Customer Service Chatbots: A Seamless GenAI Solution or a Risky Feature?

Advances in AI capabilities have vastly improved the customer service chatbot experience in the past couple of years. So why do so many consumers still avoid them? A 2023 Forrester survey found that “16% of consumers said they often used chatbots, and over a third said they never use them.” It’s not the chat format itself—59% of North American online shoppers prefer live chat to calling customer service, according to a recent international survey of consumer attitudes on ecommerce, fraud, and customer experience. That number of customer service chat fans is significantly higher for younger consumers than for baby boomer shoppers who tend to prefer talking with a representative on the phone.

by Rick Sunzeri

Even for customers who are comfortable chatting with a bot, the results aren’t always helpful, because the technology is still developing–and because companies sometimes have expectations that are too high for nascent tech. That can translate to lower customer satisfaction and reduce the return on chatbot investment. For a chatbot to elevate CX, businesses need to avoid some common pitfalls that undermine their investment in chatbot technology. 

Who uses live chat now?

The preference for chatting versus calling for help seems to follow generational lines, per the consumer attitudes survey data: 70% of Gen Z and 74% of Millennials opt for chat over calls. Some of this preference may be driven by the well-documented anxiety that many younger adults feel about making voice calls–after all, they grew up in an era of instant messages and texting. But 72% of Gen X adults also prefer chat over calls, which indicates that convenience may also play a part in the preference for chatting. 

The one generational group that’s less inclined to chat is baby boomers. Just 58% choose chat over a voice call, reflecting long standing habits and preferences.

Ideally, retailers would have chatbots that handle most inquiries easily, so shoppers can make their purchases and the company’s customer service representatives have more time to handle calls from customers who’d rather speak with someone. However, other survey data indicates that chatbots are missing the mark. One study of customer service chatbot engagements found that 66% of those interactions earned a 1 out of 5 score from dissatisfied users.

That’s a big problem in the moment for customers who can’t get answers, and in the long run for businesses that want to retain customers. 72% of U.S. and Canadian shoppers say they won’t return to an online store after a bad customer support experience. The number rises to 77% for baby boomer shoppers, while the youngest adult consumers, Gen Z are slightly more forgiving: only 66% will abandon a site after a bad support experience. 

How chatbots can go wrong

Multiple experts agree that the problem with customer service chatbots boils down to poor design. Common issues with today’s crop of chatbots include:

They sometimes give wrong answers. One airline was recently found financially liable for incorrect advice that its chatbot provided to a customer about bereavement fares. Not only did the issue create problems for the customer when he was already dealing with the loss of a loved one, the resulting tribunal case generated negative press for the company.

Punting to the FAQ page. Sending a user to the site’s FAQ page forces them to sort through all the content there–and they may have connected with the chatbot because the FAQ didn’t answer their question first.

Difficulty in reaching a human. When the chatbot isn’t getting the job done, users need to be able to quickly reach a live person, but many chatbots don’t make that easy. 

Going wildly off-script. Without the right guardrails in place, a natural language chatbot may say things that generate headlines for all the wrong reasons. Case in point: A package carrier chatbot recently told a customer that its company was “the worst delivery firm in the world.”

Designing the ideal customer service chat experience

Strategic design and realistic goals can prevent or minimize these issues and create a better customer experience, even if the use cases for the chatbot end up being more limited than initially planned. 

Manage expectations from the start. It takes a while for AI chatbots to learn the particulars of a company’s customer requests and the right answers to those questions. It’s better to start small and have, say, an effective chatbot for tracking deliveries and returns, than to offer a chatbot that promises to answer all customer inquiries but can’t give accurate information.

Do the work for the customer. Rather than directing customers to your FAQ or policies pages, train the chatbot on the information they contain. That will save customers time and avoid frustrating them.

Connect to a human quickly. While chatbots can help reduce the workload for human agents, they shouldn’t serve as gatekeepers. If a customer asks to talk with a person, the chatbot should connect them. Even better, customers who’ve tried to resolve their issue with the bot should be fast-tracked for human help rather than joining the end of the call queue.

Set limits on topics the chatbot can discuss. Have your chatbot stick to business to avoid faux pas that generate negative publicity or offend customers.

Ask for feedback. Your chatbot’s goal should be continuous improvement, and asking for feedback at the end of each chat is a quick way to gather the information needed to make those changes.

Putting it all together 

Launching a customer service chatbot is exciting, but make sure that excitement is balanced with thoughtful design and realistic expectations about what your chatbot can do for your customers, especially at first. As you gather customer feedback and other data on chatbot performance, you can expand the chatbot’s scope to include more use cases, with more confidence that the bot will handle them properly. And that’s what customers want from your chatbot–correct and helpful answers to their questions, so they can continue shopping on your site.

Original article at: https://thecustomer.net/customer-service-chatbots-a-seamless-genai-solution-or-a-risky-feature/