Call intelligence glossary
Plain-English definitions for call intelligence, voice biometrics, QA, and compliance, written to answer the question first.
Agent assist
Real-time guidance shown to an agent during a live call.
Agent scorecard
A summary of an agent’s call-quality performance across scored calls.
AI receptionist
An AI voice agent that answers, routes, and handles inbound calls.
Automated call QA
Using AI to score every call against a quality rubric, instead of a manual sample.
Call categorization
Automatically sorting calls into types or reasons, such as billing or complaint.
Call scoring
Rating a call against defined criteria, as pass/fail or on a numeric scale.
Conversation intelligence
Analyzing recorded conversations to surface insight, trends, and coaching signals.
Data subject access request (DSAR)
A request, under laws like GDPR, for the personal data an organization holds on someone.
Disclosure detection
Checking whether a required notice, like a call-recording disclosure, was actually given.
Dynamic number insertion (DNI)
Swapping the phone number shown on a site so calls can be attributed to a source.
Enrollment (voice biometrics)
Capturing a reference voiceprint so a speaker can later be recognized.
False acceptance rate (FAR)
How often a biometric system wrongly accepts an imposter as a match.
First call resolution (FCR)
The share of issues resolved on the first call, without a follow-up.
Knowledge-based authentication (KBA)
Verifying identity by asking questions only the real person should know.
Mini-Miranda
The disclosure debt collectors must give identifying the call as an attempt to collect a debt.
PLDA (Probabilistic Linear Discriminant Analysis)
A calibrated scoring method that decides whether two voiceprints are the same speaker.
Real-time transcription
Transcribing speech as the conversation happens, rather than after it ends.
Redaction
Removing or masking sensitive information from a transcript or recording.
Sentiment analysis
Classifying whether language is positive, negative, or neutral.
Speaker diarization
Splitting a recording into who spoke when.
Speaker identification
Determining who is speaking by searching many enrolled voices (one-to-many).
Speaker verification
Confirming a claimed identity from a voice (a one-to-one check).
Speech-to-text (ASR)
Converting spoken audio into written text, also called automatic speech recognition.
Structured data extraction
Pulling typed, validated fields out of a conversation for use in other systems.
Two-party consent
A recording rule requiring every party on a call to consent to being recorded.
Vishing (voice phishing)
Phone-based social engineering that tricks people into revealing information or money.
Voice biometrics
Recognizing or verifying a person from the unique characteristics of their voice.
Voice emotion recognition
Inferring emotion from how something was said, using the audio rather than the text.
Voiceprint
A mathematical representation of a voice used to recognize a speaker.
