Foley on Microsoft
Where's Microsoft's Answer to Apple's Siri?
Mary Jo Foley breaks down where Microsoft stands with its voice recognition technology and why Redmond is confident it will outperform its competitors.
- By Mary Jo Foley
Back in 2011, it was sounding as if Microsoft was poised to roll out some kind of voice-enabled personal-assistant capability. CEO Steve Ballmer was publicly talking up the idea that users would be able to tell their PCs to "print my boarding pass on Southwest" and have their systems automatically jump into action. The magic behind the scenes would be a combination of Microsoft Bing, Tellme speech technology and some natural-language-plus-social-graph concoction.
Other Microsoft officials weren't quite as brash, noting the type of scenario Ballmer was outlining was still realistically three to five years away. That would mean Microsoft's entry into this space would be a 2014 to 2016 thing.
Why the delay? After all, the voice-recognition-enabled Kinect hit the market in the fall of 2010, as did Windows Phone 7, which included some built-in voice-processing capabilities. Did all the pre-launch hoopla around Siri push Ballmer to have a senior moment and talk out of school?
Voice understanding and natural-language processing are both long-term efforts where Microsoft has been working for more than a decade. It's only been in the past year or two that the `Softies have pushed to make some of these technologies commercially available.
Microsoft moved its speech team into its Online Services unit, seemingly to facilitate work with the Bing team, at the very end of 2011. In 2012, Microsoft divested itself of part of its Tellme speech assets (the side of the business focused on interactive voice response). This year, Microsoft unveiled two key initiatives that are related to its speech and natural-language work: Bing as a platform and Power BI.
On the platform front, Microsoft is doing more than simply enabling developers inside and outside the company to license speech, translation and mapping controls. Microsoft is giving developers access to its own machine-learning assets -- the logic, context, and relevance that are designed to keep the controls and APIs fresh. The key building blocks of this "intelligent fabric," as Microsoft execs have taken to referring to the back-end services, include its Bing "Satori" knowledge base, natural UI (NUI), and geospatial- and context-awareness technologies.
It's the relevance and comprehensiveness of the information powering these back-end services that can make or break a personal assistant technology. Just ask Apple, which is going to be using Bing as the search engine behind Siri in iOS 7. While shared anti-Google sentiments no doubt played a key role in helping broker this Microsoft-Apple deal, I'd wager that Bing's growing ability to interpret intent probably influenced Apple's choice, as well.
Power BI, which Microsoft launched at its Worldwide Partner Conference (WPC) in early July, is a natural-language-powered business intelligence (BI) tool. Inside Power BI is a piece of technology that was code-named "InfoNav." Known inside Power BI as "Q&A," InfoNav is a natural-language query engine that allows users to type queries into a dialog box, which the system understands and generates answers to in the form of tables, charts and graphs. This natural-language engine was developed together by the Microsoft Research, Server and Tools Business, and Bing teams. Microsoft execs aren't saying when and if the company will roll out other services and technologies powered by InfoNav/Q&A.
Back in Windows Phone land, Microsoft has made strides toward improving its voice-recognition and understanding capabilities. In June, Microsoft acknowledged it had been rolling out updates to Windows Phone users in the United States designed to improve the speed and accuracy of voice-to-text and voice search. "Now when you compose a text message or search using your voice, Bing will return results twice as fast as before and increase accuracy by 15 percent," company officials boasted. (Outside the United States, it's worth noting, the availability of Bing-related enhancements of any kind is iffy, at best.)
Microsoft's message seems to be that even though Redmond's late to the personal-assistant party, the wait will be worth it. We'll see, possibly next year, if users agree.
Mary Jo Foley is editor of the ZDNet "All About Microsoft" blog and has been covering Microsoft for about two decades. She's the author of "Microsoft 2.0" (John Wiley & Sons, 2008), which examines what's next for Microsoft in the post-Gates era.