Democratizing AI: Microsoft’s User-Friendly Bots, Services and Frameworks

By Charles King, Pund-IT, Inc.  December 21, 2016

Something that became increasingly clear during 2016 is that artificial intelligence (AI) is considerably more than a marketing-infused buzzword. Instead AI is on a near-term track to inform and inspire solutions and services in both the consumer and business spheres.

Why is that the case? In contrast with other tech industry favorites, like virtual (VR) and augmented reality, AI-based product development is blessed by low barriers to entry. Consider the thousands of dollars (or tens of thousands in the case of business solutions) required for a quality VR headset, accessories and supporting PC.

In stark contrast, most of the major vendors focusing on AI, including Amazon, Facebook, Google, IBM, Intel, Microsoft and Salesforce are making the significant investments required for creating AI platforms, then encouraging and working closely with customers and developers to create fully formed solutions and services. That dynamic was on clear display in the AI announcements that Microsoft made in San Francisco on December 13th.

Microsoft’s AI history

Before getting into that, what sort of energy and pedigree is Microsoft bringing to the AI party? As noted by Dr. Harry Shum, the EVP who leads the 5,000 engineers, computer scientists and other employees in the recently created Microsoft AI and Research business unit, artificial intelligence has been a focus area since Microsoft Research was launched in 1991.

In fact, computer vision, speech recognition and natural language processing were among the earliest projects pursued by the group. More recently, the company has made considerable inroads in AI development with its Cortana personal assistant technology. Other Microsoft product groups incorporated in the new unit include Microsoft’s ambient computing, Bing, information platform and robotics teams.

The AI and Research unit will emphasize infusing AI capabilities and technologies into four distinct Microsoft solution and service arenas: Agents, like Cortana; Applications, including Skype and Office 365; Services that will be made available to developers worldwide; and Infrastructure – specifically the Azure cloud computing platform. Shum also noted the foundational value that the company’s quarter century of AI research will provide to these new efforts.

In addition, Microsoft plans to collaborate with other vendors to drive AI development. The day after the AI and Research business unit was launched, the company announced that it had joined Amazon, DeepMind/Google, Facebook and IBM in the new Partnership on AI, an organization that aims to study and formulate best practices on AI technologies, advance the public’s understanding of AI, and serve as an open platform for discussion and engagement about AI and its influences on people and society.

Conversational gambits

So what did Microsoft announce in San Francisco? Like IBM, Intel and Salesforce, Microsoft is pitching its strategy around AI as an effort to “democratize” the technology. That is, taking on the platform-related heavy lifting and providing customers and developers AI-related services and tools. But the announcements made in San Francisco specifically emphasized conversational computing. Those included:

  • A new chatbot named Zo that launched in the U.S. in October that the company says already has 115K users. Zo follows earlier Microsoft chatbots Xiaoice (launched in 2014 in China with over 40M current users) and Rinna (launched in 2015 in Japan; the company says Rinna has had regular conversations with 20 percent of Japan’s 127M population).
  • A Cortana Devices SDK and Skills Kit designed for OEMs and ODMs, plus a related partnership with Harmon Kardon to create a premium home speaker (similar to Amazon’s Alexa and Google’s Home solutions)
  • help, a scheduling service that enables humans in the loop to leverage AI capabilities
  • Cognitive Services for Microsoft’s Bot Framework, including the new QnA Maker and updates for the Language Understanding Intelligence Service. The company also noted that 67,000 developers are currently using the Bot Framework since its launch last March.
  • A preview version of Microsoft Translator Live which allows users of the Microsoft Translator apps and end user Web site to carry out real time, multi-lingual translations
  • Tools for partners using the Microsoft Bot Framework to build rich media cards featuring video, animated Gifs and audio clips
  • New capabilities for Skype bots, including receiving and handling Skype voice calls.

Why is voice enablement/recognition so important to AI? There is certainly an important historical angle to it with intelligently vocal computers and robots long occupying special places in modern culture, including classic television programs and films, from Star Trek to 2001: A Space Odyssey to Star Wars.

The massive success that IBM enjoyed from the debut of its Watson system on the Jeopardy! game show showed the industry a way forward for next gen AI. Plus, the deep market penetration of voice-enabled mobile solutions, like Apple’s Siri and Google Voice, and growing popularity of home devices and services, such as Amazon’s Alexa and Google Home are also important factors to consider.

For these and other points, it is reasonable to conclude that conversational computing is simply table stakes for any vendor serious about AI. Microsoft already has substantial assets, like its Cortana personal digital assistant, that it can bring to bear in such efforts. With these newly announced conversational capabilities, features and tools, the company is effectively saying that it has been, is and will continue to be a force in the marketplace.

AI’s business benefits

Microsoft’s announcements also aimed to help vendors and end-users better understand the benefits that AI-related solutions, like intelligent bots can provide businesses and other organizations. In a post related to the AI announcements, Microsoft senior director of product marketing Herain Oberoi noted that intelligent bots can deepen business engagements by:

  1. Leveraging social and historical context to better infer intent and make actionable recommendations
  2. Helping uncover new insights about customers’ challenges and preferences
  3. Driving efficiencies via automating workflows and integrating task completion

The company also cited notable examples of customers leveraging Microsoft’s Bot Framework and related technologies. Those include Australia’s Department of Human Services which is creating a bot-enabled “expert system” for responding faster and more effectively to citizen queries; the Bank of Kochi in Japan which is developing a receptionist bot; and Rockwell International which is creating a bot for production automation.

These and other examples demonstrate that bots aren’t merely tools for completing elementary tasks, like responding to simple customer requests. Instead, intelligent bots are developing a widening range of complex skills and can provide enhanced insights by analyzing rich interaction data. In other words, the intelligent bots that Microsoft envisions and helps customers develop are delivering significant, continuing value to the organizations they serve.

Final analysis

As noted before, Microsoft is hardly alone in its enthusiasm for and efforts around artificial intelligence. Formidable competitors, including IBM, Intel, Google, Salesforce and others are also in the hunt and making significant investments of their own.

That said, these efforts are not all equal or progressing at the same rate. IBM has invested extensively in its Watson cognitive platform and related technologies, including The Weather Company’s product and technology assets that it acquired in January. The company is also pursuing high profile Watson partnerships with organizations, such as the Sloan Kettering Cancer Center.

Salesforce has also bolstered its Einstein AI platform via acquisitions, including Demandware, Implisit, MetaMind and TempoAI in the last year or so. However, since the company’s AI solutions are still emerging, the value and impact of those deals are impossible to estimate.

Microsoft’s AI solutions and tools, including those highlighted in its recent announcement, show a measured and practical approach that should appeal to its business and developer customers. The company’s point is that AI isn’t a complex study that requires massive commitments or a specialized technology with more restrictions than a private country club.

Instead, Microsoft’s Bot Framework, intelligent bots like Zo, Xiaoice and Rinna, conversational computing tools and other cognitive technologies are available and applicable today, to any interested individual or organization. That is the practical, real-world effect of the democratization of artificial intelligence, and one that should serve Microsoft and its customers well as AI continues to mature and evolve.

© 2016 Pund-IT, Inc. All rights reserved.