Voice-Enabled Wikis
Why would you want to voice-enable your wiki? In what form? Lets look at a use case first: The construction industry.
Construction companies can use wikis to track commerical and residential construction projects and to build up a knowledge base of best practices. In a
StructuredWiki, project tracking and knowledge sharing can be done with simple forms based wiki applications. The challenge is connectivity. People at the construction site do not have internet access, but eveyone has a cell phone. Wouldn't it be nice to be able record a voice message to an appropriate wiki page? For example, to leave a voicemail for project 1745? Or, to call the wiki and to find out who the internal expert is on pneumatic Daytona Elevators? Or, to call the wiki, find out the recent changes, and have selected content spoken to you by a pleasant synthetic voice?
We did such an integration together with
LignUp, Inc.
, a company that provides web services around voice technology. With their technology, web developers can quickly mashup voice and rich media into web applications.
Here is how the "record voicemail" feature works: Constructions workers are registered in TWiki; their homepages list the cell phone number and a voice pin number. There are two options to record a voicemail: Call a number, or click on a "record voicemail" button.
If you are away from a computer, you call a specific number, are greeted by a welcome message asking to enter a pin number. Once identified, the voice asks for a project ID. After a beep you record the message and press the
# key.
If you are connected, you simply click on a

button. A window is popping up, informing you that your phone will ring. You answer the cell phone, and are greeted with: "Hello TWiki Guest, please record your message for project 1021 after the beep, followed by the pound sign." After pressing the
# sign, you hear: "Thank you. Your message will be posted to project 1021."
In both cases, your voicemail is posted to the appropriate TWiki topic as a file attachment in
.wav format. In the wiki, you simply click on the message to hear it.
With text-to-speech technology it is also possible to send voice messages out of the wiki. You fill out a web form that has a field for the phone number, and a text area to type or paste a short message. This will call the number, and a computer generated voice is reading the message to the person.
There are many other use cases for voice enabled wikis, here are just a few ideas:
- Call the wiki to find out recent changes.
- Remote control a wiki: Navigate to topic, add content, move topic.
- Using a cell phone, record a message and post it as a blog.
- In a photo gallery, let people annotate photos.
- Send reminders for action items with outbound voicemail.
The possibilities are endless, only bound by imagination.
I will demo the current TWiki voice integration tomorrow at the
O'Reilly Emerging Telephony Conference
in San Francisco, as part of the workshop on
Utility Computing and the LignUp Voice Application Platform
.
--
Contributors: PeterThoeny - 26 Feb 2007 - from
Voice-Enabled Wikis blog post
Discussion
The voice technology of LignUp is primarily
SaaS, hence there is not much to open source. I contributed the
UploadToTWikiAddOn to the community, the add-on is used by LignUp's software to attach voice files to an intranet TWiki.
--
PeterThoeny - 27 Feb 2007
Hi Peter, the Speak-a-Blog might be interesting to you. Found here:
http://simoncrowfoot.livejournal.com/
I've LONG had on my wishlist a
TextAloudPlugin that would help me by porting recent changes & comments to my many synthetic voices that I use with
TextAloud
. A dream come true would be the ability to assign synthetic specific voices to certain individuals and to have all other comments read in round-robin voice selection.
For example, I would love to be able to tell my TWiki to read aloud to me all of the recent changes that I'm not yet caught up on... The friendly voice would say...
"On February 28th, on the topic of VoiceEnabledWikis, KeithHelfrich said ..."
Of course, the trick with all of this is to keep the noise to a minimum. The voice would have to know how to work around patches of code and unpronounceable text. And comments left by voicemail would need to be organizable in such a way as to prevent clutter.
--
KeithHelfrich - 28 Feb 2007
Thanks for the pointers. I did not know that the voice-to-text was that advanced.
--
PeterThoeny - 01 Mar 2007
I'm just off of an emergency conference call with one of my customers. It is a very large organization, with dozens of individual IT groups each of whom has their own tasks & processes that are necessary to complete in a time critical manner where sequence is important. The reason for the call was because a construction project had accidentally chopped through multiple communications lines that connect data centers in different cities.
The severity of the outage was significant enough to prevent even backup communications channels from functioning and email had faltered & come back within the duration of the call. Due to the time critical nature of the company's operations, decisions needed to be made swiftly and accurately to recover IT processing after restoration of the network.
My idea for the perfect TWiki
MashUp for
VoiceEnabledWikis came mostly from the fact that the amount of detail being processed during the conference was such that no single individual was capable of connecting all of the individual dots in real time. At the end of the call, the person who did a very fine job of summarizing the plan of action finished by saying
"did I miss anything ?"
My brainstorm for a
KillerApp mash-up is the following :
- TWikiPhoneBot - monitors conference calls and ports text to a real-time log (IRC ?)
- Conference call logs are auto-sync'd periodically to specified TWikiTopic
- Call participants pluck text from the call logs as necessary and simultaneously SynchroEdit a plan of action
Of course, if email was down, then connection to the wiki might also have been impacted. But in the end, it doesn't take an emergency to benefit from this type of collaboration. The fact is that these days I rarely attend meetings in person, but I spend a LOT of time in conference calls. The above
MashUp would be a great solution for all
TWikiBasedPhoneMeetings, large and small.
--
KeithHelfrich - 08 Mar 2007
I had another series of very busy conference calls over the weekend, this time for a data center move. In fact there were multiple, simultaneous conference calls, with people jumping back & forth between them depending on the group that they needed to speak with. Once again I thought about how GREAT a
TWikiPhoneBot would be, that could listen in & route all of the conversation to text in a twiki topic. This would help both in real-time and also for looking back on what transpired.
--
KeithHelfrich - 19 Mar 2007
Another company offering voice to text technology: SimulScribe,
http://www.simulscribe.com/
--
PeterThoeny - 19 Jul 2007
Yesterday I went to the
MashupCamp (what is a
MashUp?) and did a
session on voice enabled wikis
. Ideas that came up at the session:
- Take pictures with your cell phone, then voice annotate and post them directly into TWiki
- Integrate voice mail into a call tracking system (TWiki app): Record voice mail and route it to the proper ticket. Useful for support center with high volume.
- Integrate voice into insurance claim handling: On PDA or laptop, fill out form at accident site, take photos, annotate photos into voicemail, interview witness into voicemail, then sync to server. Possibly with the help of GoogleGears.
--
PeterThoeny - 19 Jul 2007