{
  "version": "2.0",
  "structure": "8-Point",
  "methodology": "Workforce Planning Toolkit \u00b7 Four-phase framework \u00b7 8-Point Prompting Check \u00b7 5 Tests verification",
  "prompts": [
    {
      "id": "01",
      "name": "Charter & Scope",
      "reference": "Phase 1 \u00b7 Step 1.1",
      "category": "Drafting",
      "purpose": "Drafts a 2-3 page project charter from kickoff workshop notes.",
      "whenToUse": "After a kickoff workshop with executive sponsors and operational leaders.",
      "triggerPhrase": "Use my Charter prompt",
      "structure": "8-Point",
      "prompt": {
        "priming": "This is a drafting engagement. Treat the output as a draft for my review, not a final deliverable. We will iterate.",
        "context": "I am the Project Lead for the Barwon Water Workforce Planning programme. The audience is the executive sponsor (P&C GM, Digital GM, Finance GM) plus the MD, who will sign off the charter. They are commercially literate, time-poor, and need a charter that protects the project's authority, scope, and resourcing through to end of August. The triple mandate is: (1) build a workforce planning methodology the organisation owns, (2) deliver PS28 financial inputs, (3) take a strategic position on workforce shape under AI / future-of-work conditions.",
        "framework": "Approach this as a senior P&C consultant preparing material for executive sign-off on a strategic programme. Use the Charter format outlined in Section 04 of the Workforce Planning IIS. Apply Working Principle 01.2 (defensibility earns the right to move fast) and 01.4 (the model is the system of record).",
        "instructions": "Produce a 2-3 page charter with these headed sections, in this order: (1) Purpose (2 paragraphs); (2) Scope \u2014 in / out (bulleted); (3) Success criteria (3-5 measurable); (4) Governance (table: Sponsor / Steering / Working Group); (5) Milestones (5 phases \u00d7 dates); (6) Key risks (top 5, with mitigations); (7) Assumptions; (8) Sign-off block. Max 1,000 words.",
        "examples": "A good charter reads decisive, specific, and brief. Purpose names the triple mandate clearly. Scope is broad enough to flex but narrow enough to be defensible. Success criteria are measurable (e.g. 'PS28 financial inputs delivered by 31 August', not 'good outputs'). Risks are real (data quality, stakeholder bandwidth, strategic decision dependency) \u2014 not boilerplate.",
        "rules": "Australian English. Plain language. Active voice. None of the banned phrases in Section 04 of the IIS. Where the charter references workforce shape strategy, frame as a Board-level decision, not a P&C decision. Flag any factual claim, dollar value, date, or named party you cannot fully support from the workshop notes.",
        "iteration": "After producing the draft, ask me three questions: what to refine, what to expand, what to remove. Wait for my answers before redrafting.",
        "assumptions": "If the workshop notes don't clarify the executive sponsor, the budget envelope, the working group composition, or the Steering Group cadence, pause and ask before producing."
      },
      "fullPromptText": "PRIMING\nThis is a drafting engagement. Treat the output as a draft for my review, not a final deliverable. We will iterate.\n\nCONTEXT\nI am the Project Lead for the Barwon Water Workforce Planning programme. The audience is the executive sponsor (P&C GM, Digital GM, Finance GM) plus the MD, who will sign off the charter. They are commercially literate, time-poor, and need a charter that protects the project's authority, scope, and resourcing through to end of August. The triple mandate is: (1) build a workforce planning methodology the organisation owns, (2) deliver PS28 financial inputs, (3) take a strategic position on workforce shape under AI / future-of-work conditions.\n\nFRAMEWORK\nApproach this as a senior P&C consultant preparing material for executive sign-off on a strategic programme. Use the Charter format outlined in Section 04 of the Workforce Planning IIS. Apply Working Principle 01.2 (defensibility earns the right to move fast) and 01.4 (the model is the system of record).\n\nINSTRUCTIONS\nProduce a 2-3 page charter with these headed sections, in this order: (1) Purpose (2 paragraphs); (2) Scope \u2014 in / out (bulleted); (3) Success criteria (3-5 measurable); (4) Governance (table: Sponsor / Steering / Working Group); (5) Milestones (5 phases \u00d7 dates); (6) Key risks (top 5, with mitigations); (7) Assumptions; (8) Sign-off block. Max 1,000 words.\n\nEXAMPLES\nA good charter reads decisive, specific, and brief. Purpose names the triple mandate clearly. Scope is broad enough to flex but narrow enough to be defensible. Success criteria are measurable (e.g. 'PS28 financial inputs delivered by 31 August', not 'good outputs'). Risks are real (data quality, stakeholder bandwidth, strategic decision dependency) \u2014 not boilerplate.\n\nRULES\nAustralian English. Plain language. Active voice. None of the banned phrases in Section 04 of the IIS. Where the charter references workforce shape strategy, frame as a Board-level decision, not a P&C decision. Flag any factual claim, dollar value, date, or named party you cannot fully support from the workshop notes.\n\nITERATION\nAfter producing the draft, ask me three questions: what to refine, what to expand, what to remove. Wait for my answers before redrafting.\n\nASSUMPTIONS\nIf the workshop notes don't clarify the executive sponsor, the budget envelope, the working group composition, or the Steering Group cadence, pause and ask before producing.\n\nWORKSHOP NOTES:\n[Paste workshop notes \u2014 sponsor decisions, scope discussions, agreed governance, milestone constraints, named risks, working group commitments.]",
      "contextBlock": "WORKSHOP NOTES:\n[Paste workshop notes \u2014 sponsor decisions, scope discussions, agreed governance, milestone constraints, named risks, working group commitments.]"
