Skip to content

Insight Models

This page documents the data models related to AI-powered insights, sessions, and Q&A.


InsightSession

Represents a session of generated insights.

Properties

Property Type Description
session_id string Session identifier
status string Session status: pending, generating, completed, failed
regeneration_count integer Times insights have been regenerated (max 3)
insights Insight[] Generated insights for this session
created_at string (ISO 8601) Creation timestamp

Insight

Represents a single AI-generated insight.

Properties

Property Type Description
id string Insight identifier
category InsightCategory Insight category
category_display string Human-readable category name
title string Insight title
summary string Brief insight description
confidence number Confidence score (0.0 - 1.0)
priority integer Priority ranking (1 = highest)
is_taster boolean Whether this is a preview insight
methodology string? How the insight was derived
supporting_data object Data points supporting the insight
created_at string (ISO 8601) Creation timestamp
recent_questions object[] Recent Q&A for this insight

Example

{
  "id": "ins_001",
  "category": "financial_health",
  "category_display": "Financial Health",
  "title": "Portfolio concentration risk increasing",
  "summary": "Top 5 obligors now represent 32% of total par, up from 28% last quarter.",
  "confidence": 0.92,
  "priority": 1,
  "is_taster": false,
  "methodology": "Statistical analysis of obligor exposure trends over trailing 4 quarters",
  "supporting_data": {
    "top_5_exposure_current": 0.32,
    "top_5_exposure_previous": 0.28,
    "threshold": 0.35
  },
  "created_at": "2026-01-25T10:30:00Z",
  "recent_questions": []
}

Nomination

Represents a workbook's nomination status for insight analysis.

Properties

Property Type Description
workbook_id string (UUID) Workbook identifier
workbook_name string Workbook display name
is_nominated boolean Whether nominated for analysis
nominated_at string (ISO 8601)? When nominated

QuestionResponse

Response to a natural language question.

Properties

Property Type Description
question string Original question
answer string AI-generated answer
insight_id string? Related insight (if question was about a specific insight)
asked_at string (ISO 8601) When question was asked

Enumerations

InsightCategory

Value Display Description
financial_health Financial Health Portfolio-level financial metrics and trends
operational_efficiency Operational Efficiency Process and workflow optimization
risk_compliance Risk & Compliance Risk exposures and compliance concerns
trends_forecasts Trends & Forecasts Data patterns and forward-looking analysis
recommendations Recommendations Actionable improvement suggestions

Stakeholder Types

Value Focus
executive High-level KPIs, strategic trends
finance Detailed financial metrics, variances
operations Efficiency metrics, resource allocation
analyst Data patterns, statistical insights