“Press one to continue this call in English” is probably what comes to mind for most when we think about computer-generated help, and like most, we
“Press one to continue this call in English” is probably what comes to mind for most when we think about computer-generated help, and like most, we desperately punch zero or say “operator” to get to that actual customer service representative as soon as we possibly can. Up until the advent of Alexa and Watson, using technology to aid in customer service has left more consumers feeling less than satisfied to say the least.
Today the promise of machine-learning technologies like artificial intelligence and chatbots are being likened to modern day “fortune telling,” and doing wonders in humanizing our computer interactions. A chatbot can sell things, offer up info, or just entertain. There are chatbots that update on the weather, inform political junkies on the latest news, recommend restaurants for foodies. A chatbot can live on a website or on a platform like Facebook Messenger. There are 30,000 chatbots on Facebook messenger alone! And the most popular messaging apps already have over 4 billion monthly active users!
Chatbots have the potential to literally change the conversation for brands. The opportunity for marketers to play a role in this space is tremendous. We’re always striving to be present at that critical moment of truth and this technology enables us to be exactly right there. But we must be thoughtful in our approach. Today’s connected-consumer is paying attention more intently than ever. They do not want to be sold to. Their interaction with chatbots must feel organic, conversational, and personal, or you will lose them at hello.
Last year in an effort to make commerce more conversational, Mastercard launched bots that both merchants and issuers can leverage to build their own digital payments solutions. Our global marketing team also built a chatbot and put it into Beta testing with consumers, and we got some very interesting learnings. We built ours to live on Facebook Messenger and launched it for a closed group of our prime target: moms. We watched how they interacted; they rated the bot; we iterated; they rated us again. We’re still fine-tuning our bot before we launch it wide, but already we got some very good learnings.
Make it personal:
The more relevant the conversation is, the better. No one needs a generic conversation with a robot. Consumers loved it when our chatbot gave them relevant, new ideas. Embedding an API from brands like foursquare can help you make a bot hyper local.
Make it snappy:
Shorter is sweeter. People are busy—often using multiple devices simultaneously. A long-winded robot is not anybody’s friend.
Make it delightful:
A chatbot needs personality. One of the most downloaded chatbots is a bot that reports the daily weather called Poncho. Poncho has a big, memorable personality that is clearly defined: It’s a cat that lives in Brooklyn, and tells jokes along with the weather. Your bot needs to delight.
Make it human:
A bot should speak like a human—our bot copy was written by a team of living, breathing writers. But don’t bother hiding the fact that your bot is a bot. Consumers are intrigued by talking to a robot anyway.