    },
    {
      "id": "02",
      "name": "Stakeholder Engagement \u2014 Mode 1: Interview Guide design",
      "reference": "Phase 2 \u00b7 Step 2.3 (Mode 1)",
      "category": "Drafting",
      "purpose": "Designs the structured interview guide for 12-15 leader interviews.",
      "whenToUse": "When designing the interview guide for Phase 2 leader interviews.",
      "triggerPhrase": "Use my Interview Guide prompt",
      "structure": "8-Point",
      "prompt": {
        "priming": "This is a design engagement. Output is a structured guide for me to use as the interviewer. I will adapt phrasing in the room.",
        "context": "I am preparing to run 12-15 structured 60-minute interviews with senior leaders at Barwon Water. The interviews surface workforce demand signals, capability gaps, single points of failure, and talent management concerns. The synthesis from these interviews feeds the demand-supply driver register and the capability framework.",
        "framework": "Approach this as a senior workforce planning interviewer designing a guide for cross-functional leader interviews. Apply Working Principle 01.5 (human intelligence and data intelligence are complementary). Use the question structure pattern from the Workforce Planning IIS.",
        "instructions": "Produce a structured guide with 5 sections: (1) Business strategy assessment \u2014 3 questions; (2) Organisational assessment \u2014 3 questions; (3) Workforce requirements \u2014 4 questions; (4) Talent management \u2014 3 questions; (5) Open-ended close \u2014 2 questions. For each question: (a) the question itself; (b) a one-line 'why we are asking'; (c) 1-2 follow-up probes. Add a 5-minute opening and 5-minute close.",
        "examples": "Good interview questions are open, neutral, specific, and inviting. They surface signals that don't show up in HRIS. Example: 'Where do you have single points of failure \u2014 one person holding critical knowledge?' rather than 'Are there succession gaps?'",
        "rules": "Australian English. Plain language. No leading questions. No jargon a senior operational leader wouldn't recognise. Maintain neutrality. Flag any question that risks surfacing personal information about specific employees \u2014 those are routed to a 1:1 conversation, not the interview record.",
        "iteration": "After producing the guide, identify three questions that are likely to elicit short answers and propose a probe to deepen each.",
        "assumptions": "If the leaders' areas of focus aren't clear, ask before producing \u2014 the guide may need light tailoring (e.g. capital delivery leaders vs operational leaders)."
      },
      "fullPromptText": "PRIMING\nThis is a design engagement. Output is a structured guide for me to use as the interviewer. I will adapt phrasing in the room.\n\nCONTEXT\nI am preparing to run 12-15 structured 60-minute interviews with senior leaders at Barwon Water. The interviews surface workforce demand signals, capability gaps, single points of failure, and talent management concerns. The synthesis from these interviews feeds the demand-supply driver register and the capability framework.\n\nFRAMEWORK\nApproach this as a senior workforce planning interviewer designing a guide for cross-functional leader interviews. Apply Working Principle 01.5 (human intelligence and data intelligence are complementary). Use the question structure pattern from the Workforce Planning IIS.\n\nINSTRUCTIONS\nProduce a structured guide with 5 sections: (1) Business strategy assessment \u2014 3 questions; (2) Organisational assessment \u2014 3 questions; (3) Workforce requirements \u2014 4 questions; (4) Talent management \u2014 3 questions; (5) Open-ended close \u2014 2 questions. For each question: (a) the question itself; (b) a one-line 'why we are asking'; (c) 1-2 follow-up probes. Add a 5-minute opening and 5-minute close.\n\nEXAMPLES\nGood interview questions are open, neutral, specific, and inviting. They surface signals that don't show up in HRIS. Example: 'Where do you have single points of failure \u2014 one person holding critical knowledge?' rather than 'Are there succession gaps?'\n\nRULES\nAustralian English. Plain language. No leading questions. No jargon a senior operational leader wouldn't recognise. Maintain neutrality. Flag any question that risks surfacing personal information about specific employees \u2014 those are routed to a 1:1 conversation, not the interview record.\n\nITERATION\nAfter producing the guide, identify three questions that are likely to elicit short answers and propose a probe to deepen each.\n\nASSUMPTIONS\nIf the leaders' areas of focus aren't clear, ask before producing \u2014 the guide may need light tailoring (e.g. capital delivery leaders vs operational leaders).\n\nINTERVIEWEE CONTEXT:\n[Optional. Paste names, roles, focus areas, and any pre-briefing context. If empty, produce a generic guide.]",
      "contextBlock": "INTERVIEWEE CONTEXT:\n[Optional. Paste names, roles, focus areas, and any pre-briefing context. If empty, produce a generic guide.]"
    },
    {
      "id": "03",
      "name": "Stakeholder Engagement \u2014 Mode 2: Interview Synthesis",
      "reference": "Phase 2 \u00b7 Step 2.3 (Mode 2)",
      "category": "Synthesis",
      "purpose": "Synthesises themes from 12-15 leader interview transcripts.",
      "whenToUse": "After running 12-15 interviews. Output goes to the Steering Group.",
      "triggerPhrase": "Synthesise these interviews",
      "structure": "8-Point",
      "prompt": {
        "priming": "This is a synthesis engagement. Output is a thematic report for me to validate with leaders before sharing externally.",
        "context": "I have completed 12-15 structured leader interviews using the Phase 2 interview guide. The transcripts are pasted in the source block. The synthesis goes to the Steering Group, who will use it to validate the demand and supply driver assumptions.",
        "framework": "Approach this as a workforce planning analyst synthesising qualitative interview data into structured findings. Apply Working Principle 01.5. Use the Synthesis format pattern from Section 04 of the IIS.",
        "instructions": "Produce a synthesis report with: (1) Top 5 demand signals named across interviews, with frequency count; (2) Top 5 capability gaps with frequency; (3) Single points of failure flagged (named role, not named person); (4) Tensions or contradictions between leaders; (5) Three verbatim quotes per major theme, attributed by role only (no individual names); (6) Five implications for the workforce model. Aim for 8-12 pages.",
        "examples": "Good synthesis is honest about what the interviews said, including dissent. Frequency counts are concrete (e.g. '7 of 12 leaders named project management capability'). Verbatim quotes are short and pointed. Tensions are surfaced, not glossed.",
        "rules": "Aggregated and de-identified \u2014 no individual names, no identifying details that would point to a specific person. Australian English. Plain language. Flag any claim about specific employees, EA-sensitive material, or identifying detail \u2014 strip from the synthesis and route to me directly.",
        "iteration": "After producing the synthesis, list the three findings most likely to be challenged at the Steering Group and propose a sharper version of each.",
        "assumptions": "If transcripts contain identifying information about individuals, redact at synthesis time. Surface what was redacted and why."
