{
  "agents": {
    "Brackett": {
      "average_score": 0.7837153846153846,
      "completed_runs": 107,
      "attempted_runs": 107,
      "completion_rate": 1.0,
      "tested_use_cases": 39,
      "shared_score": 0.8396347222222222
    },
    "Claude": {
      "average_score": 0.156305,
      "completed_runs": 12,
      "attempted_runs": 60,
      "completion_rate": 0.2,
      "tested_use_cases": 30,
      "shared_score": 0.15746458333333332
    },
    "Codex": {
      "average_score": 0.581225,
      "completed_runs": 24,
      "attempted_runs": 24,
      "completion_rate": 1.0,
      "tested_use_cases": 24,
      "shared_score": 0.581225
    }
  },
  "case_scores": {
    "Brackett": {
      "Candidate–Role Rematching": 0.999,
      "Credit Release Queue": 1.0,
      "Customer Returns Inbox Triage": 0.8216,
      "Discount Request Review": 0.8352333333333334,
      "Job Offer Follow-Up": 0.6316666666666666,
      "Job Requisition Broadcast": 0.825,
      "Job Requisition Triage": 0.6466666666666667,
      "Limited-Stock Order Allocation": 0.854,
      "Low-Stock Replenishment Sweep": 0.99,
      "Order Request Replenishment": 0.82,
      "Recruiter Performance Scorecard": 0.86,
      "Accounts Receivable Collections Follow-Up 1": 0.7916666666666667,
      "Candidate–Role Rematching 1": 0.8573333333333334,
      "Credit Release Queue 1": 0.9575,
      "Customer Support Issue Triage 1": 0.7879,
      "Discount Request Review 1": 0.9566666666666667,
      "Email Purchase Invoice Posting 1": 0.76,
      "Freight Consolidation Planning 1": 0.6756666666666666,
      "Goods Receipt Shortfall Follow-Up 1": 0.8683333333333333,
      "Inbound Sales Enquiry Routing 1": 0.8793333333333333,
      "Job Requisition Broadcast 1": 0.9533333333333333,
      "Job Requisition Triage 1": 0.835,
      "Leave Request Eligibility Review 1": 0.775,
      "Limited-Stock Order Allocation 1": 0.755,
      "Low-Stock Replenishment Sweep 1": 0.7466666666666667,
      "New-Hire Onboarding Coordination 1": 0.88,
      "Order Request Replenishment 1": 0.97,
      "Purchase Invoice Reconciliation 1": 0.9666666666666667,
      "Recruiter Performance Scorecard 1": 0.8833333333333334,
      "Vehicle Maintenance Service Booking 1": 0.7213333333333333,
      "Return Eligibility and RMA": 0.4415,
      "RFQ Quote Award": 0.7925,
      "Shortage Substitution Queue": 0.646,
      "Return Eligibility and RMA 1": 0.56,
      "RFQ Quote Award 1": 0.765,
      "Shortage Substitution Queue 1": 0.4475,
      "Supplier Price Increase Review 1": 0.42600000000000005,
      "Warranty Claim Intake 1": 0.7075,
      "Weather-Aware Departure Review 1": 0.475
    },
    "Claude": {
      "Candidate–Role Rematching": 0.0,
      "Credit Release Queue": 0.43335,
      "Customer Returns Inbox Triage": 0.0,
      "Discount Request Review": 0.0,
      "Job Offer Follow-Up": 0.3027,
      "Job Requisition Broadcast": 0.0,
      "Job Requisition Triage": 0.67245,
      "Limited-Stock Order Allocation": 0.0,
      "Low-Stock Replenishment Sweep": 0.0,
      "Order Request Replenishment": 0.0,
      "Recruiter Performance Scorecard": 0.0,
      "Accounts Receivable Collections Follow-Up 1": 0.0,
      "Candidate–Role Rematching 1": 0.8975,
      "Credit Release Queue 1": 0.729,
      "Customer Support Issue Triage 1": 0.0,
      "Discount Request Review 1": 0.0,
      "Email Purchase Invoice Posting 1": 0.0,
      "Freight Consolidation Planning 1": 0.0,
      "Goods Receipt Shortfall Follow-Up 1": 0.0,
      "Inbound Sales Enquiry Routing 1": 0.0,
      "Job Requisition Broadcast 1": 0.0,
      "Job Requisition Triage 1": 0.45,
      "Leave Request Eligibility Review 1": 0.0,
      "Limited-Stock Order Allocation 1": 0.0,
      "Low-Stock Replenishment Sweep 1": 0.32415,
      "New-Hire Onboarding Coordination 1": 0.0,
      "Order Request Replenishment 1": 0.46,
      "Purchase Invoice Reconciliation 1": 0.0,
      "Recruiter Performance Scorecard 1": 0.42,
      "Vehicle Maintenance Service Booking 1": 0.0
    },
    "Codex": {
      "Candidate–Role Rematching": 0.5,
      "Credit Release Queue": 0.36,
      "Customer Returns Inbox Triage": 0.6364,
      "Discount Request Review": 0.755,
      "Job Offer Follow-Up": 0.45299999999999996,
      "Job Requisition Broadcast": 0.625,
      "Job Requisition Triage": 0.365,
      "Limited-Stock Order Allocation": 0.257,
      "Low-Stock Replenishment Sweep": 0.855,
      "Order Request Replenishment": 0.915,
      "Recruiter Performance Scorecard": 0.25,
      "Accounts Receivable Collections Follow-Up 1": 0.853,
      "Candidate–Role Rematching 1": 0.33,
      "Credit Release Queue 1": 0.737,
      "Customer Support Issue Triage 1": 0.645,
      "Discount Request Review 1": 0.488,
      "Freight Consolidation Planning 1": 0.345,
      "Goods Receipt Shortfall Follow-Up 1": 0.695,
      "Leave Request Eligibility Review 1": 0.9,
      "Low-Stock Replenishment Sweep 1": 0.29,
      "New-Hire Onboarding Coordination 1": 0.435,
      "Purchase Invoice Reconciliation 1": 0.895,
      "Recruiter Performance Scorecard 1": 0.565,
      "Vehicle Maintenance Service Booking 1": 0.8
    }
  },
  "shared_cases": [
    "Candidate–Role Rematching",
    "Credit Release Queue",
    "Customer Returns Inbox Triage",
    "Discount Request Review",
    "Job Offer Follow-Up",
    "Job Requisition Broadcast",
    "Job Requisition Triage",
    "Limited-Stock Order Allocation",
    "Low-Stock Replenishment Sweep",
    "Order Request Replenishment",
    "Recruiter Performance Scorecard",
    "Accounts Receivable Collections Follow-Up 1",
    "Candidate–Role Rematching 1",
    "Credit Release Queue 1",
    "Customer Support Issue Triage 1",
    "Discount Request Review 1",
    "Freight Consolidation Planning 1",
    "Goods Receipt Shortfall Follow-Up 1",
    "Leave Request Eligibility Review 1",
