Glossary
Words we actually use
Click a related product page after each definition.
Auto QA
Using speech-to-text and a model to grade calls against a scorecard without a human listening first.
See it on a sample call →Calibration
Scoring the same call with AI and humans, then tightening prompts until they agree.
See it on a sample call →Conversation intelligence
Topics, sentiment, objections, competitor mentions, and other insights extracted from transcripts.
See it on a sample call →Hybrid QA
AI grades binary and high-volume criteria. Humans grade nuance, disputes, and edge cases.
See it on a sample call →Scorecard
The rubric used to grade a call: questions, weights, fail conditions, and coaching notes.
See it on a sample call →Sample size
How many calls you must review to trust a metric. At 2% coverage, most teams are guessing.
See it on a sample call →Generate an AI call evaluation, for free.
Open a sample call, click every score, then email us a recording of your own.
No credit cardClick every score on a real sampleEmail support@qacalls.ai anytime