      },
      "fullPromptText": "PRIMING\nThis is a synthesis engagement. Output is a thematic report for me to validate with leaders before sharing externally.\n\nCONTEXT\nI have completed 12-15 structured leader interviews using the Phase 2 interview guide. The transcripts are pasted in the source block. The synthesis goes to the Steering Group, who will use it to validate the demand and supply driver assumptions.\n\nFRAMEWORK\nApproach this as a workforce planning analyst synthesising qualitative interview data into structured findings. Apply Working Principle 01.5. Use the Synthesis format pattern from Section 04 of the IIS.\n\nINSTRUCTIONS\nProduce a synthesis report with: (1) Top 5 demand signals named across interviews, with frequency count; (2) Top 5 capability gaps with frequency; (3) Single points of failure flagged (named role, not named person); (4) Tensions or contradictions between leaders; (5) Three verbatim quotes per major theme, attributed by role only (no individual names); (6) Five implications for the workforce model. Aim for 8-12 pages.\n\nEXAMPLES\nGood synthesis is honest about what the interviews said, including dissent. Frequency counts are concrete (e.g. '7 of 12 leaders named project management capability'). Verbatim quotes are short and pointed. Tensions are surfaced, not glossed.\n\nRULES\nAggregated and de-identified \u2014 no individual names, no identifying details that would point to a specific person. Australian English. Plain language. Flag any claim about specific employees, EA-sensitive material, or identifying detail \u2014 strip from the synthesis and route to me directly.\n\nITERATION\nAfter producing the synthesis, list the three findings most likely to be challenged at the Steering Group and propose a sharper version of each.\n\nASSUMPTIONS\nIf transcripts contain identifying information about individuals, redact at synthesis time. Surface what was redacted and why.\n\nINTERVIEW TRANSCRIPTS:\n[Paste de-identified transcripts. Use 'Leader A \u2014 Engineering' or similar; never paste a person's name.]",
      "contextBlock": "INTERVIEW TRANSCRIPTS:\n[Paste de-identified transcripts. Use 'Leader A \u2014 Engineering' or similar; never paste a person's name.]"
    },
    {
      "id": "04",
      "name": "Capability Framework",
      "reference": "Phase 2 \u00b7 Step 2.2",
      "category": "Drafting",
      "purpose": "Builds the v1 capability taxonomy from position descriptions and existing frameworks.",
      "whenToUse": "After collecting position descriptions and reviewing existing frameworks.",
      "triggerPhrase": "Use my Capability Framework prompt",
      "structure": "8-Point",
      "prompt": {
        "priming": "This is a drafting engagement. Output is a v1 framework draft for me to validate with operational leaders.",
        "context": "Barwon Water has existing capability scaffolding \u2014 a Leadership Framework, a 2030 Growth Framework with 12 organisational competencies, and 10 prioritised Organisational Capabilities. I am building a unified workforce capability taxonomy that combines four dimensions: technical/role skills, leadership, compliance credentials, and Growth Framework competencies \u2014 with a new AI / orchestration overlay added for 2026+.",
        "framework": "Approach this as a senior capability framework specialist building a taxonomy for a regional water utility. Use four proficiency levels: Foundational \u2192 Applied \u2192 Accomplished \u2192 Leading. Reference the existing Barwon Water frameworks listed in the context block.",
        "instructions": "Produce a v1 capability taxonomy with: (1) 30-50 capabilities organised across the four dimensions plus the AI overlay; (2) For each: capability ID, name, one-sentence definition, the 4 proficiency descriptors, segment classification (Strategic / Core / Requisite / Non-core); (3) A role-to-capability mapping table for 10-20 representative roles; (4) Explicit flags on the new AI / orchestration capabilities. Output in a format that copies cleanly into Excel.",
        "examples": "A good capability is durable across role changes, has clear proficiency progression, and has unambiguous segmentation. 'Asset Investment Planning' is a capability; 'Senior Asset Engineer' is a role. The proficiency descriptors describe behaviour at each level \u2014 not just words like 'better' or 'more advanced'.",
        "rules": "Australian English. Don't conflate role and capability. 30-50 capabilities is sufficient \u2014 resist over-engineering. Flag any capability that overlaps significantly with an existing Leadership Framework competency.",
        "iteration": "After producing the framework, identify the three capabilities most likely to be challenged by operational leaders, and propose a refinement to each.",
        "assumptions": "If the position descriptions don't cover a major capability area (e.g. cyber, AI, climate adaptation), surface the gap and propose how to address it before producing the full taxonomy."
      },
      "fullPromptText": "PRIMING\nThis is a drafting engagement. Output is a v1 framework draft for me to validate with operational leaders.\n\nCONTEXT\nBarwon Water has existing capability scaffolding \u2014 a Leadership Framework, a 2030 Growth Framework with 12 organisational competencies, and 10 prioritised Organisational Capabilities. I am building a unified workforce capability taxonomy that combines four dimensions: technical/role skills, leadership, compliance credentials, and Growth Framework competencies \u2014 with a new AI / orchestration overlay added for 2026+.\n\nFRAMEWORK\nApproach this as a senior capability framework specialist building a taxonomy for a regional water utility. Use four proficiency levels: Foundational \u2192 Applied \u2192 Accomplished \u2192 Leading. Reference the existing Barwon Water frameworks listed in the context block.\n\nINSTRUCTIONS\nProduce a v1 capability taxonomy with: (1) 30-50 capabilities organised across the four dimensions plus the AI overlay; (2) For each: capability ID, name, one-sentence definition, the 4 proficiency descriptors, segment classification (Strategic / Core / Requisite / Non-core); (3) A role-to-capability mapping table for 10-20 representative roles; (4) Explicit flags on the new AI / orchestration capabilities. Output in a format that copies cleanly into Excel.\n\nEXAMPLES\nA good capability is durable across role changes, has clear proficiency progression, and has unambiguous segmentation. 'Asset Investment Planning' is a capability; 'Senior Asset Engineer' is a role. The proficiency descriptors describe behaviour at each level \u2014 not just words like 'better' or 'more advanced'.\n\nRULES\nAustralian English. Don't conflate role and capability. 30-50 capabilities is sufficient \u2014 resist over-engineering. Flag any capability that overlaps significantly with an existing Leadership Framework competency.\n\nITERATION\nAfter producing the framework, identify the three capabilities most likely to be challenged by operational leaders, and propose a refinement to each.\n\nASSUMPTIONS\nIf the position descriptions don't cover a major capability area (e.g. cyber, AI, climate adaptation), surface the gap and propose how to address it before producing the full taxonomy.\n\nINPUTS:\n[Paste position descriptions, the Leadership Framework excerpts, the 12 Growth Framework competencies, the 10 Organisational Capabilities list, and any internal capability literature.]",
      "contextBlock": "INPUTS:\n[Paste position descriptions, the Leadership Framework excerpts, the 12 Growth Framework competencies, the 10 Organisational Capabilities list, and any internal capability literature.]"
