What the invention does
Muninn identifies which AI assistance each participant needs and selects where it runs, across a device, an edge server or the cloud. It adjusts service allocation as participants, network conditions and privacy requirements change.
Approved abstract
A system and method for dynamically managing artificial intelligence services during real-time communication sessions. The system identifies individual participant requirements for AI-enabled features including language translation, captioning, sign language interpretation, and real-time coaching. A resource orchestration engine determines optimal computational deployment locations for each AI service, selecting among cloud infrastructure, edge servers, and local participant devices based on latency requirements, privacy preferences, and available bandwidth. The system continuously monitors session parameters and participant changes, automatically reconfiguring AI service allocation and deployment topology in response to detected variations in network conditions, participant composition, or accessibility requirements. Adaptive load balancing ensures uninterrupted service delivery while optimizing resource utilization across heterogeneous computing environments. The invention enables personalized, context-aware AI assistance for each session participant while maintaining seamless real-time communication flow through transparent background optimization of distributed AI resources.
Filing and status
Application 63/934,090
Filed December 8, 2025
Patent pending · provisional filed