Instead of drawing its predictions from a set database of existing words, Swiftkey analyzes your particular writing style in an effort to help the future you.
Scientists call this machine learning, and under Microsoft, Swiftkey's technology stands to spread into practically every product the Redmond, Wash., company owns. Imagine if Microsoft Word or Outlook anticipated what you wanted to say in a document or email. It could be integrated into a host of other apps Microsoft has acquired, such as Wunderlist, its to-do list app, or Sunrise, its calendar app.
SwiftKey's predictive technology aligns with Microsoft's investments and ambition to develop intelligent systems that can work more on the user's behalf and under their control.
Analyzing millions of user keystrokes won't just help Microsoft supply your next turn of phrase; it'll also give the company access to troves of valuable behavioral data that is now treated like currency among tech companies large and small.
Chances are this tradeoff will be worth it for many consumers, just as many of us have willingly let Google mine our search history and Facebook see who we communicate with. But it does potentially give Microsoft another leg up in the increasingly competitive tech industry.
This is all taking place against the backdrop of a wider race to design the world's best artificial intelligence.
One of Google's longtime goals has been to design a service that can anticipate your informational needs before you even realize it - covering weather updates to traffic data to upcoming appointments.
The auto industry is throwing gobs of cash at robotics, too, in preparation for a world of driverless vehicles you can summon on-demand. Even Facebook's Mark Zuckerberg is spending the year trying to design his own, personal AI as a side project. If he succeeds, you can be sure the product will be integrated into the social network's own technology.
The AI wars have begun - and no, they don't always involve gun-toting Terminators.