    },
    {
      "id": "05",
      "name": "Demand Model assumptions",
      "reference": "Phase 3 \u00b7 Step 3.1",
      "category": "Modelling",
      "purpose": "Translates capital plan, asset data, customer projections into demand model assumptions.",
      "whenToUse": "When building the demand sub-models in the Workforce Model.",
      "triggerPhrase": "Run the demand model",
      "structure": "8-Point",
      "prompt": {
        "priming": "This is an analytical engagement producing model inputs. Output is an assumption pack for me to populate the demand sheet of the Workforce Model.",
        "context": "I am building a 5-year (2028-2033) workforce demand forecast for PS28. The model uses four sub-models per capability: asset-driven, project-driven, service-driven, strategic. Productive hours per FTE = ~1,541/year (78% of 1,976 standard hours). Capital pipeline is ~$950M committed over 5 years. Buniya is the digital transformation programme. Capability segments are defined in Section 06 of the IIS.",
        "framework": "Approach this as a workforce demand modelling analyst building defensible assumptions for a regulator-facing submission. Apply Working Principle 01.4 (the model is the system of record). Every assumption must trace to source data or a stated benchmark.",
        "instructions": "For each of 6 capability areas, propose demand model inputs across the four sub-models. For each: (1) numerical value or range; (2) rationale citing source data; (3) confidence level (high / medium / low); (4) sensitivity. For project-driven specifically, show the conversion of $M of capex into FTE demand year-by-year, including phase loading.",
        "examples": "A good demand assumption is traceable: 'Asset-driven demand for Field Operations year 3 = 138 FTE-equiv, derived from 12,400 km of network \u00d7 0.011 maintenance hours/km/year \u00d7 1.04 reactive multiplier \u00f7 1,541 productive hours per FTE. Confidence: medium. Source: Asset Management System Q1 2026 export.'",
        "rules": "Honest about uncertainty \u2014 flag low-confidence assumptions. Don't double-count. Australian English. Apply Constraint 02.4 \u2014 no identifying employee details. Show your working.",
        "iteration": "After producing the assumption pack, identify the three assumptions most sensitive to error and propose a sensitivity range for each.",
        "assumptions": "If the capital plan doesn't include phase profiles, the Buniya plan doesn't translate to FTE, or the asset data lacks intensity factors \u2014 surface the gap and propose the placeholder approach before producing."
      },
      "fullPromptText": "PRIMING\nThis is an analytical engagement producing model inputs. Output is an assumption pack for me to populate the demand sheet of the Workforce Model.\n\nCONTEXT\nI am building a 5-year (2028-2033) workforce demand forecast for PS28. The model uses four sub-models per capability: asset-driven, project-driven, service-driven, strategic. Productive hours per FTE = ~1,541/year (78% of 1,976 standard hours). Capital pipeline is ~$950M committed over 5 years. Buniya is the digital transformation programme. Capability segments are defined in Section 06 of the IIS.\n\nFRAMEWORK\nApproach this as a workforce demand modelling analyst building defensible assumptions for a regulator-facing submission. Apply Working Principle 01.4 (the model is the system of record). Every assumption must trace to source data or a stated benchmark.\n\nINSTRUCTIONS\nFor each of 6 capability areas, propose demand model inputs across the four sub-models. For each: (1) numerical value or range; (2) rationale citing source data; (3) confidence level (high / medium / low); (4) sensitivity. For project-driven specifically, show the conversion of $M of capex into FTE demand year-by-year, including phase loading.\n\nEXAMPLES\nA good demand assumption is traceable: 'Asset-driven demand for Field Operations year 3 = 138 FTE-equiv, derived from 12,400 km of network \u00d7 0.011 maintenance hours/km/year \u00d7 1.04 reactive multiplier \u00f7 1,541 productive hours per FTE. Confidence: medium. Source: Asset Management System Q1 2026 export.'\n\nRULES\nHonest about uncertainty \u2014 flag low-confidence assumptions. Don't double-count. Australian English. Apply Constraint 02.4 \u2014 no identifying employee details. Show your working.\n\nITERATION\nAfter producing the assumption pack, identify the three assumptions most sensitive to error and propose a sensitivity range for each.\n\nASSUMPTIONS\nIf the capital plan doesn't include phase profiles, the Buniya plan doesn't translate to FTE, or the asset data lacks intensity factors \u2014 surface the gap and propose the placeholder approach before producing.\n\nINPUTS:\n[Paste the capital program profile by year and discipline, the Buniya digital plan summary, asset condition data, customer growth projections, current maintenance work order patterns, and any prior demand modelling.]",
      "contextBlock": "INPUTS:\n[Paste the capital program profile by year and discipline, the Buniya digital plan summary, asset condition data, customer growth projections, current maintenance work order patterns, and any prior demand modelling.]"
    },
    {
      "id": "06",
      "name": "Supply Model \u2014 attrition, retirement, mobility",
      "reference": "Phase 3 \u00b7 Step 3.2",
      "category": "Modelling",
      "purpose": "Analyses workforce supply patterns to populate the supply model.",
      "whenToUse": "When building the supply forecast.",
      "triggerPhrase": "Run the supply model",
      "structure": "8-Point",
      "prompt": {
        "priming": "This is an analytical engagement producing supply model inputs. Output is the analysis pack for me to populate the supply sheet.",
        "context": "Building a 5-year supply forecast \u2014 what workforce we will have if no intervention. Equation: Closing supply (year y) = Opening \u2212 Attrition \u2212 Retirements + Recruitment + Mobility (net). Constraints: Geelong labour market is constrained (12,000+ new jobs competing, 18,300 worker shortfall, 7,700 retiring across the region). BAS attrition patterns differ from Barwon Water core.",
        "framework": "Approach this as a workforce supply analyst preparing a supply forecast for a regulator-facing submission. Apply Working Principle 01.4. Every projection must be segmented (capability area minimum), not org-wide averaged.",
        "instructions": "Produce a supply analysis pack with: (1) Attrition rate by capability area; (2) Retirement risk profile; (3) Recruitment capacity; (4) Internal mobility patterns; (5) Recommended supply forecast assumptions. For each: number/range, source data, confidence level, risk flags.",
        "examples": "Good supply analysis is segmented: 'Engineering attrition: 10% pa (3-year history range 8-12%). Confidence: high. Source: HRIS exit data 2023-2025.' Avoid org-wide averages.",
        "rules": "Distinguish voluntary attrition from retirement. Aggregated and de-identified only. Surface contractor footprint separately. Australian English. Flag thin data.",
        "iteration": "After producing the analysis, identify the segment with the highest projection uncertainty and propose a data uplift path.",
        "assumptions": "If the HRIS data isn't segmented to capability area, propose the mapping approach before producing."
