Chatbots and virtual assistants are magical. 44% of US consumers prefer interacting with chat bots over humans for customer relations. There’s a simple reason why – virtual assistants can solve problems easily without any effort for the consumer, just like magic does in Harry Potter’s world.
That’s why the Industry saw a virtual agent usage growth from 28% in 2012 to 37% in 2016. Brands are adopting this trend more and more. Yet chatbots and virtual assistants don’t seem to meet expectations for brands and users. “Chatbots were oversold and not that great when they first launched earlier this year,” said Facebook Messenger VP David A. Marcus during the TechCrunch Disrupt Conference. The success rate is low – Facebook reported that its chatbots failed 70% of the time — and some early adopters have dropped their chatbots due to disappointing performance.
Some vendors tried to ride on the trend dragon and created chatbots that are very narrow minded, don’t understand what the user wants, and deliver wrong answers or no help at all. Thanks to these experiences, consumer perceptions of chat bots have revealed that just 30% of consumers are confident in a chat bot’s ability to successfully help them, according to Nuance chatbot research.
The magic is missing
The missing component is the magic of sophisticated Natural Language Understanding technology, algorithms that help create a base model of a virtual assistant and then leverage humans to support the first steps. This and the continuous learning from conversations happening the moment the VA is deployed makes your customer experience with a virtual assistant truly wondrous.
Just like a complicated potion, an enterprise-ready virtual assistant needs several ingredients
It all starts with collecting information to create a base model. Machine-learning algorithms help by scanning through all the data and finding the important parts. A Professional Services team then ensures that the first version of the Virtual Assistant meets the brand’s criteria. After being let loose in the world, the virtual assistant still receives assistance from its human counterparts. Having a real person help when the virtual assistant gets stuck ensures a continuous optimization and learning loop as the insights are fed back into the system.
The result is a sophisticated and confident chatbot that understands what the user wants, can ask clarifying questions if needed and present a solution to the customer resulting in high customer satisfaction scores. A human agent will always be there, coaching the virtual assistant if needed, just like Hermione and Ron helped Harry during every challenge.
Read part II of the “Learn and grow smarter” series here.