How agents behave on a live call

Recording follows a defined consent state and does not start where it is not permitted.

• Each call tracks a recording state: not required, disclosure pending, consent requested, consent granted or declined, recording started, recording stopped.
• Recording does not proceed when platform controls determine it is not permitted.
• Customers configure recording for their use case and obtain the notices and consents their jurisdiction requires.

Requests to stop communications are honored and carried into future campaigns.

• Requests to stop communications must be honored.
• Agents are not configured to work around opt-out or do-not-contact requirements.
• Opt-out status is recorded against the contact so later outreach respects it.

Calls present accurate caller identity; origin is never misrepresented.

• Appropriate caller identification is used for every calling configuration.
• The identity or origin of a communication must not be misrepresented.
• Numbers and routing are set up with the telecommunications provider for your regions.

A caller who needs a person reaches one, with the conversation summary attached.

• Transfer rules are part of every call and chat flow.
• The person receiving the hand-off gets a summary of the conversation so far.
• Out-of-scope or sensitive requests are escalated rather than answered by the agent.

Agents identify themselves as automated where disclosure is required.

• Voice agents can state at the start of a call that the interaction is automated.
• Disclosure wording is set per market and reviewed with your team.
• Customers are responsible for meeting the disclosure rules that apply to them.

AI voice services are not a substitute for traditional emergency calling.

• Unless service documentation states otherwise, voice functionality does not replace emergency calling services.
• Agents are configured to direct callers to the appropriate emergency number.
• Voice availability can depend on third-party carriers and their coverage.