      },
      "fullPromptText": "PRIMING\nThis is an analytical engagement producing supply model inputs. Output is the analysis pack for me to populate the supply sheet.\n\nCONTEXT\nBuilding a 5-year supply forecast \u2014 what workforce we will have if no intervention. Equation: Closing supply (year y) = Opening \u2212 Attrition \u2212 Retirements + Recruitment + Mobility (net). Constraints: Geelong labour market is constrained (12,000+ new jobs competing, 18,300 worker shortfall, 7,700 retiring across the region). BAS attrition patterns differ from Barwon Water core.\n\nFRAMEWORK\nApproach this as a workforce supply analyst preparing a supply forecast for a regulator-facing submission. Apply Working Principle 01.4. Every projection must be segmented (capability area minimum), not org-wide averaged.\n\nINSTRUCTIONS\nProduce a supply analysis pack with: (1) Attrition rate by capability area; (2) Retirement risk profile; (3) Recruitment capacity; (4) Internal mobility patterns; (5) Recommended supply forecast assumptions. For each: number/range, source data, confidence level, risk flags.\n\nEXAMPLES\nGood supply analysis is segmented: 'Engineering attrition: 10% pa (3-year history range 8-12%). Confidence: high. Source: HRIS exit data 2023-2025.' Avoid org-wide averages.\n\nRULES\nDistinguish voluntary attrition from retirement. Aggregated and de-identified only. Surface contractor footprint separately. Australian English. Flag thin data.\n\nITERATION\nAfter producing the analysis, identify the segment with the highest projection uncertainty and propose a data uplift path.\n\nASSUMPTIONS\nIf the HRIS data isn't segmented to capability area, propose the mapping approach before producing.\n\nINPUTS:\n[Paste de-identified HRIS attrition data, age demographic profile by capability area, recruitment time-to-fill data by role family, internal mobility records, contractor spend by category.]",
      "contextBlock": "INPUTS:\n[Paste de-identified HRIS attrition data, age demographic profile by capability area, recruitment time-to-fill data by role family, internal mobility records, contractor spend by category.]"
    },
    {
      "id": "07",
      "name": "AI Productivity Scenarios",
      "reference": "Phase 3 \u00b7 Step 3.3",
      "category": "Modelling",
      "purpose": "Defines three AI productivity scenarios layered on demand.",
      "whenToUse": "When layering AI productivity assumptions onto demand.",
      "triggerPhrase": "Run the AI scenarios",
      "structure": "8-Point",
      "prompt": {
        "priming": "This is an analytical engagement producing scenario assumptions. Output is a scenario pack for ELT review and model population.",
        "context": "The Digital GM has framed an AI-enabled future where back-office work reduces and engineering / field / strategic capabilities are protected. I am translating this into three productivity scenarios: Low (5-10%), Medium (15-25%), High (30%+). Apply differentially by capability segment: Strategic (5-20%), Core (8-28%), Requisite (15-50%). S-curve adoption.",
        "framework": "Approach this as an AI workforce impact analyst preparing scenarios for executive choice. Apply Working Principle 01.6. Anchor in industry benchmarks. The High scenario should be a defensible stretch, not the central case.",
        "instructions": "For each of the three scenarios: (1) Productivity uplift % by capability segment, by year (S-curve); (2) Rationale grounded in named industry benchmarks; (3) Underlying assumptions; (4) Three risks; (5) Three leading indicators.",
        "examples": "Good scenarios are differentiated: 'Medium scenario, requisite capabilities, year 5 productivity uplift = 22%. S-curve: 4% / 8% / 14% / 19% / 22%. Source: Productivity Commission 2024 report on automation in service-sector roles.'",
        "rules": "Honest about uncertainty. Be specific about what 'productivity' means. Cite sources. Frame as decision support, not predictions. ESC will challenge optimistic assumptions.",
        "iteration": "Identify the three assumptions in each scenario most likely to be challenged and propose a sharper supporting argument for each.",
        "assumptions": "If the Buniya plan or digital strategy is unclear, surface the gap before producing."
      },
      "fullPromptText": "PRIMING\nThis is an analytical engagement producing scenario assumptions. Output is a scenario pack for ELT review and model population.\n\nCONTEXT\nThe Digital GM has framed an AI-enabled future where back-office work reduces and engineering / field / strategic capabilities are protected. I am translating this into three productivity scenarios: Low (5-10%), Medium (15-25%), High (30%+). Apply differentially by capability segment: Strategic (5-20%), Core (8-28%), Requisite (15-50%). S-curve adoption.\n\nFRAMEWORK\nApproach this as an AI workforce impact analyst preparing scenarios for executive choice. Apply Working Principle 01.6. Anchor in industry benchmarks. The High scenario should be a defensible stretch, not the central case.\n\nINSTRUCTIONS\nFor each of the three scenarios: (1) Productivity uplift % by capability segment, by year (S-curve); (2) Rationale grounded in named industry benchmarks; (3) Underlying assumptions; (4) Three risks; (5) Three leading indicators.\n\nEXAMPLES\nGood scenarios are differentiated: 'Medium scenario, requisite capabilities, year 5 productivity uplift = 22%. S-curve: 4% / 8% / 14% / 19% / 22%. Source: Productivity Commission 2024 report on automation in service-sector roles.'\n\nRULES\nHonest about uncertainty. Be specific about what 'productivity' means. Cite sources. Frame as decision support, not predictions. ESC will challenge optimistic assumptions.\n\nITERATION\nIdentify the three assumptions in each scenario most likely to be challenged and propose a sharper supporting argument for each.\n\nASSUMPTIONS\nIf the Buniya plan or digital strategy is unclear, surface the gap before producing.\n\nINPUTS:\n[Paste the Buniya plan summary, the Digital GM's future-of-work framing, AI productivity research, the capability segmentation.]",
      "contextBlock": "INPUTS:\n[Paste the Buniya plan summary, the Digital GM's future-of-work framing, AI productivity research, the capability segmentation.]"
