PROVENANCE · CITATIONS
Every score, sourced.
Transparent data provenance for every PACCS decision
PACCS is committed to full transparency in AI decision-making. Every data source used in analysis is documented, versioned, and traceable. This citation system is part of our commitment to explainable and accountable AI systems.
4,726+
Source records
4
Distinct data layers
By design
Decisions auditable
PRIMARY DATA SOURCES
What feeds the engine.
Internal · Curated
Comprehensive database of 4,726+ analyzed films with festival outcomes, distribution deals, and performance metrics. Used for comparable film matching and success prediction.
Live · Multi-Agent
Multi-agent intelligence engine used for narrative analysis, dialogue evaluation, and thematic assessment. Powers the Narrative Agent and Visual Language Agent.
Partner · Curated
Historical programming data from 22 anchor festivals covering 7 years, with the corpus expanding through the PIFF 2026 partner cohort. Drives the Festival Alignment Agent's pattern matching.
Multi-Agent · Live
Distributor pickup, territorial release patterns, and revenue benchmarks derived from the PHFF corpus's seven years of structured release data. Drives the Marketability Agent's revenue scenarios.