AI & call disclosure

Modular Voice combines speech recognition, reasoning and speech synthesis. Native Voice uses a real-time voice model with the shared business core. Both use the agent’s published configuration, business knowledge and action permissions.

The operator controls the greeting and should clearly identify the AI agent. Configure the appropriate disclosures and consent process for your callers, location and use case. The software does not establish consent merely by enabling a feature.

AI can misunderstand speech or produce incorrect answers. Knowledge retrieval and citations help operators inspect the source of an answer; they are not a guarantee of factual accuracy.

Confirmed requests can be saved for follow-up. Saving an appointment request does not reserve a calendar slot or confirm a booking. Connect and verify the appropriate external workflow before promising that outcome.

Configured routes can offer human handoff, including consultation and acceptance. A transfer request can fail or go unanswered; inspect its actual outcome.

Recording is optional and must be enabled in the published agent. Operators choose 7 or 30 days of audio retention and are responsible for appropriate notice and consent.