    },
    {
      "id": "08",
      "name": "Gap & Risk Analysis",
      "reference": "Phase 3 \u00b7 Steps 3.4\u20133.5",
      "category": "Analysis",
      "purpose": "Interprets gap analysis and rates capability risk across 11 dimensions.",
      "whenToUse": "After demand and supply models are populated.",
      "triggerPhrase": "Run the gap analysis",
      "structure": "8-Point",
      "prompt": {
        "priming": "This is an analytical engagement producing interpretation and risk ratings. Output is the gap interpretation and risk register for Steering Group review.",
        "context": "The Workforce Model has produced a gap analysis: AI-adjusted demand minus supply by capability and year. I need to interpret what the gaps mean and rate the capability risk using the 11-dimension framework.",
        "framework": "Approach this as a workforce gap and risk analyst preparing material for Steering Group review. Apply Working Principle 01.4. Distinguish quantity gaps, quality gaps, and timing gaps.",
        "instructions": "Produce: (1) Per-capability one-paragraph gap interpretation; (2) Risk register entry per capability rated Low/Medium/High/Critical across all 11 dimensions; (3) Top 3-5 critical capabilities; (4) Quick wins; (5) Capabilities with surplus.",
        "examples": "Good interpretation is concrete: 'Engineering & Asset Planning gap reaches +14 FTE by 2031, driven by capital program peak in 2030. Bites year 4. Consequence: capital program delays; Buniya integration at risk.'",
        "rules": "Distinguish quantity, quality, and timing gaps. Avoid uniform ratings. Australian English. No identifying detail.",
        "iteration": "Identify the three risk ratings most likely to be challenged at Steering Group and propose a sharper rationale for each.",
        "assumptions": "If the gap analysis output is missing any capability area, or the model assumptions are unclear, surface the gap before producing."
      },
      "fullPromptText": "PRIMING\nThis is an analytical engagement producing interpretation and risk ratings. Output is the gap interpretation and risk register for Steering Group review.\n\nCONTEXT\nThe Workforce Model has produced a gap analysis: AI-adjusted demand minus supply by capability and year. I need to interpret what the gaps mean and rate the capability risk using the 11-dimension framework.\n\nFRAMEWORK\nApproach this as a workforce gap and risk analyst preparing material for Steering Group review. Apply Working Principle 01.4. Distinguish quantity gaps, quality gaps, and timing gaps.\n\nINSTRUCTIONS\nProduce: (1) Per-capability one-paragraph gap interpretation; (2) Risk register entry per capability rated Low/Medium/High/Critical across all 11 dimensions; (3) Top 3-5 critical capabilities; (4) Quick wins; (5) Capabilities with surplus.\n\nEXAMPLES\nGood interpretation is concrete: 'Engineering & Asset Planning gap reaches +14 FTE by 2031, driven by capital program peak in 2030. Bites year 4. Consequence: capital program delays; Buniya integration at risk.'\n\nRULES\nDistinguish quantity, quality, and timing gaps. Avoid uniform ratings. Australian English. No identifying detail.\n\nITERATION\nIdentify the three risk ratings most likely to be challenged at Steering Group and propose a sharper rationale for each.\n\nASSUMPTIONS\nIf the gap analysis output is missing any capability area, or the model assumptions are unclear, surface the gap before producing.\n\nINPUTS:\n[Paste the gap analysis output from the Workforce Model, the capability segmentation, market intelligence, the Phase 2 interview synthesis themes.]",
      "contextBlock": "INPUTS:\n[Paste the gap analysis output from the Workforce Model, the capability segmentation, market intelligence, the Phase 2 interview synthesis themes.]"
    },
    {
      "id": "09",
      "name": "Build / Buy / Borrow / Automate Intervention Design",
      "reference": "Phase 4 \u00b7 Step 4.1",
      "category": "Planning",
      "purpose": "Designs intervention mix for each capability gap.",
      "whenToUse": "After gap and risk analysis is complete.",
      "triggerPhrase": "Design the interventions",
      "structure": "8-Point",
      "prompt": {
        "priming": "This is a design engagement producing intervention scenarios. Output is the scenario pack for Steering Group costing and review.",
        "context": "For each capability gap from Phase 3, I am designing intervention scenarios using four levers: Build (6-24mo lead time), Buy (3-9mo), Borrow (2-8 weeks), Automate (months-years). Capability segment implications: Strategic biases to Build/Buy; Core balanced; Requisite biases to Buy/Automate.",
        "framework": "Approach this as a workforce intervention designer building defensible scenarios for executive decision. Apply Working Principles 01.4 and 01.6. Most gaps need a mix.",
        "instructions": "For each capability gap: (1) Recommend an intervention mix \u2014 % by lever, year by year; (2) Cost each intervention; (3) Sequence with lead times; (4) Identify dependencies; (5) Flag risks; (6) Check inclusion impact.",
        "examples": "Good mix: 'Engineering year-5 gap of 14 FTE: Build 6 FTE (start year 1, 18-month ramp); Buy 5 FTE (years 1-3); Borrow 3 FTE-eq (years 2-4 peak); Automate \u22122 FTE demand (Buniya year 3+). Total cost ~$3.2M over 5 years.'",
        "rules": "Don't recommend single-lever solutions for gaps over 5 FTE. Don't underestimate Build lead times. Apply Constraint 02.5 (Fair Work consultation triggers).",
        "iteration": "Propose a 'low intervention' alternative for each gap that uses 30% less spend but accepts more risk.",
        "assumptions": "If the cost rates or contractor benchmarks aren't current, surface this before producing."
