An agent is the script, the voice and the behaviour of the AI on a call. You control the greeting, the questions it asks, what counts as a qualified lead, and what it should do when someone asks something it does not know.
Disclosure is enforced, not configured
Every agent states that it is an AI at the start of every call. This is not a setting, and there is no way to turn it off.
The reason is regulatory rather than editorial. The FCC's February 2024 ruling treats AI-generated voices as "artificial" under the TCPA, and the consequence is a private right of action with statutory damages per call. A platform that let you switch disclosure off would be selling you a liability.
The disclosure is recorded as its own event on the call, so a call log can show that it was given rather than merely that it was configured.
Knowledge base
An agent can answer from a knowledge base you provide. Give it the questions your customers actually ask, in the words they use.
What to write, and what not to
Write the agent as you would brief a new member of staff: what you sell, who you are calling, what a good outcome looks like. Do not write instructions that ask the agent to obscure that it is AI, imply endorsement you do not have, or promise terms you cannot honour.