This blog post is co-authored by VP and Principal Analyst Michael Facemire and Analyst Andrew Hogan at Forrester.
Google held an event in San Francisco this week to announce several new consumer products – a smartphone, a wireless speaker, Wi-Fi routers, a virtual reality headset and an updated Chromecast solution. All showcased an emerging strategic direction for Google and some killer engineering and design skills.
None of it impressed as much as the demos of Google Assistant – Google’s virtual assistant.
What is a virtual assistant, you ask? A virtual assistant is another name for an intelligent (personal) assistant. Virtual assistants orchestrate agents or services from third parties on behalf of consumers. Bots are one form of an agent. Virtual assistants rely on context (e.g., user input, localization capabilities, and access to information from a variety of data sources) to refine the quality of responses to a user’s requests. These assistants guess, but the guesses get better over time. “Virtual” implies that the service is digital and not performed by a human you’ve hired.
Google Assistant is a natural extension of Google’s path towards becoming the agent that sits between brands and their customers. The “holy grail” of becoming a consumer’s primary virtual assistant will be hard for Google to obtain, but holds unprecedented business value. Google is not alone in this race – Amazon, Apple and Facebook in the U.S. also have their sights set on being the trusted assistant for consumers.
Examples of intelligent assistants include Apple’s Siri, Amazon’s Alexa, Microsoft Cortana, Facebook M, and Google Assistant.These services will complete tasks proactively over time. For example, an assistant might create a calendar entry based on an email from a colleague or remind me when to leave for that meeting. These services will integrate third party services as well. Alexa leads the pack today with 1,900 integrations (referred to as “skills”). (Source: TechCrunch) Examples include Capital One, which allows customers to pay bills, and Campbell’s Soup, which provides recipes.
Google Assistant isn’t new, but the presence of Google Assistant on a range of consumer electronic devices built by Google from the ground up is new. By owning and controlling the entire technology stack, Google has control over the end user experience – an area where it hasn’t always excelled. Google Assistant MOSTLY integrates with Google services, as well as a hand-selected group of early partners (e.g., OpenTable, Uber).
What does this mean? Over the past eight years, consumers have been taught to download apps to get access to specific services like maps, email or music. Today, consumers choose one app to get access to content (e.g., sports scores) or a service (e.g., order take out) and hop in and out of dozens of apps each month to get everything done.
Apple, Facebook and Google have all been working hard to reduce the burden on consumers by bringing more functionality into a single platform (e.g., messaging) or the home screen. Google took a step forward on October 4th in showing off its artificial intelligence, machine learning, deep learning and natural language processing skills. Accessing Google Assistant via voice or text can control a limited number of connected devices in the home (e.g., Nest), access media and offer a rundown on an individual’s day.
For Google, the timing is good because:
- Consumers have adapted to paying $600 to $800 for smartphones.Carrier subsidies for mobile phones have faded away, and consumers must adjust to paying full price for devices even if they pay for them monthly along with the service plan.
- Amazon has taught consumers what a “virtual assistant” is.Consumers are beginning to understand the value proposition. Amazon has sold upwards of 4-5 million Echo devices and has taught consumers what a virtual assistant is. And, there’s already wide adoption of Apple’s Siri as well as the promise from an initial set of third party developers.
- Google knows more about us than anyone.Google has assembled an unbelievable consumer data graph based on our use of Gmail, Search, Calendar, Chrome, Maps, Waze, YouTube, Music, GooglePlay, etc. Google knows more about our day-to-day lives – our schedules, commute patterns, and questions, etc. – than our spouses. All offer essential context to be an effective virtual assistant.
- A strong base of developers is already in place.Five years ago, interfacing with an intelligence platform would have required highly trained data scientists. Now, platforms like Microsoft LUIS, api.ai, x.ai, and IBM Watson provide well-defined, easy-to-consume APIs. What once took specialized talent and months of time can be done in less than a week.
Google also faces challenges. They include:
- Minimal experience selling direct to consumers relative to its competition.Google lacks the go-to-market strategy of its largest competitors including both Apple and Samsung. No one sells more consumer electronics in the U.S. than Best Buy, Amazon and Apple – and Best Buy’s share is declining. Google’s “exclusive deal” with Verizon is likely as much about its poor track record with consumer devices as it is with exclusivity.
- A late start to building an installed base in messaging platforms.Google launched Allo – its instant messaging / virtual assistant in text format in the fall of 2016. In the mean time, Facebook Messenger alone has more than 1 billion monthly active users – not counting its WhatsApp audience. Audience equates to market power and expedited learning. The quality of virtual assistants depends on artificial intelligence which in turn depends heavily on interactions to get smarter – think Malcom Gladwell’s 10,000 hours times a large multiple because machines learn much more slowly than humans.
- A history of building a product, only to lose interest shortly thereafter.Remember Google Wave? Or Google Reader? Google Nexus Q? All products that inspired an immediate surge of consumer (and developer) excitement, only to be suddenly dropped by Google at a later date. Developers that consider building for Assistant-driven ecosystems must factor in the risk of that ecosystem being abandoned at some point in the future.
Google's going to need to show some commitment and unity to succeed with Google Assistant. Assistant needs to weather low early adoption by consumers and developers because it won't be available on all Android phones initially. If Assistant makes it through early troubles without losing momentum inside Google, Google's success with Chromecast (~30 million sold) provides it use cases Alexa doesn't have. If Google hopes to succeed, expect to see aggressive promotion for Home this holiday and big name third party integration announcements for Assistant – though maybe not from Samsung given its Viv acquisition. There may even be a developer program aimed at closing the gap to Alexa's 3000 "skills." Then, look to see Assistant brought to all Android phones as fast as Google can make it happen next year.