We’re still not at the point that we can have full conversations with machines but the technology behind voice recognition is getting better and better, becoming less of a limited tool and more of a real enabler. It’s something we’re seeing today in Nuance’s new Dragon Go app, a successor to Dragon Search, which dives deeper into voice activated searches by pulling up a host of websites and apps that complete a user’s query.
It’s part of the natural evolution of voice technology, which is making the necessary move from just recognizing a user’s words to completing more and more actions. This is where real promise of voice technology lies, something rivals like Vlingo and other competitors like Microsoft (s msft) and Google (s goog) have also pursued.
With Dragon Go, which is available in Apple’s App Store, Nuance is trying to take users directly to what they want to do, rather than a list of results, which is how Dragon Search operated. It uses natural language processing to understand intent and then pairs it up with the right set of options. Now, many have already gotten a taste of some of this goodness through Android Voice Actions or through Microsoft’s voice integration in Windows Phone 7, Bing and Kinect. And Vlingo and Siri, the personal assistant app bought by Apple (s aapl), which is also powered by Nuance, also does some of this.
But Dragon Go offers a more complete set of sources and does some pretty impressive work in figuring out what a user is really looking to do. It has 180 content providers that it’s working with, everything from Fandango and Live Nation (s lyv) for movie and concert tickets to Pandora (s p) and Last.fm for music, Amazon (s amzn), Milo and eBay (s ebay) for product searches and OpenTable for reservations. There’s also reliable resources like Google, Twitter, Wikipedia and YouTube in there. You just say what you want and Dragon Go speeds you to the right option. When you get results, there’s also a carousel of tabs featuring other sources that might also be relevant. So for instance, you can often see what people on Twitter are saying about your search topic.
I found it to be pretty accurate though it still needs some work. Looking for a “sport coat” pulled up Amazon and Milo, which allowed me to see where I could buy specific sport coats online and at stores nearby. Saying “Play Bruno Mars on Pandora” took me to my Pandora app. Asking for directions to the High Line Park in New York actually pulled me out of the app and into Google Maps, which was pre-populated with directions from my location. I was also able to look up friends on Facebook, though people with common names were harder to find.
Again, it’s not perfect. Trying to find information on the movie Captain America didn’t take me to IMDb, it showed me news results on America from CNN. Looking up showtimes for Transformers 3 took me to a media tab that showed me options for downloading Transformers TV shows from iTunes.
But overall, I think this shows you where voice activation is going. We need to strip out some of the extra steps in voice searches and make it more intuitive. Jumping directly into apps and websites speeds up the process a lot when the results are what you’re looking for. And if it’s done well with even more sources added, I can see Dragon Go being a major resource for mobile users. Go still needs to improve and show that it can stay one step ahead of its users. But it’s moving in the right direction and demonstrating the bigger promise of voice technology.