The world of personal computing is in a perpetual state of evolution. And there is no greater force instigating the advancement of technology than competition.
Throughout this series, we have focused on the strategies and technologies that Microsoft is bringing to the table in the industry's shift toward AI and bots. Of course, though Microsoft is an industry leader in technologies such as natural language processing, machine learning, AI and deep neural networks, the company does not exist in a vacuum.
Industry juggernauts such as Facebook, Google and Apple are positioning themselves for success in an age where communication mediums such as messaging platforms or Conversation Canvases (as Microsoft calls them), infused with intelligence are poised to become a major platform for personal computing.
The artificial intelligent digital assistant war is in full swing.
Firms are also making their respective artificial intelligent digital assistants more competitive. By enhancing them with more capabilities, merging them with other tools or opening them to third-party developers the artificial intelligent digital assistant war is in full swing. The industry giants are not the only players offering appealing solutions in this area, however. The startup Viv, founded by the creators of Siri, presents a captivating answer to the AI digital assistant.
Competition is indeed fierce. Microsoft's comprehensive and strategic approach is ambitious, but not without its challengers. In this piece, we will highlight what Microsoft's competition is bringing to the AI and bots table.
Facebook Faceoff (or Say it to our face!)
Facebook's CEO, Mark Zuckerberg, envisions a world where apps will begin to fall into the background as intelligent bots act on behalf of many services to accomplish a user's tasks. If this sounds familiar, this is the same vision Microsoft's CEO communicated during BUILD 2016. It is, in fact, the vision of several companies that are investing in technologies and platforms to offer their own flavor of artificial intelligence and bots.
Facebook began as a small social network on a college campus in 2004 and has grown into a global force in technology. The breadth of the company's investments is vast. Facebook boasts of having 1.4 billion users of the main Facebook app, plus a billion each on Messenger and WhatsApp, 700 million on Groups, and 500 million using Instagram.
Facebook broke Messenger from the main app in 2014.
Facebook broke Messenger out of the main Facebook app making it a standalone app in 2014, positioning the company to further develop Messenger as its own platform. The goal was to move Messenger beyond a tool for mere text-based person-to-person dialogue tied to the Facebook network to something more.
In April of this year, Zuckerberg announced Messenger Platform which allows developers to build bots for the Messenger app. This platform is powered by artificial intelligence which allows a developer's bots to communicate with users using natural language. Zuckerberg demonstrated receiving a daily digest directly in the Messenger app from a CNN bot. He asserted that the digest would become more personalized as the service gets to know the user over time. Furthermore, obtaining information about a particular topic is also as simple as sending a message to the bot. He also demoed ordering flowers via the 1800flowers bot. By merely messaging the bot, an order can be placed for delivery.
By partnering with a host of businesses such as Bank of America, Staples, Burger King, Fandango and others Zuckerberg believes that Messenger is going to be the next big platform for sharing privately and for connecting with a variety of services.
"M" and them, an AI and human mesh
Facebook's investments in AI and bots also led the company to the development of its own AI which is simply called M. M is a purely text-based AI, which unlike Cortana or Siri, does not receive verbal input nor does it dish it out. Another unique quality of Facebook's approach is that M is an AI-human hybrid service that is interacted with via the Messenger app.
The human component of the system consists of actual people who pick up the slack when the AI cannot facilitate a request. According to Facebook, every request that goes through the system is seen by a human "trainer", as they're called.
The role of the trainer is both to complement the AI and act as a model from which the AI will learn new tasks over time. The goal is that learned tasks would become increasingly autonomous, thereby diminishing over time the role of the humans behind the curtain.
Though Facebook plans to increase the number of trainers and expects the system to become more autonomous over time, an obvious challenge to this model is scalability and personalization. How will Facebook meet the challenge of an eventual roll-out beyond the thousands of California-based beta-testers to a very diverse set of over a billion Messenger users around the world?
Microsoft's canvas advantage
In contrast to Facebook's single-Canvas approach, Microsoft's Bot Framework makes Microsoft's solution a platform to develop bots for a range of canvases. A developer can target Slack, Line, Skype and other canvases for his bots using a single set of developer's tools. This solution is potentially much broader and more efficient than targeting a single platform such as Facebook Messenger.
Grappling with Google
Google's Google Now has been in the AI digital assistant game for some time now. Google Now has been hailed by many as the most useful of the current assistants thanks to it predictive abilities and connection to Google's industry-leading search engine and knowledge graph backbone. What has been raised as a shortcoming, however, is that it is not as personal an assistant as Cortana or even Siri. Microsoft launched Cortana as the first personal digital assistant due to her ability, supported by her Notebook, to get to know and serve a user's preferences over time.
Google Assistant, Google's evolution of its AI, has taken a play from Cortana's book and is being touted as a context aware and personal assistant that knows you. An impressive demonstration of the Assistant's context-awareness can be seen in the video below. A user simply asks, "what is this?" and the assistant correctly responds with the name of the monument that a user is standing in front of at the time.