    "Low-Stock Replenishment Sweep 1",
    "New-Hire Onboarding Coordination 1",
    "Purchase Invoice Reconciliation 1",
    "Recruiter Performance Scorecard 1",
    "Vehicle Maintenance Service Booking 1"
  ],
  "tested_cases": [
    "Candidate–Role Rematching",
    "Credit Release Queue",
    "Customer Returns Inbox Triage",
    "Discount Request Review",
    "Job Offer Follow-Up",
    "Job Requisition Broadcast",
    "Job Requisition Triage",
    "Limited-Stock Order Allocation",
    "Low-Stock Replenishment Sweep",
    "Order Request Replenishment",
    "Recruiter Performance Scorecard",
    "Accounts Receivable Collections Follow-Up 1",
    "Candidate–Role Rematching 1",
    "Credit Release Queue 1",
    "Customer Support Issue Triage 1",
    "Discount Request Review 1",
    "Email Purchase Invoice Posting 1",
    "Freight Consolidation Planning 1",
    "Goods Receipt Shortfall Follow-Up 1",
    "Inbound Sales Enquiry Routing 1",
    "Job Requisition Broadcast 1",
    "Job Requisition Triage 1",
    "Leave Request Eligibility Review 1",
    "Limited-Stock Order Allocation 1",
    "Low-Stock Replenishment Sweep 1",
    "New-Hire Onboarding Coordination 1",
    "Order Request Replenishment 1",
    "Purchase Invoice Reconciliation 1",
    "Recruiter Performance Scorecard 1",
    "Vehicle Maintenance Service Booking 1",
    "Return Eligibility and RMA",
    "RFQ Quote Award",
    "Shortage Substitution Queue",
    "Return Eligibility and RMA 1",
    "RFQ Quote Award 1",
    "Shortage Substitution Queue 1",
    "Supplier Price Increase Review 1",
    "Warranty Claim Intake 1",
    "Weather-Aware Departure Review 1"
  ],
  "untested_cases": [
    "Purchase Invoice Reconciliation",
    "Returns Inbox Sweep",
    "Vehicle Maintenance Service Booking"
  ],
  "shared_agents": {
    "Brackett": {
      "average_score": 0.8396347222222222,
      "completed_runs": 71,
      "attempted_runs": 71,
      "completion_rate": 1.0,
      "tested_use_cases": 24,
      "shared_score": 0.8396347222222222
    },
    "Claude": {
      "average_score": 0.15746458333333332,
      "completed_runs": 10,
      "attempted_runs": 48,
      "completion_rate": 0.20833333333333334,
      "tested_use_cases": 24,
      "shared_score": 0.15746458333333332
    },
    "Codex": {
      "average_score": 0.581225,
      "completed_runs": 24,
      "attempted_runs": 24,
      "completion_rate": 1.0,
      "tested_use_cases": 24,
      "shared_score": 0.581225
    }
  },
  "runs": [
    {
      "case": "Candidate–Role Rematching",
      "agent": "Codex",
      "run": 1,
      "score": 0.5,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Candidate–Role Rematching",
      "agent": "Brackett",
      "run": 1,
      "score": 1.0,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Candidate–Role Rematching",
      "agent": "Brackett",
      "run": 2,
      "score": 1.0,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Candidate–Role Rematching",
      "agent": "Brackett",
      "run": 3,
      "score": 0.997,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Candidate–Role Rematching",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Candidate–Role Rematching",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Credit Release Queue",
      "agent": "Codex",
      "run": 1,
      "score": 0.36,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Credit Release Queue",
      "agent": "Brackett",
      "run": 1,
      "score": 1.0,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Credit Release Queue",
      "agent": "Brackett",
      "run": 2,
      "score": 1.0,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Credit Release Queue",
      "agent": "Brackett",
      "run": 3,
      "score": 1.0,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Credit Release Queue",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Credit Release Queue",
      "agent": "Claude",
      "run": 2,
      "score": 0.8667,
      "completed": true,
      "model": "Sonnet 5"
    },
    {
      "case": "Customer Returns Inbox Triage",
      "agent": "Codex",
      "run": 1,
      "score": 0.6364,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Customer Returns Inbox Triage",
      "agent": "Brackett",
      "run": 1,
      "score": 0.7348,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Customer Returns Inbox Triage",
      "agent": "Brackett",
      "run": 2,
      "score": 0.8255,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Customer Returns Inbox Triage",
      "agent": "Brackett",
      "run": 3,
      "score": 0.9045,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Customer Returns Inbox Triage",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Customer Returns Inbox Triage",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Discount Request Review",
      "agent": "Codex",
      "run": 1,
      "score": 0.755,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Discount Request Review",
      "agent": "Brackett",
      "run": 1,
      "score": 0.8037,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Discount Request Review",
      "agent": "Brackett",
      "run": 2,
      "score": 0.805,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Discount Request Review",
      "agent": "Brackett",
      "run": 3,
      "score": 0.897,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Discount Request Review",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Discount Request Review",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Job Offer Follow-Up",
      "agent": "Codex",
      "run": 1,
      "score": 0.45299999999999996,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Job Offer Follow-Up",
      "agent": "Brackett",
      "run": 1,
      "score": 0.94,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Job Offer Follow-Up",
      "agent": "Brackett",
