Case Study

HUMAN-PROVEN REVIEWS

HUMAN-PROVEN REVIEWS

Pages 1 Pages

Many practices rely on call transcriptions to identify patient appointment opportunities calls. They end up making critical operational, staffing, and marketing decisions based on the assumption that patients will say the word “appointment” on every appointment opportunity call. In reality, this happens on less than half of such calls. To help providers capture more revenue and grow their patient base with detailed call analytics, Call Box looked at transcriptions from dozens of human-confirmed appointment opportunity calls. It turns out IBM Watson’s Speech to Text engine spotted the word “appointment” on only 44.8% of the test set. It took human reviewers to recognize the patient was calling for an appointment on the other 55.2% of calls.

Join for free to read