How To Hack Voice Assistants for The Music Industry

Today, Apple unveiled their smart speaker HomePod. While the tech specs are impressive for a home speaker system, I think the underlying voice technology is the long-term takeaway.

The technology itself isn't new - Siri was released 6 years ago and smart speakers were quick to follow - but Apple is notorious for mastering promising technologies changing mainstream customer behavior, especially in music.

Voice is the future

Google recently reported that one in five search inquiries are made with voice, and experts only expect that number to rise.

"Before AR and VR and AI, audio is going to be the next major platform shift for consumer attention. It’s here today! What are you going do?" -Gary Vaynerchuck

Statista recently reported that playing a song is the second most popular command for Amazon Echo, and the most repeatedly used command. This tells us that people are comfortable searching by voice, and the software is good *enough* to use over and over.

In fact, streaming music is the third most popular use for all smart speakers in the US, after asking general questions and weather inquiries.

At the WWDC keynote this morning, Apple's Phil Schiller said Sonos aren't smart speakers and Amazon Echos aren't good speakers. Here's how he predicts music commands will work on the HomePod.

This presents a new challenge for content providers, as consumers are shifting into more searches for genres, moods, and time periods than specific songs or artists. Search is moving in the direction of "deep learning" and contextual information ("play more songs by this drummer" or "play songs like One Dance by Drake"), and music marketers need to be ready with new SEO and machine learning strategies for voice.

HomePod runs on Siri, the same way Amazon Echo is built on Alexa skills and Google Home has Google Assistant. There are also a number of smaller voice-powered apps in the web store. Luckily these programs need our help to become even smarter.

What can my band do?

First, go download all the voice-search apps you can find, and experiment by commanding your band name, albums, and songs. The more requests you give, the better the machines learn. This may seem tedious, but there's a great chance companies like Spotify, Google and Apple will begin acquiring these apps to make their own software more powerful.

Here's how to "hack" each personal assistant with your music:

Melody - Voice Assistant for Spotify & iHeartRadio
Download the mobile app and search for your songs [hint: it must be on Spotify or iHeartRadio]. If Melody doesn't correctly find the name, ask Melody to "Learn [artist name]" and type in the correct text. Next time, just say "Play [artist name]" to play it.

Hound (SoundHound)
Houndify is doing some incredibly things with audio and artificial intelligence. Download the Hound app and command your music. If it searches incorrectly, you can report "Incorrect Search Results" and fill out a detailed description to its developers. You're not being a nuisance; they welcome your feedback!

(Developers can beta test new features here)

Google Assistant
Google Assistant is an evolution of Google Now, the company's first crack at a virtual assistant. Google Assistant is the backbone of Google Home, as well as Android wearable devices, cars, and other smart home devices. Search works incredibly quickly (and accurately), but can still get it wrong. You can download the mobile app and provide feedback for incorrect searches results.

(Developers can download the Google Assistant SDK here)

Cortana (Microsoft)
Cortana is Microsoft's voice assistant for Xbox One, Groove Music, Windows Phone, Windows 10, and Windows Mixed Reality. There is an app in the Apple/Android stores, and they allow feedback in the form of suggesting a new idea or reporting an issue on their mobile forum.

(Developers can use the Cortana Skills Kit here)

Siri (Apple)
The basis for the iPhone, Macs, Apple TV, and now HomePod. Unfortunately, you can only "teach" Siri your contact name pronunciations.

(Developers can use SiriKit here)

Viv (Samsung)
Created by Siri developers, Viv is a more open, intelligent platform. It is now owned by Samsung, and becoming less "open" to outside developers. There is currently no public app, but you can sign up to be notified if/when it's available.

Bixby (Samsung)
Bixby is an evolution of S Voice, Samsung's previous virtual assistant. Bixby still has trouble with the English language, but will soon be available on all Samsung Galaxy phones and smart home devices (including air conditioners, televisions, and refrigerators). It's unclear if Bixby will use technology from its new acquisition Viv. There is no mobile app and they have yet to release a third-party developer kit, but Samsung Galaxy S8 users can educate the technology using trial and error on their device.

Alexa (Amazon)
This one requires some coding knowledge and creativity, but don't let that deter you. Amazon's Alexa is built almost entirely off third-party skills, and it is learning quickly (see below). The Echo is still the leading smart speaker on the market, and works with Amazon Music, Spotify, and all the others.

(Developers can create custom Alexa Skills here)

EDIT: There will never be a substitute for making undeniably great music and building a strong fanbase around it. People will probably never stop searching for "Bohemian Rhapsody" or "Let it Be," whether it's through voice command, text search, or flipping through a jukebox. However, for those technologically inclined, the time to get started on voice is now!