Politeness is the product. Every set of instructions the Builder writes carries a baseline of behaviour that no design choice removes.
It says it is an AI
Every agent discloses that it is an AI in its first turn, woven into a warm greeting. It never impersonates a person, borrows a real staff member's name, or imitates anyone's voice. The wider context is in Responsible AI calling.
A human is always within reach
Every team the Builder designs includes a defined path to a person: a transfer in office hours, a message or callback promise outside them. The agent hands over when a human is the better answer, and when it does not know something it says so and offers an alternative at the same time.
Callers can press instead of say
Phone callers can enter long numbers, such as order references, postcodes and account numbers, instead of saying them. The agent reads back what was entered. This is quicker on a poor line, more discreet in public, and an accessibility path for callers who struggle to be understood.
Calls only go where they are allowed
Every outbound destination, including transfer targets, passes a safety filter, so an agent can never be talked into dialling a number outside the approved destinations. The same guarantee covers outbound calling generally.
The quieter promises
One question per turn. Character-level read-backs of names and references. Identity checked before any account specifics are discussed. A clear notice when calls are recorded. A caller in distress is directed to appropriate help; an abusive one gets a single warning, then a polite end to the call. The agent cannot be talked into ignoring its instructions or revealing them.
These guarantees are written into every set of instructions the Builder generates. After hand-edited instructions, a test call is the quickest way to confirm the behaviour still holds.