Reference

The conversation analytics glossary

A–Z reference. Each term defined in ≤60 words, with a link to the deeper page.

A

ACD (Automatic Call Distributor) — the system that routes incoming calls to available agents based on rules (skills, priority, wait time).

Agent assist — AI that suggests responses or actions to agents during live calls. Distinct from post-call analytics like InOro, which analyzes calls after they end.

AHT (Average Handle Time) — talk + hold + after-call work time; average across a set of calls.

AML flag — regulatory indicator triggered by patterns suggesting anti-money-laundering risk in a conversation.

ASR / STT — Automatic Speech Recognition / Speech-to-Text: converts audio to text.

Auto-QA — automated quality assurance: AI scores every call against a scorecard instead of a human reviewing a sample.

Abandonment rate — share of calls where the caller hung up before reaching an agent.

C

Call center analytics — measurement and reporting for a call center; can include phone-system, survey, and conversation-content metrics.

Call transcription — converting call audio into text, often with speaker separation.

Compliance monitoring — detection of regulatory or procedural violations in customer interactions.

Contact center — a broader term than call center; may include voice, email, chat and messaging.

Conversation analytics — analysis of customer conversations to extract topics, sentiment, quality signals.

Conversation intelligence — the AI-powered business layer built on top of speech analytics; scores, insights, actions.

CSAT — Customer Satisfaction Score, usually from a post-call survey ("How would you rate this call?").

CES — Customer Effort Score, measuring how easy it was to get a resolution.

Churn signal — language or behavior in a call suggesting the customer may leave.

D

Diarization — separating audio by speaker (agent vs customer).

F

FCR (First Contact Resolution) — share of issues resolved on the first customer contact.

I

Interaction analytics — analytics across customer interactions; often used interchangeably with conversation analytics.

K

Keyword spotting — older speech-analytics approach that flags predefined phrases; contrasted with contextual LLM analysis.

N

NPS (Net Promoter Score) — a customer loyalty index based on the "would you recommend" question.

O

Occupancy — the share of an agent's logged-in time spent on calls or wrap-up.

On-premise — deployment model where all software and data live on the customer's own infrastructure.

P

PII anonymization — masking of personal information (national IDs, card numbers) in transcripts and analytics.

Q

QA scorecard — a defined set of criteria and weights for evaluating a call's quality.

S

Sentiment analysis — measurement of emotional tone in a conversation, usually per moment or per topic.

Service level — the share of calls answered within a target time (commonly 80/20 = 80% within 20 seconds).

Speech analytics — the audio-to-data layer: transcription plus structured analysis (topics, sentiment, phrases).

V

VoC (Voice of the Customer) — customer feedback captured across channels, aggregated for insight.

W

White-label analytics — the vendor's product delivered under the reseller's brand; standard for BPO conversation-analytics offerings.

See these terms applied to your own calls

30-minute demo. 30-day pilot.