — which produces either systematic under-capitalisation (expensing development activity that should be capitalised) or systematic over-capitalisation (capitalising research activity that does not qualify). Neither error is discovered until the next external audit cycle, at which point the correction is a retrospective exercise across multiple quarters of spend.
The shift from deliverable-based to outcome-based success criteria removes the clean closure event that capital accounting is structured around. Resolving the accounting problem — rather than letting it determine the investment strategy — is what the capitalisation policy and the activity log together achieve. That resolution is the first step in the transition, not a follow-on.
Under AASB 138 Intangible Assets — which mirrors IAS 38 (International Accounting Standards Board, 2004) for organisations outside Australian reporting requirements — development phase expenditure is capitalised when the organisation can demonstrate: technical feasibility of completing the asset; the intent and ability to complete and use or sell it; the existence of a market or internal use case; the availability of adequate resources; and the ability to measure expenditure reliably. Research phase expenditure is expensed as incurred. The challenge under continuous delivery is that a single Sprint may contain both research and development activity. A product team exploring a new feature set in one Sprint and building a tested specification in the next does not have a clean phase boundary — and the capitalisation policy must account for this.
Rather than using gate approvals as the capitalisation trigger, the product team classifies each unit of work at the Sprint level as research phase (exploration, feasibility testing, discovery activities, hypothesis validation) or development phase (building to a technically feasible, defined specification). The finance function applies the capitalisation rate to each category. At each quarterly investment review, the product team provides an activity breakdown confirming the capitalisation allocation for the preceding quarter. Finance reconciles this to the ledger before the quarterly review closes.
Three things must be in place before the first continuous investment cycle begins. The capitalisation policy document must define what constitutes research-phase activity in this organisation’s product context; what constitutes development-phase activity; how time is allocated when a single team member performs both in a Sprint; the minimum documentation standard for the activity log; and the review cadence for confirming the allocation. The finance team must be briefed on the policy and able to apply it — the activity classification cannot be delegated entirely to product teams without finance involvement, because the classification question is an accounting question, not a delivery question. The audit committee must have sight of the policy before the first cycle begins, so that the quarterly review evidence sits within a framework the committee has already endorsed.
The classification criteria under AASB 138 and IAS 38 — what qualifies as research phase versus development phase — apply regardless of lifecycle stage. What changes by stage is the investment posture, and therefore the expected distribution of research-phase versus development-phase work. Table 3 in Chapter 6 is explicit: Introduction products carry a structured test with go/no-go criteria; Growth products carry scaled commitment and rolling investment; Maturity products carry optimisation-focused investment; Decline products carry an explicit funding withdrawal decision (Hodgson, 2026, Ch. 6). An Introduction product running primarily discovery Sprints will have a substantially higher research-phase allocation than a Maturity product running optimisation work — not because the criteria differ, but because the work differs. The policy documents the classification criteria once; the activity log records how each Sprint’s work maps to them.
A Decline stage product requires an explicit Pull decision before the accounting treatment changes to expensing. The product does not drift into Decline-stage accounting; the Pull posture is decided at the portfolio review. The Product Manager triggers the Pull recommendation; the executive decides it. A Decline product reports the cost of migrating its customers to whatever replaces it — not a current value or unrealised value statement (Hodgson, 2026, Ch. 15). The persistent team carries the product through every lifecycle stage, so build, operate, maintain, and retire sit in a single budget line that the portfolio review assesses against the value the product delivers. Total cost of ownership as a governed quantity is only visible under persistent product funding.
Stage-gate to quarterly review: seven steps
This transition is the first step in the structural sequence the book describes (Hodgson, 2026, Ch. 13). Investment governance reform comes before team design and before AI augmentation at scale because it changes what the organisation’s funding model makes rational — and team behaviour follows from that before it follows from any communications program or role redesign. A product team that operates under persistent, outcome-based funding holds its accountability differently from one that still runs Sprints against stage-gate approval.
Transitioning from an annual allocation cycle to rolling quarterly investment review requires seven steps in sequence. Each step has a dependency on the step before it. Compressing the sequence — specifically, attempting the governance change before the accounting policy is resolved — creates the audit liability the transition is designed to avoid.
Step 1: Document the capitalisation policy. Before any governance structure changes, the finance function documents and the audit committee approves the capitalisation policy under continuous delivery. This is a finance-function deliverable, not an IT or product-function deliverable. The accountable authority is the CFO. The output is a signed policy document. Organisations that initiate the governance change and address the accounting policy retrospectively manage a capitalisation correction across multiple quarters of spend — which introduces audit risk because the allocation decisions are made after the fact.
