{
  "meta": {
    "scenario": "Anand's headcount-allocation dashboard: Copilot first output through the four responsible-AI moves",
    "chapter": "Chapter 14 - Responsible AI in Analytics",
    "book_figure": "Figure 14.4",
    "book_figure_svg": "assets/figures/14-anand-capstone-flow.svg",
    "domain": "Financial-services analytics governance capstone",
    "synthetic": true,
    "narrative_purpose": {
      "first_output": "Copilot returns a plausible-looking one-page FTE-by-function dashboard from an eight-line brief in about fifteen seconds. Aggregate attrition sits at 8% with a green treatment; the top-line looks calm.",
      "four_moves": "Anand runs Provenance, Fairness, Decision-Boundary, and Disclosure in order. Two substantive issues surface (fairness callout revealing 22% attrition in the under-30 cohort; role-escalation clause naming the archival risk); one polish item; disclosure amended to name the fairness callout as a deliberate post-generation addition."
    }
  },
  "cfo_ask": "One-page headcount-allocation dashboard for Q4 planning; ten net-new heads to allocate across functions.",
  "copilot_first_output": {
    "prompt_lines": 8,
    "time_to_return_seconds": 15,
    "shape": "one-page FTE by function with attrition and open-req overlays",
    "aggregate_attrition_pct": 8,
    "aggregate_attrition_treatment": "green (interpreted as healthy)",
    "sort_order": "descending by absolute headcount",
    "smart_narrative_present": true,
    "provenance_footer_present": false
  },
  "four_moves": [
    {
      "id": 1,
      "move": "Provenance",
      "purpose": "Audit trail",
      "actions": [
        "Assemble the defensibility bundle: eight-line prompt, Scope restatement, decisions-made and decisions-deferred lists, source-table names, run timestamp",
        "Attach as a provenance footer to the dashboard file so a subpoena-style ask can be answered from the file itself"
      ],
      "surfaced_issue": null
    },
    {
      "id": 2,
      "move": "Fairness",
      "purpose": "Whose story gets told",
      "actions": [
        "Decompose the 8% aggregate attrition by demographic cohort (age band, tenure band, department)",
        "Under-30 cohort attrition is 22% (2.75x aggregate); over-50 cohort attrition is 3%",
        "Add a fairness callout naming the under-30 concentration and the risk it poses for the ten-new-head allocation (net add masks a net loss in a specific cohort)"
      ],
      "surfaced_issue": "Under-30 attrition at 22% would ship invisible without the decomposition; would misinform the ten-head allocation"
    },
    {
      "id": 3,
      "move": "Decision-Boundary",
      "purpose": "Match verification to role",
      "actions": [
        "Classify the artifact: recommendation (CFO makes the call), not decision-artifact grade",
        "Apply the recommendation-tier rubric: Part I visual rubric + evidence-on-every-claim + provenance footer",
        "Add an escalation clause: if the CFO's decision is archived and referenced as 'the Q4 headcount rationale' downstream, the archived version must be re-verified at decision-artifact tier"
      ],
      "surfaced_issue": "Downstream archival risk unnamed until the escalation clause was added"
    },
    {
      "id": 4,
      "move": "Disclosure",
      "purpose": "Name what changed after Copilot",
      "actions": [
        "Standard AI-directed marking present in the footer",
        "Add one additional line specific to this artifact: 'Fairness callout added post-generation; see notes for the under-30 cohort attrition context that the aggregate number understates.'"
      ],
      "surfaced_issue": null
    }
  ],
  "ship_verdict": "ship",
  "substantive_issues_surfaced": 2,
  "polish_items_surfaced": 1,
  "bundle_size_items": 6,
  "would_have_shipped_without_pass": {
    "fairness_callout": false,
    "escalation_clause": false
  },
  "source_sha256": "875527fe7bd24da1221bd7aef73a62e2932b86b6e42b645c6c5de1af1158cd45"
}