AI Review Hub
AI-powered document review built on the Document Management module's reference library. Documents are chunked (~1,500 tokens with 200-token overlap), embedded via Gemini text-embedding-004 (768 dimensions), and indexed with pgvector IVFFlat for cosine similarity search. Permit applications and compliance documents are analyzed against the reference library — building codes, ordinances, and policies — to surface relevant requirements and flag potential issues. PII detection uses pattern-based identification (SSN, email, phone, DOB, financial data) for FOIA redaction support. The interactive Redaction Studio provides a PDF overlay interface with canvas-based drag-to-draw manual redaction and permanent PDF burning. Review queues manage workload with confidence scoring and human override capability.
Key Capabilities
- Gemini text-embedding-004 with pgvector cosine similarity search
- Document chunking (~1,500 tokens, 200-token overlap) with IVFFlat indexing
- Permit/compliance analysis against reference library (codes, ordinances, policies)
- PII detection: SSN, email, phone, DOB, financial patterns for FOIA
- Interactive Redaction Studio with PDF overlay and canvas-based manual redaction
- Review queue management with confidence scoring and human override
Related Modules in Analytics & Intelligence
Reporting & Analytics
Cross-module report builder with CAFR generation and scheduled delivery
GIS Mapping
PostGIS spatial backbone with geocoding and parcel identification
Data Warehouse
ETL pipelines and cross-module data aggregation for analytics
ETL Import
Data migration with field mapping, validation, and rollback
See AI Review Hub in action
Get a personalized demo tailored to your municipality's needs.