Skip to main content

Configure caller inputs for Actions in voice conversations

You can configure Actions in your AI Agent so it can make API calls. When you have Voice enabled, for each Action that requires inputs from the caller, you need to define methods for your AI Agent to collect those inputs. Different capture methods suit different types of information being captured, so configuring appropriate capture methods help set your AI Agent and your callers up for success.

Understand voice capture methods​

When you're configuring an Action, you can add inputs so your AI Agent can ask the caller for information required to make an API call. For each input you configure in your Action, in the Capture Modes list, you can choose up to two capture methods for your AI Agent to use.

There are three capture methods you can select:

  • Speech: Your AI Agent asks the caller to speak. This is the default method when you add a new input to an Action.

    note

    Your AI Agent will allow callers to interrupt it while it's capturing information, but otherwise won't stop talking if the caller starts talking at the same time.

  • SMS Capture: Your AI Agent sends the caller an SMS message and asks them to reply.

    Your AI Agent must get the caller's consent before sending SMS messages. For more information, see Learn about SMS capture consent below.

  • DTMF Capture: Your AI Agent asks the caller to enter a number using their phone's keypad. This method is only available for numeric inputs.

Combine voice capture methods​

If required, you can combine methods so your AI Agent uses speech as the default input, with SMS or keypad entry as an alternative. These are the available input combinations:

  • Speech and SMS Capture: Your AI Agent tries up to three times to capture the input from the caller verbally. If it still can't successfully transcribe what the caller said, it sends the caller an SMS message and asks the caller to send a reply instead.
  • Speech and DTMF Capture: Your AI Agent gives the caller the choice to either say or dial in the number.
Important

You can't combine all three capture methods, or SMS and DTMF captures, for the same input. However, you can select a variety of capture methods for different inputs in the same Action.

Learn about SMS capture consent​

Before sending any SMS messages to a caller, your AI Agent must obtain consent from the caller to send them. There are some important things to note about SMS consent:

  • This consent only applies to your ability to send an SMS to the caller; you can't also use this consent to add the caller's phone number to marketing lists or send them additional SMS messages for other purposes.
  • SMS only works for mobile devices. Callers may be using landline phones, or may not want to receive SMS messages, so be prepared for callers not to be able to use SMS capture in all cases.

The first time your AI Agent needs to send an SMS to the caller in a conversation, your AI Agent asks for their consent to do so:

  • If the caller says yes, your AI Agent sends them the SMS message, and sets the sms_consent variable to true. The variable remains set for the rest of the conversation, so if other inputs require additional SMS messages, your AI Agent won't have to ask for consent again.
  • If the caller says no, your AI Agent hands them off to a human agent immediately.

Have any questions? Contact your Ada team, or email us at .