The Importance of Zig - Update on Microsoft’s Tellme Reorganization
Last week I blogged about Mike McCue’s planned departure at the end of June from Tellme and Microsoft’s plans to fold Tellme and all other things speech into one division under Zig Serafin. This includes Tellme, a research team developing text-to-speech in Beijing, and the Redmond-based Speech Components group. I’ve since talked to Microsoft and a few others, so here is a brief update on the reorganization of speech at Microsoft.
Something I sort of glossed over, but thought about after I posted, was the importance of this new group having Zig Serafin at the helm. For those of you who don’t know, Zig has spent a decade at Microsoft and is the co-founder of the Unified Communications (UC) group there. In my opinion, this is extremely important. From an industry standpoint, speech technologies have been gaining traction in pockets or silos for two decades. Speech technologies first showed up with applications such as speech-enabled IVR, and voice-activated dialing, then with directory search, and dictation. In the past two years, as unified communications evolved, speech started to proliferate everywhere from mobility applications to more complex natural language voice search. I’m not forgetting other uses such as translation either, but we have seen a more concentrated melding of multiple types of speech technologies such as speech-to-text for mobile users sending messages and in applications such as unified messaging (a component of UC) as part of the UC push.
Making Zig the leader of the combined speech groups, given the results of his efforts in UC, is a big statement as to the commitment Microsoft has made, and will continue making into the development and deployment of speech technologies across applications and platforms. For those of you who have listened to keynotes given by Gates and Ballmer in the past, this isn’t new, but it is a reminder of how far Microsoft has come in speech and its commitment to it.
As part of this, Microsoft is creating a speech center of excellence. This is really to focus their efforts on developing speech across business and consumer products. Tellme’s research and installed base of products is a big part of this as they have vast experience in many of the recipients of speech research including IVR, multi-modal applications, directory assistance, etc.
Finally, Tellme is staying put. I mean that in a physical and philosophical way; maintaining their office in Mountain View and their brand. I had hoped this was the case, and had thought so as the spring Tellme announcement really showed how far Microsoft and Tellme’s product and research groups have come working together. Folding all of them into the same division just breaks down any work barriers further. Also, that joint effort is producing a broadening of speech deployments across more products as was shown by Tellme’s Windows Mobile 6.5 announcement. This deeper melding into one group reminds me a lot of Active Voice folding itself more tightly into NEC, which I blogged about in April.

Hi there,
I think you have your timeline wrong. Speech recognition was developed first for desktop computers, starting with Dragon Dictate in the mid-80s. Commercial IVRs did not come into being until the mid-90s.
reddot
My mistake. I’ve been tracking speech technologies since the mid 80’s. I just blogged to fast. I was thinking more about proliferation and consumer awareness. We have had a lot of speech technologies with slow adoption rates, but I believe that contact centers and now UC are accelerating that.
Nancy
PS - I just loaded and tried the latest Dragon version and its light years ahead of where it was a couple of years ago.