CWM Self-Service Chatbot
Self-Service applications allow end-users to manage some of the basic operations for their devices and accounts. This reduces the overall support cost to the enterprise as this prevents calls to the IT Support team. Further, end-users can get faster resolution to the issues they are facing using a self-service application.
Collaboration Workflow Manager (CWM) provides a REST API driven interface that can work with any business process application software. CWM supports management of various versions of Cisco Unified Communication Manager (CUCM), Cisco Unity Connection (CUC), various IP phone Models, Video end-points and it can synchronize data from enterprise LDAP Servers.
As part of this project, we developed a self-service ‘Webex Teams’ chatbot called ‘jeanie’ that provides a NLP interface to end-users and allows end-users to change their device settings for their IP phones. Some of the setting supported by the NLP interface include :
-
Setting ‘Call Forwarding’
-
Extension Mobility ‘PIN’
-
Speed Dial
-
Single-Number Reach (SNR)
Figure 1: Chatbot System flow
​​
The System flow above follows these steps:
​
-
The user enters a dialog in the 'Webex team' chat room/space with chatbot and starts a dialog. The text entered is then sent to the 'Webex teams Server'. { this happens for any chat, similar to 'Slack/Microsoft teams' etc. }
-
Once the 'Teams Server' receives the text for the particular space, it will call the 'web hook' of any registered application, at which point the Chatbot server receives the 'text' entered by the user.
-
The 'chatbot' then enters a 'dialog' mode and calls an NLP engine, in our case we used IBM Watson's Assistant feature and created dialog and training phrases.
-
Once the NLP dialog reaches a completed state then Watson returns an 'Intent' response, 'Intent' responses are then translated into an API the chatbot will invoke based on the understanding of the dialog with the user.
-
The CWM self-service API is invoked with appropriate parameters by the chatbot server; CWM in-turn performs the tasks by interfacing with Cisco Unified Communication Manager or Unity Connection, once the update action is taken, a response is provided to the end-user.
The messaging below shows the interaction needed to change the call forward setting on your desk phone.
Figure 2: Chatbot message exchange
Following diagram provides the steps involved in NLP give an 'under the hood' look at what NLP engine such as IBM Watson is doing during the exchange above
