Dr. Michael Housman, Chief Data Scientist at RapportBoost.AI, speaks to General Assembly about the problem with using bots for customer chat, and why they aren't advanced enough to convey your brand's message.
"Bots really struggle to engage with consumers in an authentic way. They’re not good at selling, they often don’t reflect the brand identity. For brands that want to convey a fun, fresh personality, bots feel robotic and people pick up on that. And finally, when we create bots using decision trees, they often repeat the same answer over and over. It doesn’t customize or personalize the interaction."
In a recent post, we tested the waters ourselves to confirm that a customer support chatbot isn't up to the task of delivering your brand message. When sitting down to make a purchase online, a whole host of distractions fight for your customers' attention, like advertisements, social media feeds, and the web pages of competitors. Numbers show that placing a chat window in the right place and prompting the customer at a crucial moment in the purchasing process is an effective way to increase sales and conversions.
The reason bots fail so frequently is that they use a technology called decision trees. You can think of decision trees as a programmed menu that includes possible questions and answers. The problem arises when a customer deviates from the script (as humans often do), at which point the bot reverts back to the main menu or provides a canned answer. Studies show that customer satisfaction is driven by wait time, so fielding your customers' concerns to an ill-equipped technology is the last thing you want to do to nurture brand loyalty.
Don't trust bots with your brand's customer chat. Crunch your live chat conversation data with the team of awesome Data Scientists at RapportBoost.AI.