Microsoft Shows Off Skype Translator
An integration with AI-driven systems shows some of the exciting potential in store for enterprise communications.
This year at the Lync Conference and again at Enterprise Connect, Gurdeep Singh Pall, who leads Microsoft's Skype/Lync business unit, promised that Micrsoft was going to bring much deeper intelligence to bear on communications systems. He talked in terms of context, machine learning, and other next-gen type of capabilities of the sort he'd worked on during his hiatus with Microsoft's Bing search unit.
Yesterday, Pall and Microsoft CEO Satya Nadella offered a glimpse of what he was talking about when they demo-ed Skype Translator, an early iteration of a machine-driven near-real-time translation function that the company says will debut by the end of this year, as a beta app.
Translation is obviously an exciting application--the idea of being able to do business internationally without having to learn every language of every region you want to serve--or having to find reliable translators everywhere--would be enormously appealing to business. No doubt, this isn't an app you want to be an early adopter of, for production systems--the potential for career-limiting mistranslation will be on everyone's mind. But if the example of IBM's Watson and similar systems is any guide, there's no reason to believe that this capability won't improve fast and couldn't perform adequately for business purposes in the not-too-distant future.
And even more significant is what this application suggests about the convergence of artificial intelligence and communications. Simultaneous translation of real-time business critical communications is practically the definition of a high-risk endeavor. There are lots of things that AI engines could enable behind communications systems that are much lower risk but potentially quite productivity-enhancing.
Think of systems that could offer you screen pops for document sharing, prompted by key words as you speak: "Tom, did you get the financial projections?" is all you'd have to say for that document to pop up on Tom's screen. (Or at least a link to that document--obviously Tom doesn't want a fireworks display of screen pops exploding across his screen whenever you say the right (or wrong) words. Those are UI details anyway, something the designers will be figuring out while the engineers are making the systems work right.
That's just an example I thought of off the top of my head--you undoubtedly can reel off your own.
This all should be good news to people who build and maintain communications systems, because it makes those systems incredibly more valuable, more useful. Translation is the best example of this. Right now, why would you call a colleague on the phone if the two of you don't share a common language? It'd be pointless. But accurate simulataneous translation makes that call not only possible, but potentially valuable.
We're in the midst of some big changes for enterprise communications, and some of that can be scary. But it's also full of potential.