Using Twitter as a Real Time Channel for Chat

Anana has been quite busy over the summer deploying and realising Social Media solutions on Genesys. During the course of these projects we have enjoyed working closely with our customers on use cases, and scenario’s in which they would like to use the technology.  A more recent development is the concept of treating social media interactions as conversations rather than transactions in the form of individual interactions.

What we have certainly noted with some of our clients is that their customers actually treat Twitter differently than originally intended.  Rather than sending in a single request, that recognises the limited real-estate (character count) in a standard tweet, a user may type 3 tweets in rapid succession that target the Enterprise in rapid succession.  In order for these tweets to arrive into Customer Operations in a meaningful way we had to think of ways to deal with the scenario.

The standard, out of the box capability in Genesys (as of Genesys 8.1.1)  is ‘interaction based’. In other words, each tweet is treated distinctly, and is routed and targetted to a skilled agent distinctly.  This is fine in any use case where the customer does not Power Tweet, or indeed, use a Twitter LONG web-service that turns a long utterance into a series of tweets.

If a user sends, say, 3 tweets in rapid succession (or a longtweet that is turned into 3 of appropriate length), and these 3 tweets are picked up by the social media server in the next POLL interval, then Tweet 1 may go to Agent 1, Tweet 2 may go to Agent 3, and Tweet 3 may go to Agent 2.  As a whole, the ‘conversation’ could still be reviewed by any of these targeted agents by using the ‘interaction history’. At best though, this may result in the FIRST tweet in thread being the owned tweet, and the other two agents having to go through consultation with each other to FORWARD the other tweets that formed part of that conversation directly to the now, newly owning interaction Agent.  Sound complex?  It can be!  In simple terms, why target 3 agents with a single LONG Tweet that spans 3 interactions, when it could be easier to simply target all 3 of them at the same time to a single agent, in the right order.  That consideration gave us time to think about, what will happen if a few seconds later the subscriber tweets again?  Where should that tweet go?  A new agent, or take the first agent as the preferred agent and send it straight into that interaction?  Sounds a bit like Web Chat does it not?

Genesys has done an incredible job of offering the maturity in social engagement that is evident in the 8.1 release and we are working closely on the types of use cases cited above with Genesys Product Management, Engineering across eServices and the Interaction Workspace on identifying and documenting these new use cases so that they can be captured and become part of the roadmap where approved.

To test out the use case scenario we are discussing, our Platform Practice Team went ahead and built a new Social Media driver for Genesys based on the STREAMING API, which allows us to listen in real-time for any activity on our ABC Company Wall (in the Anana Briefing Centre) and immediately target that to an agent.  Once targetted, any subsequent tweets from that subcriber in timeframe X would be targetted to the same agent; effectively treating the interaction the same way the use-cases for Web Chat have already proven themselves against.

The video below shows the Anana manipulation of a new Driver for the Social Media Server, with Real Time Streaming API, targetted to the Genesys Strategy as a Web-Chat.  This is a basic proof of concept and would require a lot more work.  For now, it shows how useful it could be in certain use-case scenario’s where your customers want to treat social media streams of communication as near real time channels.  Please turn on the CC (Closed Captions) option, and go full screen for best viewing.  Best viewed in HD format of course if your bandwidth will allow it.

Do you think a conversational mode for interaction management in social engagement has a role to play in Customer Services?  Our clients currently think so!