Chapter 6
Embellishing Response Speech

If decades of science fiction have taught us anything, it’s that when computers or robots talk, they speak in a mechanical tone and staccato voice. Whether it’s the robot from Lost in Space or the lovable robot from Wall-E, most talking machines in science fiction do not speak in natural human tones.

In the real world, however, Alexa has a much more human voice than her sci-fi counterparts. She speaks almost as naturally as most any person you’ll meet (possibly better in some cases). Even so, despite her natural voice, Alexa doesn’t often speak with a great deal of variation or emotion. Her tone, although human-like, is at best very matter-of-fact, and at worst might even betray her as a machine.

We want the users of our skills to feel comfortable when they speak with Alexa. The more human she seems the more likely that they will continue to use our skills and continue to rely on Alexa as a relatable assistant that can help them with whatever they need.

In this chapter, we’re going to explore the Speech Synthesis Markup Language (SSML), using it to gain control over how Alexa speaks responses. We’ll use SSML to change the tone, volume, and pitch of Alexa’s voice, change how she pronounces words, and even completely change her voice altogether.

To start, let’s have a look at a simple SSML example and see how to test it with Alexa’s text-to-speech simulator.