As one of our mobility analysts at Blue Hill, I get (perhaps unduly) excited over any shiny new gadget, even though I’ve recently been disappointed by what seems like a lack of innovation in the mobile market (cue the “I miss the old Apple” rant). That’s why I was excited, but skeptical, to cover Google’s October 4th product announcement of its long anticipated “first” smartphone. But it’s really not the hardware Google is focusing on with its most recent product launch, though it could certainly seem that way from its unveiling of the Pixel (the first Google branded smartphone), the Daydream View virtual reality (VR) headset, and the Google Home smart hub.
No, it hasn’t been about hardware for the past few years for mobile device manufacturers it certainly seems, with only small incremental changes being made to hardware (or, in the case of Apple’s removal of the headphone jack, gigantic leaps and bounds of courage … but that’s another story). Most of the focus of smartphone innovation instead has been coming from within: with the software, and most recently, artificial intelligence.
Pixel is the first phone with Google Assistant built in. Google Assistant is Google’s artificial intelligence agent that can respond to natural language and text queries to complete tasks on the phone within both native and third-party apps, as well as return information that would normally require a manual internet search. But wait – I thought Google already had a voice assistant, Google Now? Yes… but Google Assistant is Google’s attempt to more fully integrate hardware and software to create a complete product ecosystem, not just a way to talk to your phone.
Google Now works on the Google app, meaning it is available on both Android and iOS devices, whereas Google Assistant is natively built into Google’s hardware: the Pixel phone and the Google Home hub. Google Now has more limited capabilities and is more of a one-way conversation; Google Assistant is designed to offer a two-way conversation between the user and the intelligent agent, be able to recall information from previous conversations and personal details about the user, and get smarter while it is used, with Google hoping that eventually Google Assistant will be able to take over more complex tasks for users. For now, the two assistants will both exist separately, although in the Pixel and Google Home, Google Assistant will replace Google Now for natural language interaction.
Google is extending its integrated hardware and software ecosystem beyond just the phone with the introduction of Google Home, a voice-activated speaker and smart home hub. With Google Home, users can interact with the device using natural language to complete a variety of tasks, including those that involve third-party applications and even physical products like lightbulbs and thermostats. Google accomplishes this through APIs and compatibility with home automation systems like Philips Hue and Nest. Google is expanding third-party app and device compatibility for Google Home in the near future.
To create this ecosystem of automation and artificial intelligence, it makes sense for Google to take control of hardware development in order to integrate software and hardware more seamlessly. In terms of software, Google has a leg up in creating a virtual assistant with access to Google’s vast data sources. Google has been pioneering search since the early 2000s and has invested heavily in deep learning and artificial intelligence. As a whole, Google has traditionally been much more of a data and software focused company than a hardware player, and even its most recent moves into hardware have an underlying software goal.
Speaking of hardware manufacturers … Samsung Electronics has also been making moves into artificial intelligence. Samsung announced on October 6 that it has acquired Viv Labs, a U.S. startup whose founders include a former Apple employee who co-created Apple’s Siri virtual assistant. Viv Labs has created an open artificial intelligence development platform for building virtual assistants and integrating natural language interaction into third-party applications and services. The platform is scalable and learns as it is used, to deliver personalized and contextual assistance to users.
With the acquisition of Viv, Samsung is hoping to create an AI-based open ecosystem across all its devices and services, including mobile devices, wearables, and home appliances. Viv will continue to operate independently under existing leadership but will work collaboratively with Samsung’s Mobile Communications business.
Google and Samsung have officially entered the fray. And it’s a fray alright. Apple continues to build out HomeKit, its smart home automation platform that connects smart devices and allows users to control them through the iPhone and its virtual assistant, Siri, announcing expanded third-party support in June. Amazon opened Alexa Voice Service, its cloud-based virtual assistant that powers its Echo home automation device, to hardware developers in June 2015, to allow connected devices to integrate with Echo through free APIs. Microsoft’s virtual assistant, Cortana, is built into its computers and mainly focuses on Microsoft’s own line of Office products. With each of these virtual assistants, their software is not supported by other manufacturers’ hardware – Alexa won’t be on the Google Home, nor Cortana on HomeKit, nor Siri on the Echo, and so forth. All these separate hardware and software platforms are creating distinct product ecosystems in which consumers become entrenched once they invest, leading to a highly fragmented market.
Right now the greatest value artificial intelligence can provide comes from actually knowing about the user, not just how to search the internet for the weather, or navigate to apps to play a song, or perform basic calculations that probably save me about 30 seconds of time. The best outcome would be that the intelligent agent could plan and make decisions for me, and perform tasks without my even having to say anything. To do that, it needs a lot of data – not just from my interactions with the assistant but also about how I use other applications and devices.
Integrating data from third-party apps and devices requires APIs, and as the market currently stands, competition is creating distinct platforms, minimal compatibility, and no common standard. Competitors aren’t looking to share their apps and devices either (Amazon wants you to use Amazon music on the Echo and not on the Google Home, and Apple Music doesn’t want you to use it on either). If most of the value of a virtual assistant is in automating quick tasks like playing a song, searching the web, or navigating within an app, that benefit is quickly lost if your preferred apps aren’t compatible. It’s leading to an arms race of sorts for the companies creating virtual assistants to snatch up API agreements and have the most third-party compatibility, but it’s resulting in slow consumer adoption.
Not only is consumer adoption slow, enterprise adoption is relatively non-existent. As the market currently stands, Microsoft has the greatest chance of success from an enterprise perspective due to Cortana’s integration with Microsoft’s line of Office applications, as well as Windows’ relatively strong market share in enterprise computer operating systems. It’s likely that Apple and Samsung will try to catch up in the enterprise market, as both have aggressively targeted businesses as a secondary smartphone market in the past. Amazon and Google seem to be less focused on the enterprise market and are instead investing more heavily in creating consumer value.
For now, the artificial intelligence market is so fragmented that it’s hard to pick a clear winner, but each of the major players has its own relative advantages. As mentioned, Microsoft has an enterprise advantage with its Office applications integration, and the availability of its assistant on computers rather than mobile devices. Google has enormous volumes of data, both from its search engine and its line of email and messaging applications, and has invested heavily in cloud infrastructure – both of which are necessary to successfully implement AI. Amazon also has significant cloud infrastructure investments to support AI, and troves of consumer and product data from its ecommerce platform. Samsung has an advantage on the hardware side, given the breadth of its product portfolio in consumer electronics, to offer Viv in more smart home appliances and wearables than some of its competitors; this might give Samsung a better chance of realizing the smart home dream, though that’s certainly still far away. And Apple … well, Apple has the iPhone – an extremely successful product – at the heart of its AI ecosystem, as well as investments in smart home automation with HomeKit, and voice assistance with Siri on the iPhone and most recently on Apple’s Mac line of computers.
Blue Hill notes that right now, it’s up in the air which artificial intelligence platform will win in the market – if any. Companies and consumers will probably have to invest in multiple platforms before they are able to discern clear differentiators between them. For companies investing in this market, like Amazon, Apple, Google, Microsoft, and Samsung, the best chance of success comes from making virtual assistants ubiquitous through hardware integration. For that effort, Google’s recent move into hardware shows that companies that are seriously investing in artificial intelligence recognize the value of owning the hardware, too. That ultimately means that once there is a dominant player in artificial intelligence, that company will likely have major market share in hardware as well. The race begins…