On January 13, 2018, Google's CEO, Sundar Pichai, skimmed the headlines pouring out of the Consumer Electronics Show (CES), held in Las Vegas, Nevada. No press conference or news release went by without some mention of Google Assistant. Just one year prior, Amazon had dominated these same headlines with news that "Amazon's Alexa Just Conquered CES 2017. The World Is Next."2 This year was slightly different, Google Assistant made it onto almost every hardware spotlight at the show, determined to show how all of these products could work together. Prominent ads throughout the city featured "Hey Google," the command used to wake up Google's intelligent assistant. Pichai wanted the world to know that it should not count Google out of the voice wars.

By almost any measure, Google's performance as a company in the past few years had been exceptional. At the close of 2017, Google's parent company, Alphabet, had achieved one of the highest market values of any public company at $727B.3In 2017, revenues for Alphabet totaled $110.9 billion, up 23% over 2016.4 Google's core advertising business, at $95.4 billion, represented 86% of Alphabet's sales in 2017 (see Exhibit 1). Pichai wondered how Google could replicated its success in search in the voice assistant space. Despite Google Assistant's rapid adoption, Google faced stiff competition: Apple was the first to market with the launch of "Siri" in 2011 on the iPhone 4S, followed by Microsoft's rollout of "Cortana" across the Windows ecosystem. However, it was Amazon that created a new market with the launch of a dedicated voice-enabled device: the Echo smart speaker powered by Amazon's intelligent assistant, Alexa. As of April 2018, Amazon Echo captured 66.7% of U.S. smart speaker users compared to Google Home's 29.5% share.5

By June 2018, Google had aggressively expanded Google Assistant's presence to more than 400 million devices, including smart speakers, phones, tablets, and watches.6 Some reports applauded the company's success: "Google won the voice assistant popularity contest at CES. Your move, Alexa."7

How could Google capitalize on the success and momentum out of CES? What strategic moves would allow Google to build out its platform to compete with Amazon as well as the growing global competition? Would voice be a core or ancillary technology for Google moving forward?

Google's History

Google co-founders Larry Page and Sergey Brin met in 1995 at Stanford University. While working on their dissertations, Page and Brin created an algorithm that used an index to sort web pages by relevance. Two years later, they decided to refine this search engine and named it "Google," a derivation from the mathematical term "googol.": the term was rumored to reflect the founders' mission to organize the infinite amount of information on the internet.8 In 2000, Google built AdWords as a pay-per-click, keyword-targeted advertising service, and their advertising revenues enabled the company to reach profitability. After bringing on former Novell CEO, Eric Schmidt as Google's new CEO, Google launched a successful IPO (ticker: GOOGL) in 2004, raising $2 billion.

Over the next decade, Google launched new products and made a series of strategic acquisitions, including mobile OS maker Android, online video platform YouTube, and smart thermostat maker Nest. As the product lines grew, Page and Brin restructured the company and formed Alphabet, Inc. in October 2015 to increase transparency and better delineate its business units. Page explained that the new structure allowed Alphabet's companies to have independence, and "we also like that it means alpha-bet (Alpha is investment return above benchmark), which we strive for!"

Pichai joined Google in 2004 and led product management on software products, including Chrome OS, Android, and Gmail. As CEO of Google, Pichai earned a reputation for championing artificial intelligence when he declared in an annual all-company memo: "We will move from mobile first to an AI-first world [...] a world where computing becomes universally availablebe it at home, at work, in the car, or on the goand interacting with all of these surfaces becomes much more natural and intuitive, and above all, more intelligent."10Pichai's biggest decision as CEO had been to set up two new Google divisions, cloud computing and hardware. In the cloud business, Pichai bet that AI- equipped tools would set Google apart from its main rivals. The hardware division introduced a diverse line of products ranging from the Google Home Mini & Max smart speakers to the Pixel line of phones, Pixelbook (laptop), Pixelbook Pen and Pixel Buds (earphones). All of these products featured integration with Google Assistant.11

Google Assistant

In May 2016, at their annual developer conference, Pichai first introduced the Google Assistant for public use: "We think of it as a conversational assistant. We want users to have an ongoing, two-way dialogue with Google. We want to help you get things done in your real world and we want to do it for you, understanding your context, giving you control of it. We think of building each user their own individual Google."12

