TwinBI: An Agentic Digital Twin for Efficient Augmented Interactions with Business Intelligence Dashboards

In three linesTwinBI is an agentic digital-twin framework coupling an LLM-based agent system with executable BI dashboard state. It unifies conversational interaction, dashboard manipulation, and provenance tracking through a shared interaction log. Benchmark: exact-match accuracy 43.3% → 63.3%, timeout rate 40% → 10%.

Summary generated by Claude — human-verified