“Computer: Open Outlook. New message to steveb@microsoft.com. Subject: Voice interface. Body: Steve, Are we really this close to a voice-powered user interface? Jason. Computer: Send message.”

If Windows had a voice interface then that’s how easy I would like it to be to send a short e-mail, and that’s the message that I would have sent to Steve Ballmer this morning (whether Steve would have personally replied is a different issue).

Yesterday, Microsoft announced that is was buying Tellme Networks, a hot player in voice-powered software, and the deal was driven by Steve Ballmer himself, according to both News.com and GigaOm. Speech recognition software has been a pet project at Microsoft for a long time (Bill Gates has been predicting that usable speech recognition was right around the corner for the past decade). Now Ballmer wants Microsoft to pair its own developments in speech recognition with Tellme, which is a leader in voice-powered customer service solutions (yes, that means those automated phone systems where you talk to a computer with voice prompts) and voice-powered search on mobile phones.

In addition to the obvious Tellme integration that Microsoft can do in its many applications and platforms (most notably, Microsoft Outlook and Windows Mobile), I have to think that part of the motivation for the Tellme acquisition is an attempt to leapfrog Google in the race for voice-powered search, especially on cell phones.

The problem is we all have the expectation of a Star Trek-like experience. That’s the vision of success for speech recognition. ZDNet’s Larry Dignan shares my skepticism about how close we really are to that kind of seemless speech recognition experience.