      },
      "fullPromptText": "PRIMING\nThis is a design engagement producing intervention scenarios. Output is the scenario pack for Steering Group costing and review.\n\nCONTEXT\nFor each capability gap from Phase 3, I am designing intervention scenarios using four levers: Build (6-24mo lead time), Buy (3-9mo), Borrow (2-8 weeks), Automate (months-years). Capability segment implications: Strategic biases to Build/Buy; Core balanced; Requisite biases to Buy/Automate.\n\nFRAMEWORK\nApproach this as a workforce intervention designer building defensible scenarios for executive decision. Apply Working Principles 01.4 and 01.6. Most gaps need a mix.\n\nINSTRUCTIONS\nFor each capability gap: (1) Recommend an intervention mix \u2014 % by lever, year by year; (2) Cost each intervention; (3) Sequence with lead times; (4) Identify dependencies; (5) Flag risks; (6) Check inclusion impact.\n\nEXAMPLES\nGood mix: 'Engineering year-5 gap of 14 FTE: Build 6 FTE (start year 1, 18-month ramp); Buy 5 FTE (years 1-3); Borrow 3 FTE-eq (years 2-4 peak); Automate \u22122 FTE demand (Buniya year 3+). Total cost ~$3.2M over 5 years.'\n\nRULES\nDon't recommend single-lever solutions for gaps over 5 FTE. Don't underestimate Build lead times. Apply Constraint 02.5 (Fair Work consultation triggers).\n\nITERATION\nPropose a 'low intervention' alternative for each gap that uses 30% less spend but accepts more risk.\n\nASSUMPTIONS\nIf the cost rates or contractor benchmarks aren't current, surface this before producing.\n\nINPUTS:\n[Paste the capability gap analysis, segmentation, current footprint, market data, EA cost rates, contractor bill rates, recruitment costs.]",
      "contextBlock": "INPUTS:\n[Paste the capability gap analysis, segmentation, current footprint, market data, EA cost rates, contractor bill rates, recruitment costs.]"
    },
    {
      "id": "10",
      "name": "Strategic Options Paper for ELT/Board",
      "reference": "Phase 4 \u00b7 Step 4.2",
      "category": "Drafting",
      "purpose": "Drafts the workforce-shape options paper for Board decision.",
      "whenToUse": "The most consequential output of the project.",
      "triggerPhrase": "Draft an options paper",
      "structure": "8-Point",
      "prompt": {
        "priming": "This is a drafting engagement for a Board-level decision paper. Output is a draft for me to iterate with the coalition GMs before submission.",
        "context": "The Board must decide what workforce shape Barwon Water adopts over 2028-2033 in an AI-enabled environment. The Digital GM has framed one pathway ('Protect & Transform'). My job is to present 2-3 credible options for Board choice.",
        "framework": "Approach this as a strategic advisor drafting an options paper for ELT and Board. Apply Working Principle 01.7 (real options, not straw men). Stress-test each option equally.",
        "instructions": "Draft 8-12 pages: (1) Executive summary (1pg); (2) Strategic context (2pg); (3) Each option in detail (2pg each): description, workforce shape, cost, capability outcome, inclusion impact, risk, change burden, conditions; (4) Comparison table (1pg); (5) Recommendation with rationale (1pg); (6) Decision required.",
        "examples": "A good options paper presents real choices, with cases for and against clearly drawn. Recommendation is decisive \u2014 one or two sentences with no hedging. The strongest option is supported with evidence, not adjectives.",
        "rules": "Australian English. Analytical, balanced, decisive. Stress-test each option equally. Apply Constraints 02.5 and 02.6. Honest about uncertainties.",
        "iteration": "Identify the three sections most likely to attract Board challenge and propose a sharper version of each.",
        "assumptions": "If you cannot articulate a clear recommendation, the options aren't well understood. Surface this before producing."
      },
      "fullPromptText": "PRIMING\nThis is a drafting engagement for a Board-level decision paper. Output is a draft for me to iterate with the coalition GMs before submission.\n\nCONTEXT\nThe Board must decide what workforce shape Barwon Water adopts over 2028-2033 in an AI-enabled environment. The Digital GM has framed one pathway ('Protect & Transform'). My job is to present 2-3 credible options for Board choice.\n\nFRAMEWORK\nApproach this as a strategic advisor drafting an options paper for ELT and Board. Apply Working Principle 01.7 (real options, not straw men). Stress-test each option equally.\n\nINSTRUCTIONS\nDraft 8-12 pages: (1) Executive summary (1pg); (2) Strategic context (2pg); (3) Each option in detail (2pg each): description, workforce shape, cost, capability outcome, inclusion impact, risk, change burden, conditions; (4) Comparison table (1pg); (5) Recommendation with rationale (1pg); (6) Decision required.\n\nEXAMPLES\nA good options paper presents real choices, with cases for and against clearly drawn. Recommendation is decisive \u2014 one or two sentences with no hedging. The strongest option is supported with evidence, not adjectives.\n\nRULES\nAustralian English. Analytical, balanced, decisive. Stress-test each option equally. Apply Constraints 02.5 and 02.6. Honest about uncertainties.\n\nITERATION\nIdentify the three sections most likely to attract Board challenge and propose a sharper version of each.\n\nASSUMPTIONS\nIf you cannot articulate a clear recommendation, the options aren't well understood. Surface this before producing.\n\nINPUTS:\n[Paste intervention scenario pack, gap analysis, capability segmentation, financial implications by option, inclusion impact modelling, Phase 2 interview synthesis themes, the Digital GM's framing email.]",
      "contextBlock": "INPUTS:\n[Paste intervention scenario pack, gap analysis, capability segmentation, financial implications by option, inclusion impact modelling, Phase 2 interview synthesis themes, the Digital GM's framing email.]"
    },
    {
      "id": "11",
      "name": "Inclusion Impact Assessment",
      "reference": "Phase 4 \u00b7 Step 4.3",
      "category": "Analysis",
      "purpose": "Models representation impact of strategic options.",
      "whenToUse": "Mandatory for any workforce reshape strategy.",
      "triggerPhrase": "Run the inclusion impact",
      "structure": "8-Point",
      "prompt": {
        "priming": "This is an analytical engagement producing impact assessment. Output is the assessment pack supporting the Strategic Options Paper.",
        "context": "Barwon Water has 2028 representation targets across 5 dimensions: Aboriginal & TSI 4.0%, All Abilities 17.5%, Neurodivergent 15%, CALD 25%, Gender 43% women / 57% men.",
        "framework": "Approach this as an inclusion and diversity impact analyst. Apply Working Principle 01.6 \u2014 model honestly. Use aggregated, de-identified data only.",
        "instructions": "For each strategic option: (1) Per-dimension impact: starting %, projected by 2028 and 2033, gap to target; (2) Mechanisms driving the impact; (3) Mitigations where worsened; (4) Permanent risks flagged; (5) Compare options.",
        "examples": "Good impact analysis is honest: 'Option A worsens gender representation by 2031: women drop from 41% to 36% as back-office reduces. Mitigations: targeted hiring in growth areas.'",
        "rules": "Aggregated and de-identified only. Apply Constraint 02.4. Honest about trade-offs.",
        "iteration": "Identify which option's inclusion outcome is most fragile and propose a strengthened mitigation for it.",
        "assumptions": "If the demographic data isn't segmented to capability or function level, propose the segmentation approach before producing."
