What the invention does
The Griot captures media, captions, translations, summaries and other outputs in a synchronized record. It tracks where each element came from and produces replay views appropriate to the viewer’s authorization and purpose.
Approved abstract
A system and method for comprehensive capture and intelligent replay of online communication sessions. The system records primary audio and video streams alongside supplementary data including real-time captions, translations, AI-generated summaries, agent coaching prompts, and session metadata in a temporally-aligned unified data structure maintaining synchronization across all captured elements. A provenance graph tracks data lineage including source identification, AI model attribution, and processing infrastructure details for each information component. The system generates role-based replay views from the unified recording, dynamically rendering different presentations based on viewer authorization and purpose. Customer-facing views exclude internal annotations, quality assurance views incorporate coaching data and performance metrics, and compliance views include complete provenance information detailing AI model versions and data processing locations. This architecture enables single-capture, multi-purpose replay functionality while maintaining comprehensive audit trails for regulatory compliance and operational analysis.
Filing and status
Application 63/934,099
Filed December 8, 2025
Patent pending · provisional filed