      "run": 2,
      "score": 0.41,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Job Offer Follow-Up",
      "agent": "Brackett",
      "run": 3,
      "score": 0.545,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Job Offer Follow-Up",
      "agent": "Claude",
      "run": 1,
      "score": 0.6054,
      "completed": true,
      "model": "Sonnet 5"
    },
    {
      "case": "Job Offer Follow-Up",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Job Requisition Broadcast",
      "agent": "Codex",
      "run": 1,
      "score": 0.625,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Job Requisition Broadcast",
      "agent": "Brackett",
      "run": 1,
      "score": 0.98,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Job Requisition Broadcast",
      "agent": "Brackett",
      "run": 2,
      "score": 0.94,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Job Requisition Broadcast",
      "agent": "Brackett",
      "run": 3,
      "score": 0.555,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Job Requisition Broadcast",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Job Requisition Broadcast",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Job Requisition Triage",
      "agent": "Codex",
      "run": 1,
      "score": 0.365,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Job Requisition Triage",
      "agent": "Brackett",
      "run": 1,
      "score": 0.84,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Job Requisition Triage",
      "agent": "Brackett",
      "run": 2,
      "score": 0.46,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Job Requisition Triage",
      "agent": "Brackett",
      "run": 3,
      "score": 0.64,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Job Requisition Triage",
      "agent": "Claude",
      "run": 1,
      "score": 0.6629,
      "completed": true,
      "model": "Sonnet 5"
    },
    {
      "case": "Job Requisition Triage",
      "agent": "Claude",
      "run": 2,
      "score": 0.682,
      "completed": true,
      "model": "Sonnet 5"
    },
    {
      "case": "Limited-Stock Order Allocation",
      "agent": "Codex",
      "run": 1,
      "score": 0.257,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Limited-Stock Order Allocation",
      "agent": "Brackett",
      "run": 1,
      "score": 0.925,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Limited-Stock Order Allocation",
      "agent": "Brackett",
      "run": 2,
      "score": 0.92,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Limited-Stock Order Allocation",
      "agent": "Brackett",
      "run": 3,
      "score": 0.717,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Limited-Stock Order Allocation",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Limited-Stock Order Allocation",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Low-Stock Replenishment Sweep",
      "agent": "Codex",
      "run": 1,
      "score": 0.855,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Low-Stock Replenishment Sweep",
      "agent": "Brackett",
      "run": 1,
      "score": 1.0,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Low-Stock Replenishment Sweep",
      "agent": "Brackett",
      "run": 2,
      "score": 1.0,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Low-Stock Replenishment Sweep",
      "agent": "Brackett",
      "run": 3,
      "score": 0.97,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Low-Stock Replenishment Sweep",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Low-Stock Replenishment Sweep",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Order Request Replenishment",
      "agent": "Codex",
      "run": 1,
      "score": 0.915,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Order Request Replenishment",
      "agent": "Brackett",
      "run": 1,
      "score": 0.81,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Order Request Replenishment",
      "agent": "Brackett",
      "run": 2,
      "score": 0.82,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Order Request Replenishment",
      "agent": "Brackett",
      "run": 3,
      "score": 0.83,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Order Request Replenishment",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Order Request Replenishment",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Recruiter Performance Scorecard",
      "agent": "Codex",
      "run": 1,
      "score": 0.25,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Recruiter Performance Scorecard",
      "agent": "Brackett",
      "run": 1,
      "score": 0.855,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Recruiter Performance Scorecard",
      "agent": "Brackett",
      "run": 2,
      "score": 0.97,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Recruiter Performance Scorecard",
      "agent": "Brackett",
      "run": 3,
      "score": 0.755,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Recruiter Performance Scorecard",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Recruiter Performance Scorecard",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Accounts Receivable Collections Follow-Up 1",
      "agent": "Codex",
      "run": 1,
      "score": 0.853,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Accounts Receivable Collections Follow-Up 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.805,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Accounts Receivable Collections Follow-Up 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.8,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Accounts Receivable Collections Follow-Up 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.77,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Accounts Receivable Collections Follow-Up 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Accounts Receivable Collections Follow-Up 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Candidate–Role Rematching 1",
