Overview

AUTO-AUCT-044, Approve High-Risk Purchase, determines whether a high-risk auction vehicle purchase should be approved, escalated, deferred, or rejected.

Decision ID

AUTO-AUCT-044

Decision Name

Approve High-Risk Purchase

Decision Preparation Model

auto-auctions-approve-high-risk-purchase-auto-auct-044 version 1.0.0

Default Profile

balanced

Default Scenario

standard

Catalog

DKR-AUTO-RUNTIME-001, version 13.9.3

Understanding This Decision

Approve High-Risk Purchase helps Compliance & Risk teams compare candidate purchase cases when profit opportunity must be weighed against risk exposure, title confidence, repair uncertainty, and management priority.

Criteria

Criterion IDs are intentionally stable machine identifiers. Display labels are for users; request values should be keyed by canonical criterionId.

Data Preparation Guide

Constraint Processing

This decision includes one catalog-defined hard constraint. Decision Catalog returns the constraint with the decision detail, and Decision Service evaluates it before criteria-based ranking.

{
  "constraintSummary": {
    "definedConstraintCount": 1,
    "activeConstraintCount": 1,
    "eligibleOptionCount": 2,
    "excludedOptionCount": 1,
    "eligibleOptions": [
      "option-1",
      "option-2"
    ],
    "excludedOptions": [
      {
        "optionId": "option-3",
        "reasons": [
          "Title confidence is below the minimum threshold of 80."
        ]
      }
    ]
  }
}

const constraints = decisionDetail.constraints || [];
if (constraints.length > 0) {
  // Decision Service applies catalog constraints before ranking.
  // Excluded candidates appear in response.constraintSummary.excludedOptions.
}

const summary = response.constraintSummary;
const excluded = response.decisionResult?.excludedOptions || [];

Excluded candidates do not participate in ranking because hard constraints are evaluated before scoring. For AUTO-AUCT-044, option-3 is excluded from the showcase sample because its title confidence is below the catalog threshold.

Decision Preparation Model

The published Decision Preparation Model validates and transforms illustrative Business Data into the canonical criteria required by AUTO-AUCT-044 before Decision Service applies ranking.

Profile ID

auto-auctions-approve-high-risk-purchase-auto-auct-044

Version

1.0.0

Options Path

$.auctionLots

Option ID Path

$.optionId

Display Name Path

$.name

{
  "requestContext": {
    "sourceSystem": "automotive-profile-factory",
    "correlationId": "factory-auto-auct-044"
  },
  "auctionLots": [
    {
      "optionId": "LOT-044-01",
      "name": "Approve High-Risk Purchase Option 1",
      "financial": {
        "expectedProfitMargin": 55000.0
      },
      "risk": {
        "riskExposure": 35.0
      },
      "compliance": {
        "and": {
          "risk": {
            "titleConfidence": 88.0,
            "repairUncertainty": 83.0,
            "managementPriority": 29.0
          }
        }
      }
    },
    {
      "optionId": "LOT-044-02",
      "name": "Approve High-Risk Purchase Option 2",
      "financial": {
        "expectedProfitMargin": 90000.0
      },
      "risk": {
        "riskExposure": 42.0
      },
      "compliance": {
        "and": {
          "risk": {
            "titleConfidence": 92.0,
            "repairUncertainty": 45.0,
            "managementPriority": 50.0
          }
        }
      }
    }
  ]
}

{
  "optionId": "LOT-044-01",
  "name": "Approve High-Risk Purchase Option 1",
  "values": {
    "expected_profit_margin": 55000.0,
    "risk_exposure": 35.0,
    "title_confidence": 88.0,
    "repair_uncertainty": 83.0,
    "management_priority": 29.0
  }
}

Profiles and Scenarios

This example selects a Profile for evaluation emphasis and a Scenario for operating context. Discover both from the selected Decision Catalog definition.

Prepared Input

For Prepared Criteria Mode, send decisionId, selected profile/scenario IDs, and option values keyed by canonical criterion ID. This request intentionally does not duplicate values under display labels.

{
  "decisionId": "AUTO-AUCT-044",
  "profileId": "balanced",
  "scenarioId": "standard",
  "algorithm": "TOPSIS",
  "weightStrategy": "Expert",
  "runSensitivity": false,
  "requestContext": {
    "correlationId": "auto-auct-044-demo-001"
  },
  "options": [
    {
      "optionId": "option-1",
      "name": "Controlled High-Margin Purchase",
      "values": {
        "expected_profit_margin": 5200,
        "risk_exposure": 12,
        "title_confidence": 90,
        "repair_uncertainty": 12,
        "management_priority": 90
      }
    },
    {
      "optionId": "option-2",
      "name": "Borderline Risk Purchase",
      "values": {
        "expected_profit_margin": 4700,
        "risk_exposure": 22,
        "title_confidence": 82,
        "repair_uncertainty": 22,
        "management_priority": 82
      }
    },
    {
      "optionId": "option-3",
      "name": "Defer Pending Documentation",
      "values": {
        "expected_profit_margin": 4300,
        "risk_exposure": 34,
        "title_confidence": 74,
        "repair_uncertainty": 34,
        "management_priority": 74
      }
    }
  ]
}

Execute

curl -X POST "https://dde.vinquery.com/api/v1/decide" \
  -H "Authorization: Bearer ${DECISIOQ_TOKEN}" \
  -H "Content-Type: application/json" \
  -H "X-Correlation-Id: auto-auct-044-demo-001" \
  --data @auto-auct-044-execute.json

Decision Service retrieves authoritative criteria, validation metadata, and constraints from Decision Catalog, excludes candidates that fail mandatory hard constraints, executes the deterministic ranking, and returns the decision result plus execution metadata.

