Schema Library

Reference Check output schema

Reference Check output schema: JSON fields, CSV headers, validation rules, and integration notes for structured AI phone call outcomes.

Browse schemas

JSON schema

{
  "outcome": "Reference check outcome",
  "fields": [
    {
      "description": "reference name captured during the call.",
      "name": "reference_name",
      "required": true,
      "type": "text"
    },
    {
      "description": "phone number from caller id captured during the call.",
      "name": "phone_number_from_caller_id",
      "required": true,
      "type": "text"
    },
    {
      "description": "candidate name captured during the call.",
      "name": "candidate_name",
      "required": true,
      "type": "date"
    },
    {
      "description": "relationship captured during the call.",
      "name": "relationship",
      "required": true,
      "type": "text"
    },
    {
      "description": "strengths captured during the call.",
      "name": "strengths",
      "required": true,
      "type": "text"
    },
    {
      "description": "concerns captured during the call.",
      "name": "concerns",
      "required": false,
      "type": "text"
    },
    {
      "description": "rehire signal captured during the call.",
      "name": "rehire_signal",
      "required": false,
      "type": "text"
    },
    {
      "description": "notes captured during the call.",
      "name": "notes",
      "required": false,
      "type": "text"
    }
  ]
}

CSV headers

reference_name,phone_number_from_caller_id,candidate_name,relationship,strengths,concerns,rehire_signal,notes

Validation rules

  • Required fields should be asked directly
  • Dates and times should include timezone when relevant
  • Urgency and handoff fields should use a controlled vocabulary
  • Unknown values should be saved as unknown, not invented

Structured output schema

{
  "outcome": "Reference Check output schema outcome",
  "fields": [
    {
      "description": "reference name captured during the call.",
      "name": "reference_name",
      "required": true,
      "type": "text"
    },
    {
      "description": "phone number from caller id captured during the call.",
      "name": "phone_number_from_caller_id",
      "required": true,
      "type": "text"
    },
    {
      "description": "candidate name captured during the call.",
      "name": "candidate_name",
      "required": true,
      "type": "date"
    },
    {
      "description": "relationship captured during the call.",
      "name": "relationship",
      "required": true,
      "type": "text"
    },
    {
      "description": "strengths captured during the call.",
      "name": "strengths",
      "required": true,
      "type": "text"
    },
    {
      "description": "concerns captured during the call.",
      "name": "concerns",
      "required": false,
      "type": "text"
    },
    {
      "description": "rehire signal captured during the call.",
      "name": "rehire_signal",
      "required": false,
      "type": "text"
    },
    {
      "description": "notes captured during the call.",
      "name": "notes",
      "required": false,
      "type": "text"
    }
  ]
}

Quality gate

  • Unique workflow artifact on the page
  • Visible script, checklist, or question set
  • Human handoff rules for sensitive calls
  • Product CTA that creates a real Call App
  • Internal links to related templates and compliance pages

How to use this resource

Build the first version

Use the page as a launch brief, not a finished policy document. Copy the prompt into the builder, keep the workflow limited to one caller intent, and test whether the first two questions collect enough context for a useful owner follow-up.

Review the output

Before publishing, check that the saved outcome has useful fields for the team: reference name, phone number from caller id, candidate name. Remove fields that are not needed, and route regulated, urgent, or ambiguous calls to a person instead of forcing automation.

Measure quality

Track whether callers finish the flow, whether owners understand the summary, and whether handoff rules trigger at the right time. A useful Call App should reduce missed context, not just answer the phone with a longer script.

Improve the page

When this workflow starts getting impressions or demo calls, add real examples from the use case: a better transcript, a refined schema, integration notes, and clearer exclusions for cases the AI phone workflow should not handle.

FAQ

Who should use Reference Check output schema?

Reference Check output schema is for teams that want a callable workflow instead of a static page or loose voicemail. It is most useful when callers need to explain context out loud and the team needs a repeatable result with reference name, phone number from caller id, candidate name.

What should I customize before launching it?

Start with the json schema, then edit the opening disclosure, required questions, handoff rules, and owner notification. Keep the first version narrow: one caller goal, one output schema, and one clear next action after the call.

How does this connect back to CallURL?

The page is written as a practical starting point for a real Call App. Use the visible prompt, schema, examples, and related workflow links to create a working phone flow, then test it with routine, incomplete, and sensitive caller scenarios.

What makes this page different from a generic AI phone agent page?

This resource is centered on a specific workflow artifact, not broad product copy. It includes concrete fields, handoff rules, related templates, and a demo entry point so a visitor can judge whether reference check output schema fits their use case.

Related pages

Create this Call App

Use the prompt, schema, handoff rules, and examples on this page as the starting point for a working CallURL Call App.