Microsoft is embarking on what it calls the “next generation of intelligent assistant technologies” today, looking to move forward to a new era in the digital brain battle as Cortana has so far not caught on the way Amazon’s Alexa and the Google Assistant have.
At its Build event in Seattle, Microsoft showed a demonstration of a more free-flowing conversation between a human and virtual assistant that goes beyond the typical single command and response and included shuffling around a schedule and creating new meetings on the go. This is just one example of what Microsoft called the “next leap in natural language interface technology.”
These announcements represent the fruits of Microsoft’s acquisition of “conversational AI” startup Semantic Machines last year. Semantic Machines aims to advance the state of voice-based AI from understanding and responding to commands to having complete conversations. The company was built by accomplished startup entrepreneurs, a former chief speech scientist for Apple’s Siri and leading AI researchers and professors from Stanford and University of California at Berkeley.
Today’s digital assistants are limited in their capabilities and “aren’t focused on learning how to do new things, or mixing and matching the things they already know in order to support new contexts,” Semantic Machines Co-founder and Microsoft Technical Fellow Dan Klein said in a blog post. He added that Semantic Machines teaches virtual assistants about context so they can grow beyond simple questions and answers.
Semantic Machines wants to teach virtual assistants like Cortana not just to open up skills to access siloed data, but to also go and gather data independently. That can stimulate more in-depth conversations.
“Everything you say is contextualized by what has come before so you can do more complicated things: you can change your mind, you can explore,” said Klein. “Moreover, once things get contextual enough, the notion of a skill begins to dissolve.”
Microsoft’s competitors are also working on advancing their digital assistants with better conversational tools. Both Alexa and the Google Assistant recently added new capabilities for follow-up questions and continued conversations.
Despite having some built-in advantages, principally an ecosystem of 800 million active Windows 10 devices around the world, Microsoft has failed to gain ground in the digital assistant race. In January, Microsoft CEO Satya Nadella noted that the company no longer views Cortana as a competitor to the Google Assistant or Amazon’s Alexa, and the priority has changed to make it a skill that works within other digital assistants across a variety of devices.
A Cortana executive last year said the company is playing the long game with its virtual brain, and Nadella has compared digital assistants to web browsers, saying they need to work together.
“The idea that you are always going to start with one wake word and one assistant is not how we start on the web,” Nadella said during his keynote at Build. “What does an open assistant future look like, similar to an open web?”
Microsoft is working with Amazon to integrate their respective digital assistants, and it’s worth watching to see how today’s announcements affect that relationship, or if Microsoft will continue down the path and pull off a similar deal with Google. Alexa is getting deep into the Microsoft ecosystem, as the digital assistant now works with mainstays like Skype and Xbox.
Under Nadella, Microsoft hasn’t been afraid to de-emphasize areas that aren’t paying off, or make a major pivot to re-invigorate a stagnant product, and that certainly seems to be happening with Cortana. In a recent Windows update, Microsoft separated Cortana from search and add the option to mute it during setup.
These changes came a few months after Microsoft’s Cortana chief left the company.