Chatbots are reigning in a new era of communication. They emulate human-like interface and engage customers conversationally. Customers can now place orders and get their issues resolved unobtrusively from someone who “listens” to them, rather than navigating menus and screens.
How can organizations ensure a successful transition to chatbots? Let’s take a look at five key factors to consider when deploying chatbots.
Assigning the right issues to a bot
Chatbots can address a specific problem set, and address them well. The challenge for organizations is to find that sweet spot where chatbots have the right and highest impact. Chatbots are good at addressing routine issues. A customer who has forgotten his password would prefer to talk to a bot and reset the password quickly rather than navigate through endless IVR menus and listening to tedious recorded messages. As bots are not human, they are not good at addressing issues that involve emotions, empathy and out-of-the-box solutions. Would you let a bot talk to your prime customer who has lost her order meant as a gift to her parents celebrating their 50th anniversary? This requires sensitivity and rushing your best delivery boy with a replacement.
We should not lose sight of the fact that chatbots are ultimately meant to enhance customer experience. Hence the need to track the customer experience as a moving target. The right metrics and measurement tools need to evolve to gauge chatbots talking to customers. How can we sense if the customer is happy or getting agitated? A human could sense this through the tone of voice. Can we train chatbots to sense the sentiment with choice of words, for instance? Sentiment analysis is critical to the success of chatbots.
Broadly, there are two types of chatbots – conversational and transactional. While enterprise chatbots are oriented towards transactions, it would be good to add a touch of personality with the ability to hold conversations. Amazon customer support representatives are known to converse on a wide variety of topics, including Game of Thrones and Thor. While it could take some time to build this level of conversational sophistication and humour, we could have bots detect where the customer is from, for example San Francisco, and open the dialogue with “Awful weather we are having today.”
A well-timed, smooth cutover to a human when the bot is not helping can do wonders to customer satisfaction. A non-intrusive sentiment monitoring tool can trigger a handover if or when the situation is getting out of hand. The bot could sense the customer dissatisfaction and initiate handover saying “My manager will talk to you now,” or a human could interrupt proactively and take over the conversation.
Insightful and intelligent analytics are essential for technology to be effective in any field. Chatbots are no exception. The right chatbot analytics platform can keep track of engagement, retention, sentiment, bottlenecks and provide summarized reports on transcripts.
With chatbots, it is important for an organization to pick the battles. Chatbots cannot solve all the problems, but if used appropriately, they can be potent tools to automate prosaic tasks, thus freeing the workforce for more strategic activities. Some of the key factors to derive the best of chatbots are discussed here. What do you think? How can we get more out of our chatbot friends?