      "agent": "Codex",
      "run": 1,
      "score": 0.33,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Candidate–Role Rematching 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.885,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Candidate–Role Rematching 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.875,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Candidate–Role Rematching 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.812,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Candidate–Role Rematching 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.885,
      "completed": true,
      "model": "Sonnet 5"
    },
    {
      "case": "Candidate–Role Rematching 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.91,
      "completed": true,
      "model": "Sonnet 5"
    },
    {
      "case": "Credit Release Queue 1",
      "agent": "Codex",
      "run": 1,
      "score": 0.737,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Credit Release Queue 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.945,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Credit Release Queue 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.97,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Credit Release Queue 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.73,
      "completed": true,
      "model": "Sonnet 5"
    },
    {
      "case": "Credit Release Queue 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.728,
      "completed": true,
      "model": "Sonnet 5"
    },
    {
      "case": "Customer Support Issue Triage 1",
      "agent": "Codex",
      "run": 1,
      "score": 0.645,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Customer Support Issue Triage 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.7602,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Customer Support Issue Triage 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.8265,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Customer Support Issue Triage 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.777,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Customer Support Issue Triage 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Customer Support Issue Triage 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Discount Request Review 1",
      "agent": "Codex",
      "run": 1,
      "score": 0.488,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Discount Request Review 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.97,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Discount Request Review 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.97,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Discount Request Review 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.93,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Discount Request Review 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Discount Request Review 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Email Purchase Invoice Posting 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.825,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Email Purchase Invoice Posting 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.785,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Email Purchase Invoice Posting 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.67,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Email Purchase Invoice Posting 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Email Purchase Invoice Posting 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Freight Consolidation Planning 1",
      "agent": "Codex",
      "run": 1,
      "score": 0.345,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Freight Consolidation Planning 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.39,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Freight Consolidation Planning 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.94,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Freight Consolidation Planning 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.697,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Freight Consolidation Planning 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Freight Consolidation Planning 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Goods Receipt Shortfall Follow-Up 1",
      "agent": "Codex",
      "run": 1,
      "score": 0.695,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Goods Receipt Shortfall Follow-Up 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.89,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Goods Receipt Shortfall Follow-Up 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.88,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Goods Receipt Shortfall Follow-Up 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.835,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Goods Receipt Shortfall Follow-Up 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Goods Receipt Shortfall Follow-Up 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Inbound Sales Enquiry Routing 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.94,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Inbound Sales Enquiry Routing 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.92,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Inbound Sales Enquiry Routing 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.778,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Inbound Sales Enquiry Routing 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Inbound Sales Enquiry Routing 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Job Requisition Broadcast 