There’s no doubt that “chat bots,” or automated virtual assistants based on chat, are a hot technology. Since Facebook (News - Alert) opened its Messenger platform to businesses, those companies have used the technology to launch an astonishing 33,000 bots. Other platforms has followed: Kik now supports 19,000 bots, and Line another 10,000. Chatbots are being proposed as the answer for taking care of simple, quick-response customer needs, which leaves contact center agents more time to focus on complex customer demands and high-touch interactions.
But while chatbots continue to grow in popularity, customer support experts warn that there are limitations to the technology, and that is still has a lot of room for improvement. In a recent post for Fonolo (News - Alert), blogger Shai noted that too many expectations were placed on the technology, and companies need to be in a position to measure whether their bots are actually working and benefiting customer relationships (and not harming them).
“Now that we’re past the initial crush, it’s time for bots to start the slow climb to a mature technology. One of the clearest sign-posts along that climb is the metrics that are reported,” wrote the blogger. “We definitely have quantity. But do we have quality? In other words, are these bots accomplishing something?”
One way to measure the success of chatbots – at least those deployed through Facebook -- is to use the analytics support tools that the social media giant began offering late last year. Facebook Analytics for Apps designed to help businesses measure, understand, and optimize their customers' journey across apps and Web sites, including any chatbots that have been deployed from Facebook Messenger. The analytics tools allow users to keep track of a metric called “containment rate,” which is the number of calls for which the bot eliminated the need for a live agent conversation via telephone or any other channel.
Another way companies can measure the effectiveness of their chatbot deployment is third-party remote call monitoring, which can help companies skip over the hype of how many chatbot interactions occurred, and instead track the more meaningful metric of the percentage of chatbot interactions that were contained to an automated transaction.
Third-party remote call monitoring can add a “human element” to the oversight of automated customer support interactions. The service can check chatbot-based interactions to determine the reason for the interaction as well as the outcome to determine if the customer went away satisfied or grumbling and vowing to take business elsewhere (after all, how would you know otherwise?)
“In our rush to automate everything, push customers online and reduce expensive human contact, we must not forget the value these interactions bring to our business,” wrote Andrew Mutch, Chief Customer Officer of EMEA for BPA Quality in a recent blog post.
Automation isn’t a bad thing. In fact, customers sometimes prefer to help themselves with the aid of a self-service solution. Just be sure you’re not deploying automation and assuming that it’s working fine.
Edited by Stefania Viscusi