Voice AI glossary

An AI voice agent is software that makes and answers phone calls, understands what the caller says, and replies in a natural spoken voice in real time.

Conversational AI is technology that lets a computer hold a multi-turn dialogue with a person in natural language, keeping track of context across turns.

Barge-in is the ability of a voice system to stop speaking immediately when the caller starts talking over it.

Turn-taking is how a voice system decides when the caller has finished speaking and it is its turn to reply.

ASR, or automatic speech recognition, is the technology that converts spoken audio into text.

TTS, or text to speech, is the technology that converts written text into spoken audio.

An LLM agent is a system built on a large language model that can take actions — calling tools, querying data, updating records — rather than only producing text.

RAG is a technique where a language model retrieves relevant documents from a knowledge base and uses them to ground its answer, instead of relying only on what it memorised during training.

A webhook is an automated message sent from one system to another when a specific event occurs, allowing software to react in real time.

CRM integration is the connection between a calling system and a customer relationship management platform, so call data flows into customer records automatically.

Voice cloning is generating synthetic speech that reproduces the sound of a specific person's voice from a sample of their recorded audio.

Latency, in voice AI, is the delay between a caller finishing speaking and the agent beginning to reply.

Lead response time is the elapsed time between a prospect submitting an enquiry and a business making first contact.

Connect rate is the percentage of outbound call attempts that result in a live conversation with the intended person.

Containment rate is the percentage of calls an automated system resolves fully without transferring to a human.

Call deflection is diverting an incoming call away from a human agent to an automated channel that resolves it.

First call resolution is the percentage of customer issues fully resolved during the first call, with no follow-up needed.

AHT, or average handle time, is the mean duration of a customer interaction including talk time, hold time and after-call work.

CSAT is a measure of how satisfied customers are with an interaction, usually collected as a rating immediately afterwards.

Sentiment analysis is the automatic detection of emotional tone — positive, negative or neutral — in text or speech.

IVR is an automated phone system that plays recorded prompts and routes callers based on keypad presses or simple spoken commands.

An outbound dialer is a system that places calls automatically from a contact list and connects answered calls to an agent.

A DID number is a dedicated phone number that routes directly to a specific destination without going through a switchboard.

A SIP trunk is a virtual connection that carries phone calls over the internet instead of over traditional physical phone lines.

Code-switching is alternating between two languages within a single conversation, often within a single sentence.

Tenglish is the mixed Telugu-English register used in everyday speech across Telangana and Andhra Pradesh, where English words are embedded inside Telugu sentence structure.

DLT registration is the mandatory Indian regulatory process, run on a blockchain-based Distributed Ledger Technology platform, that businesses must complete before sending commercial SMS or making commercial voice calls.

TRAI regulations are the rules issued by the Telecom Regulatory Authority of India governing telecom services, including how businesses may contact consumers commercially.

The DNC registry is the Indian list of phone numbers whose owners have opted out of receiving unsolicited commercial communication.