Glossary
Speech analytics
Software that transcribes and analyzes spoken audio, such as call center recordings, to extract sentiment, keywords, and compliance signals.
Speech analytics is software that processes recorded or live spoken audio, most commonly customer service calls, transcribes it, and analyzes the transcript and audio signal for content, sentiment, and compliance.
A speech analytics pipeline typically combines automatic speech recognition to produce a transcript with text analytics techniques like sentiment analysis and keyword spotting applied to that transcript, and can also analyze acoustic features directly, tone, pace, silence, that a transcript alone would miss. This differs from audio analytics in general in that speech analytics is specifically focused on spoken language content rather than sound more broadly.
Contact centers use speech analytics to monitor agent compliance with required scripts, flag calls at risk of escalation, and feed conversation intelligence tools that surface coaching opportunities, often alongside metrics like average handle time. The main pitfalls are transcription errors compounding into wrong downstream analysis, particularly with accents, background noise, or industry jargon, and privacy or consent requirements around recording and analyzing customer conversations, which vary by jurisdiction.
Last reviewed September 22, 2026