Chatbots Falling Short of Expectations

Interesting stats.  Chatbots clearly need to be better, more empathetic, more contextually focused, with better memories and links to required and up to date metadata.   When required with links to sensors as part of an IOT digital twins network than can remotely test and interact with context.  A lot of the pieces are possible, but not connected well enough as yet. 

Where are chatbots falling short of consumer expectations?  by Tom Ryan in Retailwire, with further expert comment.

According to a survey from Uberall, 80 percent of consumers who have interacted with an AI chatbot say the experience was generally positive.

When asked what elements of chatbots need improvement, 43 percent said their accuracy in understanding what customers are asking or looking for. Other areas cited included the ability of customers to get “a human customer rep involved where needed,” 27 percent; and “ability to hold a more human-sounding, natural conversation,” 19 percent.

Drift’s “2018 State of Chatbots” report found different inhibitors to chatbot adoption when it asked 1,000 consumers, “What would stop you from using a chatbot?” The top answer was, “I’d prefer to deal with a real-life assistant,” 43 percent; followed by, “I’d worry about it making a mistake,” 30 percent; “If I could only access it through Facebook,” 27 percent; “I’d prefer to use a normal website,” 26 percent; and “If it wasn’t able to ‘chat’ in a friendly manner,” 24 percent. ... "