Can Chatbots Fully Replace Humans? Not Yet

Customer service departments in all industries are increasing their use of chatbots, and we will see usage rise even higher over the next year as companies continue to pilot or launch their own versions of these rules-based digital assistants. Forrester defines chatbots as autonomous applications that help users complete tasks through conversation.

While Forrester data reveals that 60 percent of U.S. online adults already use online messaging, voice, or video chat services, there are challenges to even more widespread adoption. We reached out to online consumers through our ConsumerVoices qualitative online community to better understand consumer impressions of chatbots, and we found that our respondents had a difficult time identifying clear benefits to interacting with them. Many prefer to communicate with representatives who can show real empathy, address more complex needs, and offer them assurance.

We saw similar sentiments across different ages, though younger groups more readily came up with potential use cases and benefits. For example, a male consumer in the 25- to 34-year-old age group said, “I can see a chatbot being helpful for simpler interactions, like asking when something will be delivered, or money will be transferred over, or about scheduling. Simple things like that seem fine to me.”

A female consumer in the 65-and-over group agreed with the utility of chatbots for basic tasks but felt more complex requests were beyond them: “It might be helpful to use a chatbot for straightforward questions and/or commands, like checking balances, but I think more detailed questions and commands may be challenging for a chatbot to handle because of the difficulty in deciphering and/or understanding the variety of commands that the chatbot might get.”

It’s not just consumers who are thinking about chatbots’ link to customer experience. Earlier this month I attended the Qual360 2017 conference in Washington, D.C., where industry executives highlighted their experiences with chatbots from a research methodology perspective, while also being candid about their shortcomings. We saw some overlap between their sentiments and those of Forrester’s online community members. Paul Hudson of FlexMR and Tom de Ruyck of InSites Consulting focused on how chatbots are very good at capturing large amounts of information and addressing simple needs, using rules-based logic and natural language processing capabilities. Yet Ross McLean of Over the Shoulder noted that chatbots are unable to truly empathize with the respondent or customer—while they can mimic human empathy, they can’t pick up on nuances in conversation that fall outside their rules-based logic.

This leads me to a few points of guidance for businesses considering launching chatbot services:

• Make sure your customers know that they’re speaking with chatbots and are aware of their limitations. A big barrier to customers using chatbots is that they are often unsure whether the bot will be able to meet their specific needs. Be specific about what it can and cannot do. This will minimize any negative sentiment or fallout resulting from feeling misled.

• Enable chatbots to identify when customers need to speak to live representatives as early in conversations as possible—and make the transfer seamless. Provide customer service reps with transcripts of conversations with bots to minimize any need for customers to repeat information. This will ensure that customers have a great experience in terms of ease of use (and speed of resolution), effectiveness in meeting their needs, and generating an overall positive emotional sentiment.

While chatbots can serve as an efficient way to offer customers solutions to their problems, their future success will depend on how thoughtfully companies leverage them to meet customers’ needs. After all, delivering a great customer experience depends on it.

---

Kristopher Arcand is a data analyst at Forrester focusing on qualitative insights. Follow him on Twitter at @kristophersays.