How then can the customer service industry reconcile consumers’ stated preferences with chatbots’ inexorable and continued rise? How exactly should a brand announce that it’s launching a public-facing chatbot for consumers who claim not to want one?
Chatbots are undeniably useful
There are now over 100,000 chatbots on Facebook Messenger and, when polled, consumers do offer praise — they appreciate that bots are available 24 hours with zero wait time.
Consumers also needn't give up human interactions to use chatbots, which work best when they augment agents, not replace them. They’re great at routine tasks such as welcoming customers, diagnosing their issues, answering common questions, and escalating the conversation to a human if the request is too complex. Most requests are not complex, however: Forrester found chatbots are well-suited to handle as much as 80 percent of inbound queries. And as we’ve written before, if chatbots help humans provide faster support, consumers quickly learn to love chatbots.
Brands should not be frightened of using chatbots to help improve the CX experience. Therefore they should communicate the benefits. When announcing your chatbot to the world, then, announce it as you would any new initiative to better support customers. Remind them that it supplements humans, not replaces them. Then, give your bot the greatest chance of success by doing three things:
Be clear about the chatbot’s limitations
When brands don’t educate customers about a chatbot’s limitations, consumers treat the bot as either exceptionally smart or exceptionally dumb. They either expect too much and grow frustrated, or they think the bot can’t possibly help them and mash buttons to reach an agent.
To help consumers use your bot, label your bot as a bot and be clear about what it can and cannot do. Provide users with simple use cases where they can try it out, and program the bot to guide users with multiple choice prompts such as “You can ask me things like ‘access my account’ or ‘billing dispute.’”
The most successful bots are clearly marked so there’s no mistaking them. Take the Dutch airline KLM’s chatbot Blue Bot, for instance, which answers millions of customer queries every year. It has “Bot” in its name and introduces itself in
Source: Mobile Marketing Magazine
KLM has also launched Blue Bot on Facebook Messenger, Google Assistant, and Google Home. Just like human agents, KLM’s consumers can find Blue Bot in multiple places and learn to treat it like they would any other customer resource.
Connect it to your CRM
Chatbots thrive on context and, often, the difficulty of programming them to have authentic-seeming conversations is that they don’t have enough of it. For example, if an airline chatbot doesn’t know that the customer behind an inbound Twitter DM has a reason to be upset, it may ask insensitive questions and amplify their frustration. But if it knows that that user’s handle belongs to a customer whose flight is delayed, it’s likely to send more relevant messages. Whatever your industry, make sure your chatbot is connected to your CRM and has the foresight of unified customer information.
Never stop testing
Chatbots may have made it through the boom-and-bust hype cycle, but they’re still not a fully established technology. They require plenty of testing and companies should view their chatbots as experiments, where the development team is constantly learning, taking feedback from both customers and support agents, and improving the service until it delivers consistently satisfying interactions.
Consumers may prefer to talk to a person, but what they really crave are swift and efficient resolutions. When brands are upfront about their reasons for launching a chatbot and transparent about what it can and cannot do, consumers think of it as no different than a mobile app – fast, great for routine support, and only a swipe away.