SiPAJu — Judicial Propensity Prediction
Equatorial Energia / ANEEL P&D
Highlights
- Consumer risk profiling from multi-source CRM/ERP data
- Two-stage pipeline: topic profiling + active risk scoring
- Reduces legal costs and improves conflict resolution efficiency
SiPAJu (Judicial Actions Prediction System) was developed to help Equatorial Energia anticipate and prevent consumer lawsuits before they are filed.
Stage 1 — Consumer Profiling: Historical records spanning energy outages, improper billing, and other complaint categories are extracted from CP-PRO, SAP-CCS, SAP-CRM, and judicial tribunal (TI/TRF) databases. A topic modeling pipeline assigns each consumer a multidimensional dissatisfaction profile.
Stage 2 — Active Risk Scoring: Active consumers are scored by complaint topic, producing a risk tier (low, medium-low, medium-high, high). A management information layer translates risk scores into prioritized action plans: proactive outreach, maintenance scheduling, billing review, or relationship management interventions.
Expected outcomes: improved distributor image, higher consumer satisfaction, minimized legal costs, reduced operational expenses, identification of service failure patterns, and efficiency gains in resolving legal conflicts.