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.95,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Job Requisition Broadcast 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.96,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Job Requisition Broadcast 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.95,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Job Requisition Broadcast 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Job Requisition Broadcast 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Job Requisition Triage 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.95,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Job Requisition Triage 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.83,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Job Requisition Triage 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.725,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Job Requisition Triage 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Job Requisition Triage 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.9,
      "completed": true,
      "model": "Sonnet 5"
    },
    {
      "case": "Leave Request Eligibility Review 1",
      "agent": "Codex",
      "run": 1,
      "score": 0.9,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Leave Request Eligibility Review 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.785,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Leave Request Eligibility Review 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.78,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Leave Request Eligibility Review 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.76,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Leave Request Eligibility Review 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Leave Request Eligibility Review 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Limited-Stock Order Allocation 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.84,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Limited-Stock Order Allocation 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.84,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Limited-Stock Order Allocation 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.585,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Limited-Stock Order Allocation 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Limited-Stock Order Allocation 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Low-Stock Replenishment Sweep 1",
      "agent": "Codex",
      "run": 1,
      "score": 0.29,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Low-Stock Replenishment Sweep 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.665,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Low-Stock Replenishment Sweep 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.745,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Low-Stock Replenishment Sweep 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.83,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Low-Stock Replenishment Sweep 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.6483,
      "completed": true,
      "model": "Sonnet 5"
    },
    {
      "case": "Low-Stock Replenishment Sweep 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "New-Hire Onboarding Coordination 1",
      "agent": "Codex",
      "run": 1,
      "score": 0.435,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "New-Hire Onboarding Coordination 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.955,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "New-Hire Onboarding Coordination 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.91,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "New-Hire Onboarding Coordination 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.775,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "New-Hire Onboarding Coordination 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "New-Hire Onboarding Coordination 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Order Request Replenishment 1",
      "agent": "Brackett",
      "run": 1,
      "score": 1.0,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Order Request Replenishment 1",
      "agent": "Brackett",
      "run": 2,
      "score": 1.0,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Order Request Replenishment 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.91,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Order Request Replenishment 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Order Request Replenishment 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.92,
      "completed": true,
      "model": "Sonnet 5"
    },
    {
      "case": "Purchase Invoice Reconciliation 1",
      "agent": "Codex",
      "run": 1,
      "score": 0.895,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Purchase Invoice Reconciliation 1",
      "agent": "Brackett",
      "run": 1,
      "score": 1.0,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Purchase Invoice Reconciliation 1",
      "agent": "Brackett",
      "run": 2,
      "score": 1.0,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Purchase Invoice Reconciliation 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.9,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Purchase Invoice Reconciliation 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Purchase Invoice Reconciliation 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Recruiter Performance Scorecard 1",
      "agent": "Codex",
      "run": 1,
