Jev in front of a support inbox cuts 70% of premium-model tickets
An arXiv paper cited in the post says 70% of an AI support inbox never needed a model, with Jev making one cheap classification call per message for intent, urgency and complexity before simple tickets go to a database lookup.
This is insane and almost nobody has seen it.. A paper that hit arXiv on October 12 says 70% of an AI support inbox never needed a model in the first place. Their example store gets 100,000 support messages a month and sends every single one to a premium model. Bill: $7,950. Then Jev goes in front. One cheap classification call per message reads intent, urgency and complexity, and attaches a confidence score. -> 45% are "Where is my order?" A database lookup answers them for $0. -> 25% are p