The Platform PM

Glossary and sources

Field Guide

Voice AI: real-time voice agents

10

Glossary and sources

7 min read · Chapter 10 of 10

Words that mean something else here

TermWhat you'd assumeWhat it means here
LatencyOne numberAt least six: model inference, first audio, first token, end of turn, the platform's turn gap and mouth to ear. Ask which, and at which percentile
RealtimeInstantOpenAI's product name, any streaming speech-to-speech model, or loosely anything under a second
InterruptionThe caller cutting inEither direction: the caller interrupting the agent (wanted) or the agent cutting off the caller (not wanted); an "interrupt rate" can mean either
HandoffPassing to a personUsually a transfer, by SIP; in agent frameworks, a switch to another AI agent
AgentThe AIAn AI agent, a human contact center agent, or in SIP any endpoint. In "transfer to an agent", it's the human
MinuteSixty seconds of a callCarrier minutes, stream time with silence, audio tokens, or minutes of generated speech; "$0.05 a minute" can cover very different scope
ContainmentResolutionCalls not transferred to a person, abandoned calls included
Artificial voiceRobotic-soundingUnder the TCPA, any AI-generated or cloned voice, a live conversational agent included
Attestation AThe call is legitimateThe carrier vouches for the customer's right to the number, not the voice or the content
TranscriptWhat the model readIn a chained stack, yes; in speech-to-speech, a separate recognizer's guess at what was said

What surprised me

Placeholders in your voice, drafted from the research and the earlier guides. Rewrite each with your own moment.

The rate deck. In telco I could multiply minutes by a price per minute. Here one four-minute call is metered five ways by five vendors, and on a speech-to-speech model the fifth minute costs more than the first.

"Faster means better." The fastest agents in the benchmark still got about half the tasks wrong on phone audio. Speed got them into the conversation and logic decided the outcome.

"Inbound is the safe side." The TCPA barely touches calls the business answers, and then a California court treated the AI vendor on the line as an eavesdropper.

Caller ID. In the telco guide, attestation was a technical detail. In the Biden robocall it was the one penalty that got paid.

"The cheap part is the AI." The AI minute was a rounding error next to the human minutes that follow every call it can't finish.

Sources

Undated entries were read on October 8, 2026; "search result" means seen only as a search snippet or summary. Company figures are self-reported unless they come from a filing or a regulator, and vendor benchmarks measure the vendor's own products or customers.

Regulators, laws and courts

Labs and speech model companies: documentation and pricing

Telephony and voice agent platforms

Finished agents, contact centers and deals

Voice security and fraud

Benchmarks, research and outcomes

Field Guides are learning notes, not legal or compliance advice. Rules and fees change; check the cited primary sources before you act on anything here.