Intelligent Assistants

Intelligent assistants had become commonplace in premium tier smartphones and other assistant- enabled devices by 2018 (see Exhibit 3 for installed base figures across competitors). Intelligent assistants (also known as virtual or voice assistants) were computer software programs that translated conversational text or spoken language requests into completed services. By early 2017, Android users

conducted 20% of searches by voice through Google Assistant instead of text.13By 2020, estimates projected that 50% of all searches would be voice searches, signaling a world where intelligent assistants played an ever increasing role in users' lives.14

A trigger, such as "Hey, Google" or "Alexa" or "Siri," activated the intelligent assistant on certain devices. The assistant filtered out distracting audio signals (voice echoes, background noise, the device's own loudspeaker) in order to capture the user request.15This request was then converted to text and run through a natural language processing algorithm on the cloud for analysis. Natural language processing included the breakdown of phrases into intents (what is the request's purpose?) and entities (does the request include a date, time or a contact?).16 These intents and entities were used to access a database of possible responses, which was then converted into an audio clip issued back to the user. Over time, intelligent assistants became smarter as more data fed into the algorithm.

Three primary technological advancements enabled the growth of intelligent assistants. Access to data was no longer a hurdle given the amount of information generated on the internet and mobile networks. Many companies, such as Netflix, sponsored competitions to encourage labelling of data or usage of data for recommendation systems.17In addition, computer scientists had made major breakthroughs in machine learning (ML), a specific AI sub-category that improved task performance with increased access to contextual experience and data.18 In 2009, Nvidia's Graphics Processing Units (GPUs), originally developed for video games, achieved the processing power capabilities needed to train deep neural networks, which were ML computing systems loosely modeled on the human brain.19These three factors together led to the explosion of machine learning, which enabled many of the fields associated with modern AI capabilities, including computer vision, speech recognition, and natural language processing. Google had invested heavily in AI and had started building its own Tensor Processing Units (TPUs) to compete with GPUs and to focus specifically on neural network machine learning.20

Competition in the Intelligent Assistant Landscape

Google Assistant initially debuted on Google's messaging app Allo in 2016 and was exclusive to the Pixel smartphone for a period before becoming more widely available on Android devices. To speed adoption, Google positioned Google Assistant across multiple devices and channels, which included preloading Google Assistant on Android smartphones and launching a Google Assistant app on iOS. As a result of these efforts, Google Assistant accounted for 46% of all smartphones equipped with an intelligent assistant by May 2018.21Google Assistant grew from being available on just one device in one language in 2016 to being available on multiple devices with translations available in 22 languages by 2018.22The tech giant predicted that the Google Assistant would be available in 52 countries and support over 30 languages by the start of 2019.23

In addition, Google started to open up the Google Assistant ecosystem in December 2016. The company launched "Actions on Google", a platform allowing services enabled by third-party developers. Google's Software Development Kit (SDK) allowed third-party hardware to run Google Assistant, and Google incentivized developers to build actions for the platform through competitions and prizes. Google kept tight control over which actions were approved on Google Assistant and added to the ecosystem gradually over time. In early 2018, Google Assistant had over a million conversational action-based services enabled, but only 1,830 voice actions (e.g. "Hey Google, dim the lights").24 Notable manufacturers such as Sony and Panasonic partnered with Google to make Google Assistant accessible on their own devices, such as smart speakers and TVs.25

Google worked aggressively to grow its assistant's market share, but there were several companies vying to be the dominant player in the emerging intelligent assistant market.26

Apple Apple was first to market with its voice-activated intelligent assistant, Siri, which it launched in October 2011 with the release of the iPhone 4S. Siri was a voice-activated intelligent assistant that Apple had acquired the previous year.Apple maintained a closed ecosystem, making Siri available only for Apple products until 2016, when it opened up its ecosystem to third-party applications.28 However, Apple restricted integration to specific application categories, which included voice calling, ride-hailing, messaging and search. Apple was renowned for its strict stance on protecting consumer privacy, and tagged Siri searches with random user identifiers to protect consumer privacy and stored tagged data for only six months before deleting it. Observers noted that this decision reduced Apple's ability to use the data to evaluate software updates and improve Siri's functionality.29According to some analysts, Siri's technology paled in comparison to other smart assistants: a 2017 study of the accuracy of answers to questions showed that Siri answered 62.2% of questions correctly, while Google Assistant, Alexa and Cortana were all above 80%.30 Thanks to the iPhone's massive installed base, however, Siri was the most widely used in terms of unique monthly users in 2017, with 41.1 million users. By 2018, Siri was also accessible in 21 languages in 36 countries.31

