I've lived with the latest incarnation of SwiftKey for about a week now and, like most reviewers, love the alternative Android keyboard. My colleague, James Kendrick, gave SwiftKey 4 a glowing review, calling it the "best mobile keyboard" and I couldn't agree with him more. In particular, its latest feature, called "SwiftKey Flow", allows users to mix regular tapped typing with Swype-style gesture-based typing, making text input extremely fast, especially on smaller devices. The company likes to point out that it has saved its collective user base a millennium of typing time (SwiftKey also collects extensive metrics as you type).
This is all well and good (and it is very, very good), but what interests me most is the AI engine behind SwiftKey that allows it to learn from your typing, from things that you write. You can connect it to your Gmail, Twitter, and Facebook accounts, as well as to an RSS feed and your text messaging, giving the software a huge amount of data to learn commonly used words and phrases. The result is predictive text that works well for even the sloppiest and fastest of thumb typers or sliders, and it makes touch typing on a tablet almost a pleasure.
Where this starts getting really cool, though, is when you start thinking about the application of this sort of machine learning and predictive input to various vertical markets and specialized use cases. In general, for example, healthcare providers love tablets for the ease of access to patient information, medical research, medical imaging, etc, but hate the finicky text input that comes with most health management systems on tablets. Doctors can barely write a legible prescription, let alone tap out notes during the limited time they're with patients. With SwiftKey, though, the combination of its predictive engine and ability to accept hastily scrawled input and turn it into the words you actually meant to enter can make tablet input simple and accurate for even the busiest of doctors or most overworked of nurses.
SwiftKey has actually done quite a bit of work in the healthcare field specifically because of this potential, developing an entire healthcare application. Obviously, the video I've embedded below is a nice piece of marketing, but it gives a sense that SwiftKey's vision is not just to be a better, prettier version of Swype. Note, by the way, that the SwiftKey 4 keyboard is only available for Android; its healthcare solution is available on both iOS and Android, because the predictive input is embedded within a larger application framework, but still uses the iOS keyboard.
It isn't at all difficult to imagine applications outside of healthcare. This is a no-brainer for educational settings, especially in special education, where students with motor impairments would benefit tremendously from the predictive text. Similarly, teachers and school clinicians collecting data and documenting individual educational plans (IEPs) for students could take advantage of the capability for specialized language. Even outside of special education, with a focus increasingly on having students create content rather than merely consume it, SwiftKey's dramatic improvement to the usability of touchscreen-based keyboards is quite significant.
Engineering, manufacturing, quality assurance, and virtually any other market that uses specialized vocabulary and benefit from keeping users in the field highly mobile could easily tap into the potential of SwiftKey. In fact, SwiftKey has an SDK and partners with a variety of organizations to create applications around their predictive input engine.
One of the major drawbacks to touch has always been the lack of a keyboard. Even young people, who tend to be remarkably adept with this sort of thing, will usually turn to a full-blown PC for any serious writing, leaving their touch devices for texting and Facebook updates. However, SwiftKey is changing the way we think about "typing" and making increasingly ubiquitous touch devices much more usable for all of our computing needs.