      "score": 0.565,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Recruiter Performance Scorecard 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.925,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Recruiter Performance Scorecard 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.925,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Recruiter Performance Scorecard 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.8,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Recruiter Performance Scorecard 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.84,
      "completed": true,
      "model": "Sonnet 5"
    },
    {
      "case": "Recruiter Performance Scorecard 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Vehicle Maintenance Service Booking 1",
      "agent": "Codex",
      "run": 1,
      "score": 0.8,
      "completed": true,
      "model": "Not recorded"
    },
    {
      "case": "Vehicle Maintenance Service Booking 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.6,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Vehicle Maintenance Service Booking 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.73,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Vehicle Maintenance Service Booking 1",
      "agent": "Brackett",
      "run": 3,
      "score": 0.834,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Vehicle Maintenance Service Booking 1",
      "agent": "Claude",
      "run": 1,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Vehicle Maintenance Service Booking 1",
      "agent": "Claude",
      "run": 2,
      "score": 0.0,
      "completed": false,
      "model": "Sonnet 5"
    },
    {
      "case": "Return Eligibility and RMA",
      "agent": "Brackett",
      "run": 1,
      "score": 0.355,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Return Eligibility and RMA",
      "agent": "Brackett",
      "run": 2,
      "score": 0.528,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "RFQ Quote Award",
      "agent": "Brackett",
      "run": 1,
      "score": 0.7,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "RFQ Quote Award",
      "agent": "Brackett",
      "run": 2,
      "score": 0.885,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Shortage Substitution Queue",
      "agent": "Brackett",
      "run": 1,
      "score": 0.687,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Shortage Substitution Queue",
      "agent": "Brackett",
      "run": 2,
      "score": 0.605,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Return Eligibility and RMA 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.48,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Return Eligibility and RMA 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.64,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "RFQ Quote Award 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.73,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "RFQ Quote Award 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.8,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Shortage Substitution Queue 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.44,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Shortage Substitution Queue 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.455,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Supplier Price Increase Review 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.462,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Supplier Price Increase Review 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.39,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Warranty Claim Intake 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.73,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Warranty Claim Intake 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.685,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Weather-Aware Departure Review 1",
      "agent": "Brackett",
      "run": 1,
      "score": 0.47,
      "completed": true,
      "model": "Fused/multiple models"
    },
    {
      "case": "Weather-Aware Departure Review 1",
      "agent": "Brackett",
      "run": 2,
      "score": 0.48,
      "completed": true,
      "model": "Fused/multiple models"
    }
  ],
  "model_labels": {
    "Brackett": {
      "Fused/multiple models": 107
    },
    "Claude": {
      "Sonnet 5": 60
    },
    "Codex": {
      "Not recorded": 24
    }
  },
  "metadata": {
    "snapshot_date": "2026-09-12",
    "source": "Brackett ShowTell Benchmark Results.xlsx",
    "source_sha256": "7034b3049c4f1090b512b82996dca9840c197cbe0b8daf9db850d400e42f1d18",
    "sheet": "Overview",
    "score_source": "Overview run cells, cross-checked against matching Detailed Analysis rows; workbook average formulas are not used.",
    "status": "Preliminary; selected attempts; non-canonical grading (manual grades and imported model-assisted grades, as recorded in the workbook notes).",
    "methodology": "NA = attempted and incomplete, scored zero. Blank = untested, excluded. Average attempts within each use case, then weight use cases equally. Shared cases require an attempt from every system.",
    "completion_definition": "An attempt has a numeric benchmark score; not a measure of operational task success.",
    "system_configuration": "System/model labels: Brackett: Fused/multiple models (107 selected attempts). Claude: Sonnet 5 (60 selected attempts). Codex: Not recorded (24 selected attempts). Brackett uses the author-specified fused/multiple-model description. Other model labels are workbook-reported, not verified model IDs; model and adapter versions are not fully pinned. Grader names in comments do not identify the evaluated model.",
    "model_source": "Brackett: author-specified Fused/multiple models system description; component model labels are not published. Other systems: Detailed Analysis, Model column; missing labels remain Not recorded.",
    "uncertainty": "No confidence interval is reported for this selected-attempt snapshot.",
    "colors": {
      "Brackett": "#f5b840",
      "Claude": "#a89acd",
      "Codex": "#65aeb5"
    }
  }
}