RE: use case -- intelligent customer service

Debbie,

 

Thanks for the use case. It seems well grounded in a specific task. 

 

I'm approaching this from the point of view that there are no tools for
conversation management that are fundamentally different from the state full
directed dialog systems like VoiceXML. Clearly we need some new and
qualitatively different methodology to break out of the rut we're in. I like
to think about this problem from the perspective of "what if we had a new
method/functionality X? What can we build with this new tool?" 

 

We clearly need tools that will change "what" we can build. Much like 3-D
printing allows us to build parts that are impossible to cast, mold, or
machine. I think all of us in this group have been in the trenches fighting
with the issues of dialog. I like to think that we're beginning to peer at
the more sophisticated problem of conversation. I want to rise up out of the
plane (plain?) of the VoiceXML Flatland. I think a new perspective is in
order.

 

One method/function/paradigm I would like to have would be "generic"
behaviors. Some of those generics might be appropriate for things like
ordering carry out food, or describing products in an online catalog, or
explaining paintings in an art gallery, etc. there are general ways in which
humans approach these conversational problems. Humans can easily repurpose
the generics they use: if you have a job selling office supplies you could
easily modify and extend your techniques to sell medical supplies.

 

What kind of fundamental structure will help us accomplish this in an
automated way?

 

 

From: Deborah Dahl [mailto:dahl@conversational-technologies.com] 
Sent: Thursday, June 23, 2016 12:05 PM
To: public-voiceinteraction@w3.org
Subject: use case -- intelligent customer service

 

Here's an example of a use case for an enterprise virtual assistant. 

A user contacts an e-commerce company and talks to an intelligent enterprise
assistant about a problem with a product that the user has recently
purchased. The contact can be by phone, on a website, using an app, or by
text. The assistant knows about the item (when it was ordered, shipped,
warranty status, etc.) and about the user's previous interactions with the
company, by whatever channel. The user starts by asking for troubleshooting
help,  and the assistant helps with that, but then the user changes their
mind and decides to return the product. The assistant switches gears and
helps the user with the return (return authorization, shipping and refund,
for example)  and finding a replacement. If a GUI interface is available,
the assistant can show the user pictures of possible replacements and
explain how they differ from the original item.

  _____  

No virus found in this message.
Checked by AVG - www.avg.com
Version: 2016.0.7639 / Virus Database: 4598/12406 - Release Date: 06/12/16
Internal Virus Database is out of date.

No virus found in this message.
Checked by AVG - www.avg.com
Version: 2016.0.7640 / Virus Database: 4613/12478 - Release Date: 06/23/16