Interpret the Result

The successful response includes both the deterministic decisionResult and a top-level explanation object. The explanation adds context but does not select, rerank, recalculate, or override the ranking.

{
  "service": "decisioq",
  "version": "7.6.3",
  "requestId": "0HNE...",
  "operation": "Decide",
  "success": true,
  "decisionType": "AUTO-AUCT-044",
  "decisionVersion": "13.9.3",
  "timestampUtc": "2026-07-19T00:00:00Z",
  "decisionResult": {
    "winner": "option-1",
    "confidence": 91,
    "ranking": [
      {
        "optionId": "option-1",
        "score": 0.91,
        "breakdown": {
          "expected_profit_margin": 0.25,
          "risk_exposure": 0.30,
          "title_confidence": 0.15,
          "repair_uncertainty": 0.15,
          "management_priority": 0.15
        },
        "normalizationBreakdown": {}
      },
      {
        "optionId": "option-2",
        "score": 0.84,
        "breakdown": {},
        "normalizationBreakdown": {}
      }
    ],
    "excludedOptions": [
      {
        "optionId": "option-3",
        "reasons": [
          "Title confidence is below the minimum threshold of 80."
        ]
      }
    ]
  },
  "explanation": {
      "summary": "The selected purchase case provided the strongest overall balance of margin, risk, confidence, and management priority.",
      "whyRecommended": "It met the hard title-confidence threshold and ranked highest among eligible cases.",
      "keyDrivers": [],
      "tradeoffs": [],
      "competitors": [],
      "sensitivitySummary": "Sensitivity analysis was not included in this response.",
      "scenarioSummary": "The standard scenario was selected.",
      "risks": [],
      "nextSteps": [],
      "assumptions": []
  },
  "constraintSummary": {
    "definedConstraintCount": 1,
    "activeConstraintCount": 1,
    "eligibleOptionCount": 2,
    "excludedOptionCount": 1,
    "eligibleOptions": [
      "option-1",
      "option-2"
    ],
    "excludedOptions": [
      {
        "optionId": "option-3",
        "reasons": [
          "Title confidence is below the minimum threshold of 80."
        ]
      }
    ]
  },
  "warnings": [],
  "requestContext": {
    "correlationId": "auto-auct-044-demo-001"
  }
}

Sensitivity Analysis

runSensitivity is an optional execution flag supported by both Business Data Mode and Prepared Criteria Mode. Set it to true when the client wants recommendation-stability information in the same response. The sensitivity engine analyzes the prepared criteria produced by either input path, perturbs criterion weights by controlled factors, and reports whether the winner remains stable.

{
  "sensitivityResult": {
    "stableWinner": true,
    "winner": "option-1",
    "mostSensitiveCriterion": "risk_exposure",
    "confidence": 95,
    "criterionImpacts": {
      "risk_exposure": 0.1842
    },
    "winnerChangeCounts": {
      "risk_exposure": 0
    }
  }
}

Use sensitivity output to decide whether a recommendation is robust enough for automation or should be reviewed by a person.

Explanation of Decision Result

Every successful response includes a provider-neutral top-level explanation object. It explains the already-finalized deterministic result and does not change the ranking.

{
  "explanation": {
      "summary": "The selected purchase case provided the strongest overall balance of margin, risk, confidence, and management priority.",
      "whyRecommended": "It met the hard title-confidence threshold and ranked highest among eligible cases.",
      "keyDrivers": [],
      "tradeoffs": [],
      "competitors": [],
      "sensitivitySummary": "Sensitivity analysis was not included in this response.",
      "scenarioSummary": "The standard scenario was selected.",
      "risks": [],
      "nextSteps": [],
      "assumptions": []
  }
}

Tracing and Logs

Use request identifiers to connect client, catalog, and execution activity during support or integration testing.

X-Request-Id

Optional client-supplied request ID. If omitted, the server generates one.

X-Correlation-Id

Optional client correlation value propagated from Decision Service to Decision Catalog.

requestContext.correlationId

Optional payload value echoed in the response and used for tracing.

configurationUsed

The authoritative effective weight strategy, ranking algorithm, profile, scenario, sensitivity setting, and source for each value.

diagnostics

Safe execution counts returned only when responseOptions.includeDiagnostics is true.

Code Examples

These examples demonstrate the current DecisioQ flow:

1. Request jwtToken from https://identity.vinquery.com/connect/token.
2. Load AUTO-AUCT-044 metadata from https://dks.vinquery.com/decisioncatalog/decisions/AUTO-AUCT-044.
3. Execute a Prepared Criteria Mode at https://dde.vinquery.com/api/v1/decide.

Set these environment variables before running any companion example:

DECISIOQ_CLIENT_ID
DECISIOQ_CLIENT_SECRET

Optional:
DECISIOQ_AUDIENCE=vinquery:api:decisioq
DECISIOQ_IDENTITY_URL=https://identity.vinquery.com/connect/token
DECISIOQ_DKS_URL=https://dks.vinquery.com
DECISIOQ_DDE_URL=https://dde.vinquery.com

Download the source files directly:

Troubleshooting

Production Checklist