Amazon Alexa, Amazon's cloud-based voice technology, had its roots in the technology of a little-known British company called Evi.32Evi launched its voice assistant in 2012 and was positioned as a competitor to Apple's Siri before Amazon acquired Evi that same year. Initially, Amazon planned to leverage Evi's technology to build an artificial speech-based book reader.33 This narrow vision later evolved into an idea to have a new platform that would be powered by a combination of Amazon Web Services (AWS), speech recognition, and high-quality speech synthesis and would be tied to an affordable piece of dedicated hardware, ultimately producing the Alexa-powered Amazon Echo Smart Speaker, which launched in late 2014. The next year, Amazon announced the Alexa Skills Kit (ASK) - a collection of self-service APIs and tools that made it easy for third-party developers to create new Alexa skills. This open platform strategy accelerated the number of Alexa skills to about 5,000 by the end of 2016 and over 25,000 by the end of 2017, far more than its rivals. Amazon offered a wide variety of skills such as playing games such as Jeopardy, ordering an Uber ride, and asking about the weather and news. However, Alexa had limited availability across phones and tablets and as of 2018, it was available only on the Amazon Fire, HTC U11, and the Essential phone. Additionally, it was available as an application for Android and iOS. Alexa was available in over 80 countries and supported 4 languages.

MicrosoftIn April 2014, Microsoft launched its voice enabled digital assistant, called Cortana,as part of an update to Microsoft's mobile operating system. In 2015 and 2016, Microsoft expanded Cortana to iOS and Android through mobile applications.36 In May 2017, Microsoft publicly launched the Cortana Skills Kit for developers looking to build new features and apps for Cortana on desktop, mobile, Xbox, and the newly-announced Cortana-powered smart speaker launching in fall 2017. As of January 2018, Microsoft Cortana had 235 skills, representing a 305% growth in the number of skills in the last 6 months of 2017.37However, many major brands that had created skills for other intelligent assistants had not created skills for Cortana. Some developers claimed the certification process was inefficient and less clear than those offered by Amazon and Google.38In May 2017, Microsoft announced that 500 million Windows 10 devices were integrated with Cortana, with 145 million monthly active users of Cortana worldwide and 60 million MAUs in the U.S.39As of March 2018, Cortana was available in 13 countries and supported 8 languages.40

Samsung Samsung Electronics was a multinational electronics company headquartered in South Korea. In 2017, Samsung led the overall global smartphone market with 18% market share, selling a wide variety of smartphones that spanning a range of price points.41 In 2016, Samsung acquired Viv, a voice assistant start-up from the makers of Siri, for integration with Samsung's products. Samsung further developed its AI voice assistant (renamed "Bixby") to work across Samsung devices. As of 2018, Bixby became the standard on Samsung's high-end Android phones, which also supported Google Assistant.

Voice start-ups By 2018, a number of smaller players had entered the voice market such as X.Ai, Ozlo and Mycroft. Mycroft, an open source software and hardware AI platform, raised a total of $1.75M in seed funding to help realize its vision.42 Mycroft's open sourced approach allowed users to develop hardware and software on top of the platform to explore more use cases for the voice AI technology. Mycroft had landed a partnership with Mozilla to help the company build out its deep search library for speech-to-text transcription. The CEO of Mycroft, John Montgomery's vision focused on flexibility and customization: "The advantage of working with our technology is it can be customized, and the companies can build a user experience that represents their brand."43

Competition Grows in China

By early 2018, China had emerged as the major locus of competition to U.S. tech giants, including in the intelligent assistant market. China boasted over 100 companies that had entered the race to capture a share of the voice assistant industry. The Chinese companies spent aggressively on research and development for their versions of voice-activated devices, and some had even expanded to the

U.S. market.44Perhaps of most concern to Google were Alibaba, Baidu, and Tencent, given their strategic assets and the progress they had already made.

Alibaba Group Holdings Alibaba unveiled its intelligent assistant called AliGenie in mid- 2017. AliGenie offered much of the same functionality as Amazon's Alexa, such as ordering from Tmall, smart home control, weather, news, music, etc. AliGenie's voiceprint recognition feature offered security by only allowing authorized users to place online orders. Later in 2017, Alibaba launched an open-development platform at its annual Computing Conference.45By May 2017, AliGenie had over 200 applications, of which less than 50 were developed by Alibaba.46

Baidu The dominant internet search operator in China, Baidu launched the Duer app - a mobile based virtual AI assistantin 2015. Two years later, Duer OS, an open AI operating system, was launched at CES in January 2017. According to Baidu, DuerOS had accumulated more conversation- based skill sets than any company in the world10 major domains and over 100 sub-domains of conversational skills.47The company also launched a solution called TurnKey, which provided backend support for manufacturers to speed up enterprise demand for AI integration.48 One of Baidu's major partnerships was with Xiaomi to bring DuerOS to Xiaomi's IoT platform of 85 million connected devices.

