Amazon Polly Adds Custom Voice
Is this the beginning of branded voice as a standard feature for voice assistants? (We hope so!)
About two years ago, we started our first experiments with building synthetic voices. One of the first custom voices we made was a “southern” voice for a company that wanted to pitch a custom voice skill to KFC. Here’s what we came up with:
That deal didn’t work out, but it set us on our path to building tools for custom voice in mobile apps. In our prior incarnation, we made smart speaker apps like The Bartender. After we created the southern voice, we started experimenting with adding custom voices to Alexa skills. We used that Col. Sanders voice in beta skill version of the Bartender and it was great! However, for technical reasons with how Alexa skills work, it wasn’t realistic to think we could commercially release the custom voice in The Bartender. That was one of many reasons we started looking outside of Alexa and the smart speaker market to build VUIs. More on that in another post.
It is great news that Amazon would like brands to build and deploy branded voices in skills. We hope they will push brands to make brand-enhancing and identifiable voices, and make it easier for developers to add custom voices in their skills. It will be a much better user experience and more incentive for brands to invest in voice.
Shameless plug: If you would like to talk to Spokestack about helping your brand have a custom voice and NLU where you can actually own the conversation with your customer (instead of Alexa), email me.
OK, here’s our favorite voicefirst reads today:
Amazon is helping brands build custom text-to-speech voices for their Alexa skills - by Jay Peters for the The Verge
Why All the Smart Home Platforms Suck and How They're Going to Fix Them - by David Nield for Gizmodo
Spoiler alert! It’s open protocols. See yesterday’s post.
ML-fairness-gym: A Tool for Exploring Long-Term Impacts of Machine Learning Systems by Hansa Srinivasan, Software Engineer, Google Research
This recommendation came from Will Rice, our data scientist.
Finally, Elizabeth Ropp, our designer and product marketing lead, pulled up this amazing presentation by Karen Kaushansky on designing the future with Voice. This is an absolute must-watch presentation from 2018’s Front Conference.
This is issue #5 of SpokeDaily, a daily round of #voicefirst reads that Spokestack is talking about today. Let us know what you think by leaving a comment or hit us up on Twitter!