Step 2: Assess portfolio readiness. Each product in the portfolio is assessed against two dimensions: whether the value stream has a Product Manager accountable for the investment envelope, a maintained Product Backlog, and a regular review cadence; and whether the investment data is reliable enough to support the activity-based capitalisation treatment. The Product Manager’s accountability includes owning the investment hypothesis and reporting against the evidence cadence the portfolio requires (Hodgson, 2026, Ch. 15). Products that do not meet the threshold on both dimensions remain on stage-gate governance until they do. The transition is a product-by-product migration, not a portfolio-wide cut-over. Appendix A provides the Layer 1 Investment Governance diagnostic for organisations that have not yet run a formal readiness assessment.
Step 3: Issue the delegation instrument. The accountable authority signs a delegation instrument authorising the quarterly review as the investment decision mechanism. This instrument specifies: the Product Manager investment envelopes for each product (the committed funding within which the team operates — not a per-cycle approval cap); the investment hypothesis format and evidence cadence standard the portfolio review requires; the composition and authority of the portfolio review, which reallocates funding across products but does not approve individual product decisions; and the posture vocabulary — persevere, pivot, pause, pull (Hodgson, 2026, Ch. 10). The Pause and Pull postures change the investment hypothesis; the Product Manager triggers them, the executive decides them. The delegation instrument is the primary document an audit committee will examine if a quarterly review decision is subsequently challenged.
Step 4: Run the shadow period. For the first two quarters after the delegation instrument is signed, quarterly review decisions are made in parallel with the existing stage-gate process. Stage-gate decisions are implemented; quarterly review decisions are documented but not yet binding. The shadow period produces the evidence set for the first real quarterly review and identifies products where the activity log quality is insufficient to support the capitalisation treatment. Any product that cannot produce a reliable activity breakdown by the end of the shadow period is not ready to transition.
Step 5: Execute the first live quarterly review. In the first live cycle, stage-gate governance is suspended for products that passed the shadow period. The portfolio review issues investment decisions. The primary audit artefacts are the investment hypothesis, the evidence cadence, and the decision record at each review — recording what evidence was presented, which posture was taken, and who decided (Hodgson, 2026, Ch. 15). The KVA (Key Value Area) evidence varies by lifecycle stage (Scrum.org, 2024): Introduction products report against Time to Market and Ability to Innovate; Growth products against the Current Value threshold in the investment hypothesis; Maturity products against whether Unrealised Value still justifies development investment; Decline products against the cost of migrating customers. The investment hypothesis anchors which KVA evidence is primary at each stage.
Step 6: Conduct the retrospective audit check. At the end of the first live quarter, the internal audit function conducts a spot check on three products: one with a high capitalisation rate, one with a low capitalisation rate, and one that changed phase during the quarter. The check confirms that the activity logs support the capitalisation treatment applied and that the evidence package meets the standard in the delegation instrument. Deficiencies are corrected before the second quarterly review. This step is not optional — it is the mechanism that confirms the transition is not creating an audit liability.
Step 7: Decommission stage-gate governance for transitioned products. Once the retrospective audit check returns a clean result for three consecutive quarters, the stage-gate approval process for transitioned products is formally closed. The quarterly review is the sole investment governance mechanism for those products from that point. Sprint-based delegation — team decisions within the Sprint, Product Manager decisions within the investment envelope, portfolio review decisions on reallocation — is the key financial mechanism persistent funding requires (Hodgson, 2026, Ch. 15).
Audit compliance during the transition
The audit requirements do not change when the investment model changes. The obligation to demonstrate proper use of resources, sound financial management, and accountable decision-making applies equally under both models. What changes is the evidence that satisfies those requirements.
Three forces protect project-based investment governance and create structural resistance to this transition: financial frameworks calibrated to annual capital budget cycles and project-level expenditure tracking; procurement arrangements built around statements of work with predefined deliverables and payment milestones; and audit and assurance methodologies developed for project-based investment (ANAO, 2026; Hodgson, 2026, Ch. 6). None of these forces is immovable, but each requires deliberate executive action to address. Persistent product funding produces a stronger audit position than stage gates — not a weaker one — because the organisation can demonstrate continuous accountability for outcomes against the investment, rather than point-in-time accountability for deliverables against a business case whose assumptions may have shifted years ago (Hodgson, 2026, Ch. 6).
The primary audit artefacts under continuous investment governance are the investment hypothesis, the evidence cadence, and the decision record at each portfolio review (Hodgson, 2026, Ch. 15). These constitute artefacts more auditable than a business case written before the organisation understood what it was building, because they include what the organisation learned and what it did in response. The delegation instrument, the quarterly review minute, and the activity log remain as supporting documents, but the investment hypothesis and the decision record are what demonstrate that a qualified decision-maker made an informed investment decision on the basis of credible evidence.