This new conversational assistant was described by the company as "building each user their own personal Google." Given Google's market power in search, this motto communicates a very appealing promise to users of a more personal Google. This more personal, conversational approach also positions Google to intensify the commitment of it's user base to its search engine. As a conversational assistant, Google's goal is to develop a two-way, ongoing dialogue between the Assistant and the user.
Google Assistant is like giving each user their own individual Google.
Furthermore, Google hopes that the Assistant will become more aware over time. The company presented a scenario where the Assistant will be able to suggest movies that are playing in the area for instance — and the user could note that they're bringing the kids with them for the night and the system would then adjust its suggestions to a curated list of child-friendly movies.
Following the conversation
Google's Assistant is touted as a conversational assistant. The AI's conversational prowess was demonstrated when follow-up queries to an original question were tossed at the AI without the user having to repeat the subject. For instance, the Assistant was first asked, "who directed the Revenant?" to which it replied with the correct answer (Alejandro Gonzalez Inarritu). It was then asked, "Show me his awards." Google remembered the context of the previous question and gave the appropriate answer (4 Oscars and a slew of other awards).
This ability, of course, is not unique. Microsoft demonstrated that Cortana's could do the same during her debut at Build 2014. Cortana's "Chit Chat" function also demonstrates this ability. Admittedly, tests with Cortana on my Lumia 1520 running Windows 10 gives spottier success with follow-up questions than I remember when using Cortana on Windows Phone 8.1.
Google claims to be an order of magnitude ahead of rivals with follow-on queries.
Google claims its investments over the last ten years have made its systems the best natural language processing technology in the industry. They boast that what they've achieved with follow-on queries is an order of magnitude beyond its competitors. If the company's advancements in natural language processing are what Mountain View claims, the Assistant's presence in the company's new messenger app should equip the platform with a unique advantage over its rivals.
Did you get the message?
Google's new messenger app Allo is an acknowledgment that Google recognizes the move of the industry to messaging as a platform.
By imbuing this new app with the intelligence inherent to its advanced AI, Allo learns how a user communicates over time. As a result, the system becomes efficient at suggesting words or phrases based on past experiences. It will even give suggestions for images.
Google's new Assistant, like Amazon's Alexa, has also been given a more ambient "home" beyond the smartphone. Google Home is a standalone, always listening home-based unit. The device positions the Assistant to serve and learn from users in a "non-mobile" setting where assistants have traditionally "flourished."
Microsoft's Cortana can
Cortana's Notebook provides many categories such as movies, sports, news, and more that can be used to save a user's interests. For example, Cortana will proactively suggest movies that are playing near a user based on their interests. She will also recommend places around the user to dine for breakfast, lunch or dinner (based on time of day).
Cortana proactively uses the Notebook to provide highly personal responses on her interface or Live Tile. She even uses her context-awareness to provide coupons based on a store a user is visiting.
Apple's AI attack
Siri, of all of the AI digital assistants, dominates consumer mindshare, though you could make a case that this has has more to do with Apple's position in the market and its almost magical marketing power than any advantages Siri might have over rivals.
In fact, Siri's association with Apple as both the target of a media that is alternating between fawning and breathlessly alarmed has brought the assistant its fair share of criticism over the years.
It seems that Apple has taken notice and has quietly improved upon the assistant's abilities. Back in 2014 Apple followed Microsoft's lead and invested in neural networks as the technology upon which Siri was built. This change has made a noticeable improvement to Siri's accuracy. Furthermore, Apple has integrated AI throughout iOS 10, making the OS more "aware" and capable of providing proactive assistance within various functions.
Apple opened Siri to third-party develpers at WWDC.
Moreover, in a move that betrays the perception of Apple as a "walled garden", they recently opened Siri to third-party developers. Now, as with Cortana, developers will be able to connect their apps directly to Apple's digital assistant. This advancement will allow a user to ask Siri to perform a task, and she will be able to connect to an app to do so.
Due to Apple's strong developer support, there will likely be a flood of available apps (unlike with Cortana) that will take advantage of this feature.
Alway's messaging around
As with Facebook, Microsoft and Google, messaging as a platform has taken a front row seat for Apple. Moreover, iMessage as the most-used app on iOS is indicative of the importance of messaging for consumers. Usage data combined with the industry's trend toward messaging as a platform has driven Apple to treat iMessage to a much-deserved overhaul in iOS 10.
Some may dismiss most of what Apple presented by way of advancements in iMessage as unnecessary eye-candy. I would contend many of those changes have made iMessage even more engaging for users. Of course, engaged users is something every platform needs.
Apple has made iMessage a platform for third-party developers.
Apple has also opened iMessage to third-party developers. Developers can now add their flair to the messaging interface through additions such as interactive conversation bubbles and more.
iMessage is Cupertino's answer to bots. Third-party developers will be able to build apps that will allow users to get things done such as send money, order food and more, directly from Messages. Due to the company's strong developers and consumer relationships, iMessage apps will likely be well supported and well received.
