Last month, Sonos started beta testing Google Assistant support for its Sonos One and Sonos Beam speakers. While both devices already work with Amazon Alexa voice commands, adding Google Assistant is without precedent. With every other smart speaker–whether it’s made by Google, Amazon, or a third party–you’re locked into one assistant or the other.
Still, the feature in its current form has one major limitation: There’s no way to use both assistants interchangeably by saying either “Alexa” or “Hey Google.” Instead, you must assign a single assistant to a speaker, and use the Sonos app to switch between them.
Being able to use both assistants without a cumbersome switching process would have some clear benefits: You could shop on Amazon without giving up Google Assistant’s superior search capabilities, or tap into Amazon’s larger number of third-party skills without losing voice control over Chromecast devices. You could also listen to Amazon Music and YouTube Music on the same device, and broadcast one-way messages to other speakers around the house, regardless of whether they’re Amazon Echos or Google Homes.
So why isn’t Sonos allowing both voice assistants to run simultaneously? And for that matter, why aren’t more companies aren’t trying to support multiple assistants on their voice-enabled products? It’s hard to get a straightforward answer to either question, but here’s what I’ve come up with after asking Sonos and others:
Explanation 1: Users might get confused
Lidiane Jones, Sonos’ vice president of software product, says Sonos started with one assistant at a time to avoid potential confusion. If you set a timer, for instance, you might not remember whether to invoke Alexa or Google when the alarm starts ringing.
“When they are interchangeable, it is more complex for customers to remember which assistant you asked for what,” she explains. “For us that are overly techie, we might feel like, ‘Oh of course, it makes sense.’ But when you put it in homes where people are not, it gets overwhelming.”
Jones says Sonos isn’t wedded to the one-assistant-per-speaker approach, though. One reason the company is still beta testing Google Assistant support is to figure out whether people want to use both at the same time.
“There are all sorts of different ways that people are using multiple [assistants] at home already, so we’re trying to learn from that,” she says. “We want a lot of customer feedback to help us determine the right path.”
Explanation 2: There are minor technical hurdles
Jones and other experts I’ve spoken to say that listening for two wake phrases–one for Alexa, one for Google Assistant–is not technically difficult. Doing so may require more memory and processing power, but smart speakers like the Sonos One already have enough of both.
Still, each assistant does have its own system for handling certain commands, and supporting both on the same device is at least a small technical obstacle.
“The biggest challenge is actually keeping track of things like timers and alarms, because you do need to keep internal state of those in case there’s a Wi-Fi outage and things like that,” says Jones.
Explanation 3: There are other, unrelated holdups
Sonos isn’t the only company that wants to integrate its products with multiple assistants. Last week, a startup called Orro started selling a connected light switch with a built-in speaker and microphones, and has promised the ability to speak with either Amazon Alexa or Google Assistant in the future.
But while Orro is already beta testing Alexa support, the company doesn’t have a timeline for adding Google Assistant. That’s because Google doesn’t yet support certain aspects of Orro’s hardware at a technical level, says CEO Colin Billings.
“Their program is just less mature than Alexa’s third-party program, and so they are still working things out in some of the technical aspects of how they want to work with third parties,” Billings says.
Headphone makers are also interested in supporting multiple voice assistants with different wake words, so users don’t have to press any buttons or switch between assistants in an app. Peter Hartmann, the senior director of consumer partnerships and strategic alliances for headphone maker Jabra, says the company is still waiting for new Qualcomm chips that can support wake words in the first place. Jabra had originally planned to support “Alexa” and “Hey Google” voice commands on its upcoming Elite 85h headphones, which it announced at the CES trade show last month, but has now pushed that feature back to a future product.
“Most of us in the Bluetooth industry are using chips from Qualcomm, and honestly we have not been able to find a decent way to implement a headphone that’s always listening without having a severe negative consequence on power,” Hartmann says.
Wild speculation: Maybe the tech giants don’t want it
Last May, The Information’s Aaron Tilley reported that Google didn’t want to share space with Alexa on Sonos speakers in the first place. But Sonos holds several key home audio patents, which helped it strongarm Google into getting on board as an alternative to lawsuits or a licensing deal. Variety’s Janko Roettgers also reported last year that Google didn’t want its assistant to be non-exclusive on smart speakers, but that Sonos “got the tech giants to play nice by flexing its patent muscles.”
This is speculation on my part, but I wonder if Google is still resisting the idea of using Alexa and Google Assistant simultaneously, and Sonos’ current beta test is about determining whether to push Google even further. If the company can demonstrate that customers want easy access to both assistants, it might more willing to flex its patent muscles again.
When I asked a Google representative if the company has any policies or guidelines precluding device makers from supporting multiple voice assistants, the rep declined to give a statement on the record, but pointed out potential confusion and technical hurdles, and said that Google was still working with brands like Sonos to gather feedback.
Amazon, meanwhile, provided the following statement when asked the same question:
No. We believe in providing choice and flexibility for customers, and we’ve actively encouraged projects that allow Alexa to function seamlessly alongside other voice services. There will be multiple successful voice services in the future, each with its own set of skills and capabilities, and we believe device makers should make it as easy as possible for customers to access the service they’d like for a given task.
Amazon also noted that Alexa already co-exists with other voice assistants today. Facebook’s Portal smart display, for instance, responds to both “Alexa” for general queries and “Hey Portal” for features specific to that device. French telecom company Orange is working on a smart speaker that works with both Alexa and its own Djingo assistant. Alexa and Microsoft’s Cortana can also talk to one another, albeit in a somewhat cumbersome way.
As for Sonos, the company provided a non-answer when I asked if it alone would have the authority to support Alexa and Google Assistant simultaneously: “[W]e have great relationships with both Google and Amazon and are in regular communication with our partners there. We’ve always been clear from day one that we planned for our products to support multiple voice assistants.”
For what it’s worth, Jabra’s Peter Hartmann says he’s never had a conversation with either company that made him believe multiple assistants wouldn’t be supported on headphones. But when I float my theory by Orro’s Colin Billings, he doesn’t shoot it down.
“I think that that’s a fair read, from what I also know,” he says. “I think you’re seeing that correctly.”
There may also be a more charitable explanation: Coexistence just isn’t a priority for Google and Amazon, even if they’re not actively opposed to it. When a product gets certified for Google Assistant or Alexa, device makers have to specify how those products will implement voice commands, and peaceful interoperability just isn’t something that the tech giants are factoring into their certification.
“Those conversations are generally focused around exclusive operation at any given time,” says Billings.
I don’t think the one-assistant-per-speaker paradigm will last forever. Having to use multiple smart speakers is already a pain, but as Amazon and Google push device makers to built voice control into other products like headphones and light switches, the pressure to support both at the same time will only increase.
As for the present, Sonos’ Lidiane Jones says things could go either way.
“We’ve gotten quite a bit of mixed feedback,” she says, “so we’ll see.”