The war for voice control dominance continues at pace and as adoption rates continue to rise, a plethora of opportunities and threats are opening up across nearly every business and consumer touchpoint. The ultimate goal for the world’s largest technology brands is to have their Virtual Personal Assistant (VPA) in your home, phone, car and office, acting as both butler and gatekeeper to as many daily tasks as possible. VPAs are appearing at both work and home. The enterprise sector is growing steadily, with both Gartner and Tracitica predicting VPAs in the workplace to account for a quarter of all use by 2021, but it’s the battle for the consumer where the real spoils are to be had.
The brands that succeed in getting their VPAs into our lives will control the next generation of digital engagement. Those that fail will be locked out of an ecosystem where services are dictated by only a handful of players. The stakes couldn’t be higher.
The belligerents
There are countless players in the market vying for supremacy, spanning small valley startups like SapientX and Soundhound, to industry heavyweights such as Apple, Google and Samsung. Some have developed VPAs with a view to push existing services — as Amazon has with Alexa, others have focused on specific sectors, think Nuance with automotive.
While the smaller players will actively try and reach a degree of ubiquity, it’s fair to assume that they will follow the standard startup model - Either be swallowed up by the larger players as soon as they hit a certain number of users, land enough partnerships with massive blue chips, or create a must-have piece of unique IP.
For Apple and Samsung, their respective VPA’s Siri and Bixby are a driver to add value to the hardware they push and in Apple’s case it’s also its music streaming platform which is duking it out with Spotify. Amazon wants to sell you stuff. Lots of stuff. And they want Alexa to do the selling. Microsoft’s Cortana is, well it’s there, but beyond integration with a handful of smart speakers there aren’t showing any signs of a clear route to market. When Javier Soltero left the Cortana team late in 2018, it signalled a climb down from Microsoft. I doubt we’ll see Cortana gain any significant market share in the future.
On the automotive side; the larger car makers are frantically developing or improving the third party VPAs they rely on. Mercedes MBUX made a splash last year and has been widely praised. Not to be outdone BMW followed with their own assistant and have stated that in the future they want theirs to not only be in the car but also on your phone and in your home. This push by the OEMs to develop their own systems is driven by an innate fear of Apple and Google, who have made no secret of their desire to own the car VPA experience through their own platforms.
Finally it’s worth highlighting Facebook’s presence in the VPA space, or more accurately it’s absence. Back in the summer of 2015 Facebook announced M, a VPA that was to live inside it’s wildly successful messaging platform, however after a trial with 10,000 users it was shuttered early 2018.
It felt like a surprising move at the time. Considering Facebook’s dominance of the social graph and clear desire to be at the front of emerging technology (Think VR and Oculus), I’d have thought they would have pushed on with M, even if it Zuckerberg himself claimed he wasn’t entirely sure of it’s purpose.
The next technical hurdle
Relentless investment in VPA technology is fueling growth, however in technical terms the VPA market is still in its infancy. The early days of VPAs spawned many a viral video of the systems failing to recognise accents or commands, often to comic effect, but in the last two years the systems have matured to the point that it’s generally considered the actual voice and word recognition is a solved problem. The next challenge to overcome is understanding and executing the intent of the user. In essence: what is the meaning behind the command or question? If I ask my car:
“Play me some Metallica, actually perhaps Iron Maiden” Can the system recognise that the word actually is me changing my intended choice of music?
Now throw in command stacking:
“Play me some Metallica, actually play Brave New World by Iron Maiden, and text my husband telling him I’m playing his favourite album” and the system needs to not only spot the change of choice, but understand that the word album is a word to be included in a text, and not a queue to make another album choice. Some brands such as Mercedes have made headway here, with their VPA handling stacked commands. Tesla in contrast still can’t manage this, neither can Alexa, ask:
“In ten minutes time start playing Run to the Hills”, it’ll just start playing the song, despite the fact there is a timer skill included as standard. Advancements in AI and neural networking, combined with natural language processing will see these barriers removed in coming years, with the computational heft required being provided by cloud computers. This all poses an existential threat to the auto makers VPA offerings, as they fight to keep pace with the larger tech vendors.
Three’s a crowd
Over the next five years the field of VPAs will thin out as the number of consumers using VPAs increases. This poses the question, How many VPAs will we want to interact with? One for home, one for car, one for phone? Or will we opt for one for everything?
My money’s on a single VPA being the standard choice for most consumers. If every VPA offered the same set of services and matured at the same rate then in theory all I’ve got to do is change the wake word on each to the same name, however there’s a clear reluctance from manufacturers to allow this. Mercedes and BMW are sticking to their brand names as wake words. Some consumer devices like Sonos, for example, are already set up to use multiple VPA suppliers based on user preference, but only have Alexa installed at present.
Maybe our VPAs will all just get along? You can comically set them up on infinite loops, and some experts claim VPA backed service partnerships such as Amazon Prime arriving on Apple TV and Apple Music being available on the Echo Dot could be read as a sign of a connected future. I’m not sure though, as while this may look like progress my guess is that this is an uncomfortable short term move for the two behemoths to stave off a greater enemy.
One VPA to rule them all
The strength of the VPA is not only reliant on how well it understands you and your intent, it also has to have a body of actions and services it can provide that make the VPA relevant in the first instance. This is where the likes of Google has a huge advantage over the competition. They own the email, search, calendar and mapping systems used by the majority of us, and so can create fast and reliable connections to services we have already signed up to. In short they hold a decks worth of cards when it comes to functionality, and this home advantage will allow them to bolster VPA utility as the intent aspect of the technology is improved.
It’s worth remembering that in the race to catch Apple in the smartphone space they open sourced Android allowing any manufacturer to build phones using it. Then as market share grew, they slowly stopped updating the core apps for the open source market — choking these third parties as they geared up to launch their own phones. Now firmly making hardware in the form of the Pixel and Google Home, I’d expect Google to be become increasingly selective as to which other VPA providers they provide services to as their own offering improves.
A VPA Future
For all Google’s power, I believe VPAs powered by white labeled providers offer a unique opportunity for brands to differentiate themselves, creating avatars and personalities that directly engage with their users. Take Nomi, the cute tamagotchi like character that lives on the dashboard of the Nio. She offers a soft edge to a new brand, becoming an instant differentiator in the rapidly crowding EV market. AI driven chatbots online and in mobile apps are rising in popularity; Compare the Market’s Auto Sergei is now front and centre in their ad campaigns.
Antitrust laws should keep Google from sweeping up all of the smaller third party VPA providers, and with Mycroft, the first open source VPA platform gaining interest from developers I hope there will continue to be options for brands of all sizes to get to create their own bespoke VPAs.
Smart speakers don’t offer visual interfaces, and so while Alexa is queen of my Echo, I’m free to download and interact with any apps on my phone. It’s not hard to imagine avatar driven VPAs living in branded apps.
Finally, with HMI IDEs such as Kanzi and Altia offering ever improving 3D graphics support we’ll soon have the ability to create real time 3D VPAs, opening up a whole new world of creativity to interface designers. It couldn’t be a more exciting time to be on the front-line of this transformative technology.