Microsoft Bot Framework paints a bigger picture
As with other rivals, Microsoft's platform approach via its Bot Framework and Conversation Canvases strategy gives Redmond a wider net to enlist developers than Cupertino's iMessage approach provides. Still, Apple's relationship with developers is undeniably strong.
Can't Viv with'em can't Viv without'em
Viv is a startup cofounded by the creators of Siri, who left after selling the company to Apple (Siri was originally an app on the iOS App Store before being purchased and integrated into iOS). Viv Cofounder Dag Kittlaus explained that Viv means life and that the company's goal is to "breathe life into inanimate objects and devices in our lives through conversation." Viv's creators are taking an ambitious approach seeking to provide the go-to AI for the entire industry.
According to Kittlaus, no one company will be able to plug into every single service that a user will want to use. It is for this reason that Viv, according to Kittlaus, has been in discussion with major companies. Their hope is to make Viv such a ubiquitous industry standard that the Viv logo on a new a product will signal to consumers that they can talk to that product.
Viv's goal is ubiquity.
Viv will be unbounded and will know a user across devices (not unlike Satya Nadella's vision for Cortana). Kittlaus described the simplicity of buying a new product, logging in, and having immediate access to Viv and any personal information associated with the AI. Viv's goal is to enable users to talk to virtually any product.
Speaking of talking, Viv uses Nuance to power its natural language understanding. Kittalus demonstrated talking to Viv to order flowers (a pretty popular theme it seems). He also asked the AI about the weather two weeks from that date, after 5 pm at a particular location. Viv responded promptly and accurately to this highly specific query.
Viv's abilities seem to be comparable to its rivals. The major advantage Viv seems to have over its rivals, however, is the ability for the system to "program itself." Kittlaus demonstrated that Viv could derive responses to queries from Dynamic Program Generation. This "self-programming" is a capability which is clearly beyond the very rigid, "some query related to some domain" model that governs the responses of other AIs and bots.
Through Dynamic Program Generation, which Kittlaus described as a breakthrough for the company, the system can essentially program itself after understanding the user's intent. So rather than being constrained by a rigid query-response structure, in the demo above, according to Kittlaus, the Viv AI automatically wrote a 44-step program in 10 milliseconds in order to find the answer a query.
Viv can program itself.
As an unbounded, OS-agnostic AI, Viv could be a formidable rival to Cortana, Siri and Google Now. Moreover, in a quest for ubiquity the company's "device-less" approach is akin to the strategy Microsoft is employing for Cortana. As a marketing message to consumers, Kittlaus described Viv as the intelligent interface for everything. For developers, the AI is described as the next major marketplace and channel for offering content, commerce and services.
When asked what his response would be if Mark Zuckerberg offered him "all the kingdoms in the world" for Viv, Kittalus replied:
Our goal for this is ubiquity. And were going to follow the path to ubiquity. We've had acquisition offers in the past that we have not gone with. We're friends with all these guys and there's a lot of interest and active stuff going on this in this area. We're going to stay true to what we think is the right way to get to get to ubiquity. We're not going to predetermine what path that is, but we're determined to finish the job, for sure.
Will Viv live at Microsoft?
Of all of the big technology firms, no company's mission speaks to ubiquity like Microsoft's platform-company approach to personal computing does. From it's cross-platform apps and services like Cortana and Office, to Windows as a dev box for cross-platform app development, to a robust OS-agnostic Cloud platform, Microsoft is all about ubiquity.
I don't know if Microsoft is one of the company's that has been in talks with Viv, nor if the technology behind Viv could be integrated smoothly into what Microsoft has already established with AI and bots. However, Microsoft has shown no hesitation in purchasing smaller companies that help to facilitate their mission to be the platform that helps people gets things done.
That said, from the outside looking in, Viv and Microsoft look like a match made in heaven. Then again, Viv's apparently advanced self-programming AI could find a home in any Silicon Valley giant. Could Kittlaus' statement have been a subtle petition for a partnership with someone like Microsoft? Maybe.
There war is on!
Some techies and enthusiasts are of the opinion that there is no future for AI and bots. They assert that the "warehouse of apps model" is alive and well and will not be altered. It is evident, however, by the massive research and development investments by Microsoft, Apple, Google, Facebook and others, that powerful industry drivers are of a different mind. That does not mean that they are right, of course. The actions of each of these companies do, however, force innovation, research and investments in this new frontier.
Consequently, the compelling vision of AI and bots coupled with the steady advancements of integrated intelligence in messaging platforms and other Canvases is hard for the media to ignore. Coverage of these advancements brings our progressive approach to the borders of this new frontier to the mainstream. This promotion results in the ever-important "Joe-consumer" becoming an aware and often interested party in the technology. In essence, demand is provoked in part for what the industry's powerhouse's push.
I think the shift to a more intelligent "app" model hosted by AI's and bots is inevitable. "The real question is, "Who will emerge as an industry leader in this new frontier?"
Who do you think will win the AI and bot war? What company's messaging platform is best positioned to take advantage of a shift to messaging as a platform? Sound off in comments and on Twitter!
If you missed any part of this series or these related pieces catch up here!: