cs.AI, cs.CL

Evaluating AI Meeting Summaries with a Reusable Cross-Domain Pipeline

arXiv:2604.21345v2 Announce Type: replace
Abstract: Industrial teams often deploy large language model features before stable regression or model selection evaluation exists. We present a reusable evaluation system for AI meeting summaries that combin…