Skip to main content
Configure the entry point that greets the caller, establishes initial context, and begins the first conversational turn.

​
Prerequisites

  • A workflow draft.
  • An approved AI disclosure and an optional greeting style.

​
Configure the entry point

  1. Add or open the Start Call node.
  2. Choose the required AI disclosure behavior and write a short opening prompt.
  3. Select Text (TTS) or Pre-recorded Audio for an optional greeting.
  4. Enable interruption or delayed start only when the call experience needs it.
  5. Add tools, documents, or a pre-call fetch when the opening turn needs them.

​
Constraints and failure states

Every workflow needs exactly one entry node, and it cannot have an incoming edge. Text greetings are not supported with realtime speech-to-speech models. Delayed start accepts a short wait between 0.1 and 10 seconds. A pre-call fetch that fails contributes no variables; see Pre-call data fetch.

​
Protect your data

The first prompt can be spoken aloud and recorded. Do not put secrets, private URLs, or unnecessary personal data in the greeting or disclosure.

Related topics

Pre-call data fetchWorkflow fundamentalsTurn interruption settingsPre-recorded audioEditing a workflow