      },
      "fullPromptText": "PRIMING\nThis is an analytical engagement producing impact assessment. Output is the assessment pack supporting the Strategic Options Paper.\n\nCONTEXT\nBarwon Water has 2028 representation targets across 5 dimensions: Aboriginal & TSI 4.0%, All Abilities 17.5%, Neurodivergent 15%, CALD 25%, Gender 43% women / 57% men.\n\nFRAMEWORK\nApproach this as an inclusion and diversity impact analyst. Apply Working Principle 01.6 \u2014 model honestly. Use aggregated, de-identified data only.\n\nINSTRUCTIONS\nFor each strategic option: (1) Per-dimension impact: starting %, projected by 2028 and 2033, gap to target; (2) Mechanisms driving the impact; (3) Mitigations where worsened; (4) Permanent risks flagged; (5) Compare options.\n\nEXAMPLES\nGood impact analysis is honest: 'Option A worsens gender representation by 2031: women drop from 41% to 36% as back-office reduces. Mitigations: targeted hiring in growth areas.'\n\nRULES\nAggregated and de-identified only. Apply Constraint 02.4. Honest about trade-offs.\n\nITERATION\nIdentify which option's inclusion outcome is most fragile and propose a strengthened mitigation for it.\n\nASSUMPTIONS\nIf the demographic data isn't segmented to capability or function level, propose the segmentation approach before producing.\n\nINPUTS:\n[Paste current demographics by capability, attrition patterns segmented by demographic, hiring data, the strategic options being modelled, the People Strategy 2028 targets.]",
      "contextBlock": "INPUTS:\n[Paste current demographics by capability, attrition patterns segmented by demographic, hiring data, the strategic options being modelled, the People Strategy 2028 targets.]"
    },
    {
      "id": "12",
      "name": "PS28 Submission Synthesis",
      "reference": "Phase 5 \u00b7 Step 5.1",
      "category": "Drafting",
      "purpose": "Drafts the regulatory submission narrative.",
      "whenToUse": "When drafting the PS28 narrative for ESC.",
      "triggerPhrase": "Draft the PS28 narrative",
      "structure": "8-Point",
      "prompt": {
        "priming": "This is a drafting engagement for a regulator-facing document. Output is a draft to be reviewed by Pricing/Strategy before lodgement.",
        "context": "The Essential Services Commission applies the PREMO framework to Victorian water price submissions. Workforce costs are a major component. The narrative accompanies the workforce financial inputs handed to Link 8 (Revenue Model).",
        "framework": "Approach this as a regulatory submission writer with deep familiarity with PREMO. Apply Working Principle 01.4 and 01.2.",
        "instructions": "Draft 15-25 pages: (1) Workforce overview; (2) Strategic context; (3) Demand drivers; (4) Supply context; (5) Capability investment cases; (6) Productivity assumptions; (7) Workforce shape strategy; (8) Financial trajectory; (9) Sensitivity and risk; (10) Customer outcomes link.",
        "examples": "Good regulatory narrative is plain, evidence-led, traceable. 'FTE increases from 446 in 2026 to 487 by 2033 driven by [specific driver]. Supported by [specific evidence], traceable to [specific model output].'",
        "rules": "Plain language. Every assertion ties to evidence. Honest about assumptions. Australian English. Pricing/Strategy peer-reviews before lodgement.",
        "iteration": "Identify the three sections most likely to attract ESC challenge and propose a sharper version of each.",
        "assumptions": "If the approved strategic pathway, model outputs, or sensitivity analysis is missing or unclear, surface the gap before producing."
      },
      "fullPromptText": "PRIMING\nThis is a drafting engagement for a regulator-facing document. Output is a draft to be reviewed by Pricing/Strategy before lodgement.\n\nCONTEXT\nThe Essential Services Commission applies the PREMO framework to Victorian water price submissions. Workforce costs are a major component. The narrative accompanies the workforce financial inputs handed to Link 8 (Revenue Model).\n\nFRAMEWORK\nApproach this as a regulatory submission writer with deep familiarity with PREMO. Apply Working Principle 01.4 and 01.2.\n\nINSTRUCTIONS\nDraft 15-25 pages: (1) Workforce overview; (2) Strategic context; (3) Demand drivers; (4) Supply context; (5) Capability investment cases; (6) Productivity assumptions; (7) Workforce shape strategy; (8) Financial trajectory; (9) Sensitivity and risk; (10) Customer outcomes link.\n\nEXAMPLES\nGood regulatory narrative is plain, evidence-led, traceable. 'FTE increases from 446 in 2026 to 487 by 2033 driven by [specific driver]. Supported by [specific evidence], traceable to [specific model output].'\n\nRULES\nPlain language. Every assertion ties to evidence. Honest about assumptions. Australian English. Pricing/Strategy peer-reviews before lodgement.\n\nITERATION\nIdentify the three sections most likely to attract ESC challenge and propose a sharper version of each.\n\nASSUMPTIONS\nIf the approved strategic pathway, model outputs, or sensitivity analysis is missing or unclear, surface the gap before producing.\n\nINPUTS:\n[Paste the Workforce Model PS28 Outputs sheet, the approved strategic pathway from Phase 4, Strategy 2030 priorities, the customer outcomes framework, EA outcomes, the sensitivity analysis.]",
      "contextBlock": "INPUTS:\n[Paste the Workforce Model PS28 Outputs sheet, the approved strategic pathway from Phase 4, Strategy 2030 priorities, the customer outcomes framework, EA outcomes, the sensitivity analysis.]"
    }
  ]
}