Tencent Holdings Tencent, creator of popular messaging and social media app called WeChat, had been working on its AI assistant called Xiaowei, which was powered by data from Tencent's sprawling ecosystem (WeChat, TenCent Music, Tencent Video, and Youtu Labs).49 Xiaowei supported the usual applications such as weather reports and music requests and came with voice recognition, facial recognition, and an SDK. In the future, Xiaowei would also be able to conduct multilingual translation and object detection.

Voice Wars: Intelligent Assistants Move Beyond Mobile

After launching their own respective intelligent assistants, all the major players began focusing on manufacturing their own voice-enabled devices and/or partnering with other companies to increase their user base. More broadly, the push towards industry dominance led players with intelligent assistants to move beyond initial use cases. With a crowded landscape, Google faced large well-funded domestic and international technology giants on the one hand and ambitious lean upstarts on the other. Google had to think beyond the initial use cases and ecosystems for the Google Assistant to achieve Pichai's vision to make voice and the Google Assistant ubiquitous.

Smart Speakers

Smart speakers represented the major hardware platform for voice assistants beyond smartphones and other mobile devices. Amazon catalyzed the discussion regarding the potential applications of intelligent assistants with the launch of the Echo, a smart speaker that became publicly available in July 2015.54 This first smart speaker entrant opened up the realm of possibilities for voice assistants beyond phones, tablets, and computers. In March 2016, Amazon followed up the Echo Dot (retail price: $50), a smaller version of the original Echo. A few months later, Amazon went on to release an entire suite of hardware products powered by Alexa . In June 2017, Amazon released the Echo Show (retail price: $230) with a seven-inch screen to integrate voice and visual user interfaces for the first time. This device enabled users to watch videos and view product details on potential purchases in conjunction with voice commands.

In November 2016, Google followed suit and launched its own smart speaker, the Google Home (retail cost: $129) powered by Google Assistant. A year later, Google released two new sizes: the Google Home Mini (retail cost: $49) and the Google Home Max (retail cost: $399) in time for the 2017 holiday season ( for Google Home product portfolio). Google proclaimed to have sold more than one Google Home (largely mini speakers) per second in the fourth quarter of 2017.55Over the holiday season, Amazon responded by dropping the price of its Echo Dots to $29.99, despite estimates which suggested an Echo Dot cost $34.87 to produce.56

By the time Apple entered the market in February 2018 with the HomePod (retail cost: $349), Amazon had 66.6% of the market (down from 71% in 2017), Google 29.5%, and other players only 8.3%.57Apple tried to distinguish itself from the competition by boasting sound quality akin to Sonos or Bose and marketed the HomePod as a complement to Apple Music and iTunes. In response, Google announced partnerships with higher quality sound and smart screen display products made by JBL, Lenovo, LG and Sony, which further expanded its smart speaker features.

Microsoft, Mycroft, and big Chinese manufacturers also sought to play in the smart speaker market. Microsoft first entered the smart speaker market through a partnership with Harman Kardon (the home and car audio equipment manufacturing company), releasing Invoke (retail cost: $199) in October 2017

Despite Pichai's optimism about the future of voice and Google Assistant for the company, Pichai's team had not yet created a monetization strategy for voice. Amazon announced a monetization strategy for Alexa in May 2018, allowing skills developers to offer in-skill purchases, i.e. selling premium digital content within an Alexa skill. Amazon received 30 percent of each sale.104Google had explored monetizing voice search with advertisements. In one experiment in March of 2018, consumers would ask Google Assistant a question and then hear something like, "By the way, Disney's live action Beauty and The Beast opens today."105 The test soon ended due to user complaints and negative press.

Questions

1. If consumers used voice to search, and got their answers with a voice response, how would that impact Google's core search business? ( where as their core revenue comes from ad's)
2. Should Google keep the technology in-house to drive Google services and Google products? Should Google license the technology broadly to third parties?
3. How would the future of voice intersect with the growing number of Alphabet's businesses ( Google Parent Company)?