Introducing the Xurrent Virtual Agent

Many people have been looking forward to see how Xurrent’s development team would incorporate native chatbot functionality in the Xurrent service.  Today, everyone can see how the Xurrent’s service-oriented data structure and tightly integrated KCS-based knowledge management functionality have made it possible for Xurrent customers to deploy a fully functional virtual agent in less than 5 minutes.  Best of all, this is a free chatbot, which is makes it unique in the enterprise service management industry.

Rather than requiring organizations to spend weeks training a bot for every knowledge article and standard request, Xurrent’s virtual agent already has enough information to understand the context of the conversations it has with the enterprise employees it supports.  Most importantly, it already knows which services someone is covered for by an active SLA.  This dramatically improves the relevance of the virtual agent’s responses when it uses natural language processing to propose the knowledge articles and standard requests that are most likely to help the user.

Before offering any knowledge articles or standard requests, though, the virtual agent first determines whether the user is trying to accomplish something that it can assist the user with without having to look through the knowledge base or request templates.  For example, the user may want to change the picture in her Xurrent profile.

After submitting this information, the virtual agent will respond by offering the user a link to the section in Xurrent that allows the user to update her profile.

There are already several other things that the virtual agent is able to assist end users with.  For example, it can look up the items in the user’s inbox.

Specialists can also use the virtual agent in a similar way to check their inbox.  The bot will always show the three items that are at the top of the specialist’s inbox.

Pressing on one of these three inbox items, opens it so that the specialist can start to work on it.  The specialist can press the fourth option to open the complete ‘My Inbox’ view.

But when a user is not looking for help with one of these standard activities, the virtual agent will look through the knowledge base for articles that might allow the user to help him- or herself.

If a relevant knowledge article cannot be found, the virtual agent attempts to find a request template (i.e. a standard request) that may help the user.

By first looking for ways to assist the user with standard activities in Self Service, or else trying to find a relevant knowledge article, the virtual agent helps to reduce the number of requests that get submitted to the support organization. And when a suitable knowledge article cannot be found, the agent continues by looking up a request template that will collect all the information from the user that the support organization needs to quickly complete (or automate the fulfillment of) the user’s request.

But if the virtual agent is also not able to offer a useful request template, it will offer to assist the user with the registration of a new request.

If the ‘Yes’ option is pressed, the virtual agent will guide the user through the steps that will allow it to apply the correct service instance to the new request, select the most likely impact level and assign it to right team.

To activate the virtual agent, administrators and designers can go to the ‘Self Service Design’ section of the Settings console.  There they can insert the {{virtual_agent}} widget in the Homepage HTML.

In the same ‘Self Service Design’ section, it is also possible to inject CSS to change the look of the agent and to style the presentation of the conversations in Xurrent Self Service.

The virtual agent currently speaks English, French and Dutch.  More languages will be added in the coming weeks.