Emotions in AI Calls with Outreach Master

Overview

Sonic-3, the speech model behind Outreach Master’s AI Calls, enables natural-sounding, emotionally expressive speech.
You can control emotions, volume, and speed directly via SSML tags in the transcript or through the user interface.
This makes your AI calls sound more human, context-aware, and authentic.

The parameters in Sonic-3 are interpreted as guidelines, not strict controls.
Experiment with different combinations to achieve the most natural and effective delivery.

Controlling Speed & Volume

You can adjust speech speed and volume directly in your text using SSML tags:

<speed ratio="1.5"/> I like to speak quickly because it makes me sound smart.

<volume ratio="1.5"/> And I can be loud, too!

Speed: Values between 0.6 and 1.5
Volume: Values between 0.5 and 2.0

Example: 1.5 corresponds to roughly 50% faster or louder than the default voice setting.

Controlling Emotions (Beta)

Sonic-3 automatically infers emotions from the text content.
However, for precise control, you can manually define an emotion using an SSML tag.

Example

<emotion value="angry" /> How dare you speak to me like I'm just a robot!

Supported Emotions

Sonic-3 supports a wide range of emotional tones — from neutral and calm to highly expressive delivery.

Primary Emotions

These provide the most consistent and natural results:

neutral
angry
excited
content
sad
scared

Extended Emotions

Sonic-3 can also recognize and render many subtle emotional variations:

happy, enthusiastic, elated, euphoric, triumphant, amazed,
surprised, flirtatious, joking/comedic, curious, peaceful, serene,
calm, grateful, affectionate, trust, sympathetic, anticipation,
mysterious, mad, outraged, frustrated, agitated, threatened,
disgusted, contempt, envious, sarcastic, ironic, dejected,
melancholic, disappointed, hurt, guilty, bored, tired,
rejected, nostalgic, wistful, apologetic, hesitant, insecure,
confused, resigned, anxious, panicked, alarmed, proud,
confident, distant, skeptical, contemplative, determined

These emotions influence tone, rhythm, and vocal energy.
Sonic-3 dynamically blends them with the meaning and emotional intent of the text for natural, expressive delivery.

Laughter (Nonverbalisms)

In addition to emotions, Sonic-3 supports nonverbal expressions such as laughter.
You can insert laughter simply by adding the [laughter] placeholder into your transcript:

That meeting was... unexpected. [laughter] Anyway, let's continue.

The model generates natural-sounding, context-appropriate laughter,
making your AI calls feel warmer, more engaging, and human-like.