The quarterly review evidence standard must be defined in the delegation instrument. The minimum evidence structure covers: a current value statement, an unrealised value estimate, a cost and capitalisation statement, and a forward investment request. The investment hypothesis is the interpretive frame that tells the reviewer whether the evidence confirms continued investment. The posture vocabulary — persevere, pivot, pause, pull — replaces the previous language of “continuation, reduction, or stop” throughout. Pause and Pull postures change the investment hypothesis and require the portfolio review to decide; the Product Manager triggers them, the executive decides them. A quarterly review that does not produce a completed decision record for each product, recording which posture was taken and who decided, is not producing what an audit examination will look for.
Three controls must be maintained throughout the transition regardless of which products are on which governance model. The investment register must reflect the current approved investment level for every product, updated after each quarterly review. The capitalisation ledger must be reconciled to the activity logs at each quarter end. The delegation authority must be current — if the accountable authority changes, the delegation instrument is reissued before the next quarterly review.
The transition period presents a specific compliance risk: products on quarterly review governance and products on stage-gate governance coexist in the same portfolio. The internal audit function must be briefed on which products are on which model and in which quarter the transition occurred. Without this briefing, an audit inquiry that examines the investment trail across the portfolio will encounter two different governance structures and may characterise the quarterly review evidence as a departure from the approved model rather than an approved alternative — a distinction that matters significantly if a quarterly review decision is subsequently challenged.
Worked example: twelve-product portfolio
A central government agency with twelve products across all four lifecycle stages illustrates how the transition sequence operates. The portfolio consists of three Introduction products (I1, I2, I3), four Growth products (G1, G2, G3, G4), three Maturity products (M1, M2, M3), and two Decline products (D1, D2).
The CFO initiates the transition in Q2 Year 1 by directing the finance function to produce the capitalisation policy document. The policy is presented to the audit committee in Q3 and approved. The portfolio readiness assessment is completed across all twelve products by the end of Q3.
Six products meet the readiness criteria and are cleared for transition: G1 and G2 (Growth), M1, M2, and M3 (Maturity), and D1 (Decline). Six are deferred: I1, I2, and I3 (Introduction — product teams do not yet have maintained backlogs or a regular review cadence); G3 and G4 (Growth — Sprint-level activity tracking does not distinguish research phase from development phase reliably); and D2 (Decline — no activity log maintained). Cleared plus deferred equals twelve.
The delegation instrument is signed in Q3 Year 1 for the six cleared products. It specifies Product Manager investment envelopes for each product, not per-cycle approval caps; the investment hypothesis format and evidence cadence standard; and the portfolio review’s authority to reallocate funding across products. The portfolio review does not approve individual product decisions.
The shadow period runs in Q4 Year 1. M1, M2, M3, and D1 produce activity logs of sufficient quality and pass. G1 and G2 produce incomplete logs — research-development classification was applied inconsistently across Sprint team members — and continue in the shadow period for one additional quarter.
The first live quarterly review runs in Q1 Year 2, covering four products: M1, M2, M3, and D1. G1 and G2 complete their second shadow quarter in parallel.
Before the Q1 Year 2 portfolio review, M2 states its investment hypothesis: “We are investing because this product’s Current Value is sustained by three annual enterprise licence renewals totalling $4.2 million. We expect to maintain that value while reducing per-transaction support cost by 15% through the optimisation Sprint cycle. Funding continues while CV holds and support cost trends downward. If CV falls below $3.5 million or support cost does not move within two review cycles, the Product Manager reports which explanation applies and the portfolio review decides the posture.”
Posture decisions at Q1 Year 2: M1 and M2 persevere — CV stable, Unrealised Value still justifies optimisation investment. M3 receives a Pull decision: Current Value has been declining for two consecutive quarters and there is no Unrealised Value pipeline that would justify the current investment level — the diagnostic confirms genuine stall, not lagging indicators. The Product Manager triggers the Pull recommendation; the executive decides it at the portfolio review. M3 enters wind-down. D1 also receives a Pull decision: the primary KVA evidence is the cost of migrating customers to the successor platform, and continued investment in the existing system no longer serves the portfolio strategy. The Product Manager triggers; the executive decides. D1 enters customer migration. G1 and G2 pass their second shadow quarter. Q1 Year 2 totals: four products in live quarterly review, two in shadow, six on stage-gate — twelve.
The retrospective audit check in Q2 Year 2 examines three products. M1 — high capitalisation rate, 85% development phase — has clear activity log evidence supporting the classification. M2 — low capitalisation rate, 30% development phase — has a classification anomaly: two Sprints were recorded as research phase despite the product having confirmed technical feasibility. Finance corrects the allocation and updates the policy guidance. G1 — which transitioned from research to development phase mid-quarter — has a correctly maintained log recording the transition date and the activity split for each period. The audit check returns a qualified result: one correction required, one policy clarification added. A qualified result is not a clean quarter. The three consecutive clean quarters required for Step 7 begin from Q3 Year 2.
