Recipe catalog / claim-facets
Label the facets an insurance claim narrative states
Which of these does claim state: when the incident happened, where it happened, what caused it, what was damaged or lost, and whether there are witnesses or evidence?
A claims intake assistant receives a free-form loss description and needs to know which standard first-notice details are already present so it can pre-fill the claim and ask only for what is missing.
Explore this recipe interactively ยท Source and implementation guide
Use claim-facets in TypeScript
Install with npm install jev-recipes. Requires Node.js 22.9 or newer and ES modules. Set TYPESAFE_API_KEY in your server environment for live calls, which send input to TypeSafe and use API quota. See the installation guide.
import { claimFacets } from 'jev-recipes/claim-facets';
const result = await claimFacets({
"claim": "On the night of Sunday 8 March, sometime after we went to bed around 11 pm, a pipe in the upstairs bathroom at our home at 27 Alder Court burst. By the time we woke at 6 am, water had come through the ceiling of the living room directly below. The living room ceiling plaster has collapsed in two places, the hardwood floor is warped across roughly half the room, and the sofa and a wool rug are soaked. We shut off the main and called a plumber, who replaced the split section of pipe that morning.",
"minConfidence": 0.8
});
console.log(result);Input contract
| Field | Type | Needed |
|---|---|---|
| claim | string | Required |
| minConfidence | number | Optional |
Full input and result schemas
{
"input": {
"$schema": "https://json-schema.org/draft/2020-12/schema",
"type": "object",
"properties": {
"claim": {
"type": "string"
},
"minConfidence": {
"type": "number",
"minimum": 0,
"maximum": 1
}
},
"required": [
"claim"
]
},
"result": {
"$schema": "https://json-schema.org/draft/2020-12/schema",
"type": "object",
"properties": {
"model": {
"type": "string"
},
"usage": {
"type": "object",
"properties": {
"input_tokens": {
"type": "integer",
"minimum": 0,
"maximum": 9007199254740991
},
"output_tokens": {
"type": "integer",
"minimum": 0,
"maximum": 9007199254740991
}
},
"required": [
"input_tokens",
"output_tokens"
],
"additionalProperties": false
},
"status": {
"type": "string",
"enum": [
"ready",
"review"
]
},
"detected": {
"type": "array",
"items": {
"type": "string",
"enum": [
"statesWhen",
"statesWhere",
"statesCause",
"statesDamages",
"statesEvidence"
]
}
},
"labels": {
"type": "object",
"properties": {
"statesWhen": {
"type": "object",
"properties": {
"status": {
"type": "string",
"enum": [
"ready",
"review"
]
},
"verdict": {
"type": "string",
"enum": [
"present",
"absent"
]
},
"probability": {
"type": "number",
"minimum": 0,
"maximum": 1
},
"confidence": {
"type": "number",
"minimum": 0,
"maximum": 1
}
},
"required": [
"status",
"verdict",
"probability",
"confidence"
],
"additionalProperties": false
},
"statesWhere": {
"type": "object",
"properties": {
"status": {
"type": "string",
"enum": [
"ready",
"review"
]
},
"verdict": {
"type": "string",
"enum": [
"present",
"absent"
]
},
"probability": {
"type": "number",
"minimum": 0,
"maximum": 1
},
"confidence": {
"type": "number",
"minimum": 0,
"maximum": 1
}
},
"required": [
"status",
"verdict",
"probability",
"confidence"
],
"additionalProperties": false
},
"statesCause": {
"type": "object",
"properties": {
"status": {
"type": "string",
"enum": [
"ready",
"review"
]
},
"verdict": {
"type": "string",
"enum": [
"present",
"absent"
]
},
"probability": {
"type": "number",
"minimum": 0,
"maximum": 1
},
"confidence": {
"type": "number",
"minimum": 0,
"maximum": 1
}
},
"required": [
"status",
"verdict",
"probability",
"confidence"
],
"additionalProperties": false
},
"statesDamages": {
"type": "object",
"properties": {
"status": {
"type": "string",
"enum": [
"ready",
"review"
]
},
"verdict": {
"type": "string",
"enum": [
"present",
"absent"
]
},
"probability": {
"type": "number",
"minimum": 0,
"maximum": 1
},
"confidence": {
"type": "number",
"minimum": 0,
"maximum": 1
}
},
"required": [
"status",
"verdict",
"probability",
"confidence"
],
"additionalProperties": false
},
"statesEvidence": {
"type": "object",
"properties": {
"status": {
"type": "string",
"enum": [
"ready",
"review"
]
},
"verdict": {
"type": "string",
"enum": [
"present",
"absent"
]
},
"probability": {
"type": "number",
"minimum": 0,
"maximum": 1
},
"confidence": {
"type": "number",
"minimum": 0,
"maximum": 1
}
},
"required": [
"status",
"verdict",
"probability",
"confidence"
],
"additionalProperties": false
}
},
"required": [
"statesWhen",
"statesWhere",
"statesCause",
"statesDamages",
"statesEvidence"
],
"additionalProperties": false
}
},
"required": [
"model",
"usage",
"status",
"detected",
"labels"
],
"additionalProperties": false
}
}Saved example result
This hand-authored response demonstrates the contract. It is not a model accuracy measurement. Run it without an API key: npx jev-recipes demo claim-facets.
{
"model": "demo-fixture",
"usage": {
"input_tokens": 0,
"output_tokens": 0
},
"status": "ready",
"detected": [
"statesWhen",
"statesWhere",
"statesCause",
"statesDamages"
],
"labels": {
"statesWhen": {
"status": "ready",
"verdict": "present",
"probability": 0.96,
"confidence": 0.96
},
"statesWhere": {
"status": "ready",
"verdict": "present",
"probability": 0.95,
"confidence": 0.95
},
"statesCause": {
"status": "ready",
"verdict": "present",
"probability": 0.94,
"confidence": 0.94
},
"statesDamages": {
"status": "ready",
"verdict": "present",
"probability": 0.96,
"confidence": 0.96
},
"statesEvidence": {
"status": "ready",
"verdict": "absent",
"probability": 0.08,
"confidence": 0.92
}
}
}
Evaluation evidence
No verified live accuracy measurement is available. Evaluate representative cases before using this decision in your workflow.
Use the evaluation guide to measure this decision on your own labeled cases.
Limitations
- Reports only whether the wording states each facet, not whether the details are true, covered, or plausible. It makes no coverage, liability, or fraud judgment.
- Labels are independent, so a narrative can state several facets or none. Parse actual dates, addresses, and amounts in application code.
Related recipes
- symptom-facets: Use symptom-facets for the parallel intake check on a patient's symptom description rather than an insurance loss narrative.
- clarify: Use clarify to decide whether to ask the claimant a follow-up question about the facets this recipe finds missing.