Streaming music service Pandora has launched a new voice mode as part of its mobile apps, allowing users to search for songs, podcasts and more with simple voice commands. The new feature is being made available to a small subset of Pandora’s users at launch, with plans to gradually bring it to all of the service’s 69 million monthly active users.
Pandora is launching voice control in part as a response to the growing popularity of smart speakers, explained chief product officer Chris Phillips in a recent interview with Variety. “Pandora is already used by millions of people on voice-enabled devices,” he said.
But it’s not just a growing acceptance of voice control. Smart speakers and better AI assistants are also changing the way consumers are interacting with voice control, period. Previously, consumers were using voice input for very specific requests. “It’s almost like talking to a robot,” said Phillips.
As voice technology advances, consumers are starting to expect technology to be more conversational, and talk to devices like they would talk to their friends. “More and more consumers are going to become more and more vague,” Phillips said.
Pandora is responding to this trend by combining natural language processing and machine learning it licensed from Soundhound with its own music genome data to match up even vague requests with the right soundtrack.
For instance, the service may deduct from previous listening that a user has kids, and play family-friendly tunes when he or she asks for a cooking playlist. Or it may respond with two completely different soundtracks to the very same request from two different users. “We are applying personalization to all the asks,” Phillips said.
Users who have been enrolled into the new voice mode can access it on their iOS or Android phone with the wake phrase “Hey Pandora.” It allows you to ask for the title of a song, direct Pandora to add a song to a playlist, or to play a specific playlist or podcast. Users can even thumb-up a song by saying “Pandora, I like this song.”
Pandora’s voice efforts are for now focused on mobile, but Phillips said that the company plans to bring as much of this functionality as possible to smart speakers as well. However, he admitted that it’s a bit harder for the company to innovate on third-party devices, as smart speaker makers only share limited data with service providers.
For instance, Pandora doesn’t know what happened if a smart speaker fails to recognize a voice command. However, being able to respond to failures is especially important to improve a voice-based service, and make sure that users aren’t forced to jump through extra hoops just to get to the result they want. Said Phillips: “Nobody wants to voice-babysit their speaker.”