AT&T announced today they have developed software that allows iPhones to recognize voice commands.  The software is called AT&T WATSON Speech Mashups based on AT&T’s WATSON technology. WATSON is hosted on the web allowing iPhone access to the software “without the need to install, configure, and manage speech recognition software and equipment” according to AT&T.

The WATSON Speech Mashups is based on established web programming models: SOA, REST, AJAX, JavaScript and JSON and requires a fast connection to AT&T's servers for native applications that don't include web code. WATSON isn’t capable of voice dialing or other direct speech recognition features. Sorry, iPhone users, you asked for voice dialing, instead you get this.  

AT&T was nice enough to post a video on the software in action, a man with prefect diction entering the business name and location into text fields by speaking at the appropriate time into an iPhone in a quiet office. The software doesn't create a completely hands-free iPhone, you still need your iPhone in front of your face.

In theory it seems nice but without a real world demonstration -- like on a city bus with a multitude of dialects speaking simultaneously, while the user tries find to their favorite pizza spot -- the software appears light years from useful real world applications. Still, AT&T claims the software is faster and more convenient than typing. If you're in a quiet office and speak with perfect inflection.