Q3 Year 2 is the first clean quarter. I1 and I2 (Introduction) have now developed maintained backlogs and review cadences; they enter the shadow period. Q4 Year 2 is the second clean quarter. Q1 Year 3 is the third — three consecutive clean quarters confirmed. Stage-gate governance is formally decommissioned for M1, M2, G1, and G2.
End state: four products on quarterly review with stage-gate closed (M1, M2, G1, G2); two Introduction products in the shadow period (I1, I2); four products on stage-gate with documented readiness pathways (I3, G3, G4, D2); two products in managed wind-down (M3, D1). Total: twelve. The transition does not require all twelve products to move simultaneously — it requires the capitalisation policy to be in place before any of them do.
Artefacts the finance function needs before day one
Three artefacts must exist — not be drafted in parallel with the governance change — before the first continuous investment cycle begins.
Capitalisation policy outline
The policy document answers these questions, in this order:
- What constitutes research-phase activity in this organisation’s product context? (Exploration, feasibility testing, discovery Sprints, hypothesis validation.)
- What constitutes development-phase activity? (Building to a technically feasible, defined specification.)
- How is time allocated when a single team member performs both in the same Sprint? (Practical default: Sprint-level declaration against a defined classification threshold, or a weekly time log where declaration is unreliable.)
- What constitutes a minimum documentation standard for the activity log? (What evidence is required to support the capitalisation treatment — sufficient for an internal audit spot check.)
- What is the review cadence for confirming the allocation? (Quarterly, at the portfolio review.)
- Who approves the policy? (CFO. The audit committee reviews it before the first cycle begins.)
Product teams complete the classification for their context. Finance validates and applies the policy. The accounting question cannot be delegated entirely to delivery teams.
Activity classification table
The table maps common product team activities to their phase classification:
| Activity | Classification |
|---|---|
| Discovery Sprints | Research phase |
| Feasibility testing | Research phase |
| Hypothesis validation | Research phase |
| User research and prototyping | Research phase |
| Design-approved build Sprints | Development phase |
| Technical architecture (feasibility confirmed) | Development phase |
| Iteration on shipped feature against defined specification | Development phase |
| Bug fixes on released feature | Development phase if producing technically defined improvement; expense if maintenance — check policy |
Product teams complete the table for their context. Finance validates it. The classification is an accounting question, not a delivery question.
Decision record template
The decision record is the primary audit artefact under continuous investment governance (Hodgson, 2026, Ch. 15). A quarterly review that does not produce a completed decision record for each product is not producing what an audit examination will look for.
Minimum fields:
- Product name and lifecycle stage
- Investment hypothesis in force (the four sentences: rationale, measurable evidence, continuation threshold, decision rule)
- KVA evidence reviewed (the measures and their current values, lifecycle-differentiated)
- Posture taken: persevere / pivot / pause / pull
- Decision authority: who decided
- Next review date
The decision record replaces the stage-gate approval trail as the primary evidence of accountable investment governance. Its completeness is the test of whether the transition has been executed or merely announced.
The accounting policy must be signed before the first continuous investment cycle begins — not drafted once the delegation instrument is already in place. Finance functions that start with the governance change and address capitalisation retrospectively manage a correction across multiple quarters of spend that surfaces at the next external audit cycle. The delegation instrument, the investment hypothesis, and the decision record together replace the stage-gate approval trail as the evidence of accountable investment governance. The finance function that has all three in place before the first cycle does not discover the accounting problem at audit — it does not have one.
References
- AASB. (2021). AASB 138 Intangible Assets (Compilation 4). Australian Accounting Standards Board. https://www.aasb.gov.au/admin/file/content105/c9/AASB138_08-15_COMPmar20_07-21.pdf
- Australian National Audit Office. (2026). Artificial intelligence use in IP Australia. Australian Government. https://www.anao.gov.au/work/performance-audit/artificial-intelligence-use-ip-australia
- FinOps Foundation. (2024). FinOps framework. FinOps Foundation. https://www.finops.org/framework/
- Hodgson, M. (2026). Evolve: The operating model AI demands. Zen Ex Machina.
- IFRS Foundation. (2004, updated 2026). IAS 38 Intangible Assets. IFRS Foundation. https://www.ifrs.org/issued-standards/list-of-standards/ias-38-intangible-assets/
- Scrum.org. (2024). The evidence-based management guide: Improving value delivery under conditions of uncertainty. Scrum.org. https://www.scrum.org/resources/evidence-based-management-guide
