Customer service and support leaders are embracing generative artificial intelligence-enabled chatbots and voicebots to handle customer interactions. However, a poorly designed interaction can reinforce customer perceptions of AI's limitations and negatively impact customer experience.
Customer service leaders can improve automated interactions by making intentional choices about conversation design. Conversation design is a field that combines psychology, user experience, and linguistics to create more natural interactions between humans and AI. Conversation design decisions range from how bots greet customers to how bots respond to customers in distress.
These decisions about how the natural language interaction between customer and AI plays out should not be an afterthought. Too often they are outsourced to vendors and product teams rather than driven by service leaders who are closest to customer needs.
By making design decisions that prioritize customer needs during AI interactions, organizations can significantly improve customers' success in self-service.
Below, we share five principles of good conversational design that customer service and support leaders can use to design new bots or transform bad bots into better ones.
1. Transparency
Companies should design bots to look and act like bots, not hyper-realistic humans. After all, if customers discover that they are talking to a bot when they thought they were talking to a human, they can feel deceived, which drives mistrust. Customer service and support leaders can evaluate transparency in their organizations' bot-guided customer journeys by asking questions like, "Is it clear to customers they are engaging with a bot?" and "Does the voicebot mimic a real human voice?"
Leaders can improve transparency by disclosing to customers that they are speaking to a bot rather than a human in the bot's initial greeting. Rather than striving to imitate a real human as closely as possible, companies should strive for a bot with a consistent persona that aligns with brand image.
2. Clarity
Customers should know what to expect from the outset of an automated interaction. That means organizations have to explicitly communicate their bots' capabilities. By explaining to customers what a particular bot can help them accomplish, organizations can frame customer expectations appropriately and avoid disappointment.
To design automated experiences that prioritize clarity, service and support leaders should update the bot's greetings so that its scope is clearly expressed. It's also important to track customers' attempts to resolve issues outside the bot's current scope. That way, leaders can better frame the bot's capabilities while working to expand them down the line.
3. Depth
Chabots that provide specific answers, including the why behind them, reassure customers that their needs are fully understood.
If customers regularly contact human agents to double check the bots' answers, that can be a signal to leaders that the depth of the bots' responses don't match customers' needs.
A response that is either too long or too short can lead to customer misgivings. Instead, the length of the response should mirror what you'd expect in a natural language conversation. That way, it demonstrates sufficient comprehension of a user query and provides specific and relevant information in response.
4. Understanding
An important way to build trust during an automated conversation is to design a bot that echoes back customer intent at the outset of the conversation. This repetition of the customer need helps the customer feel heard and also provides a chance for the bot to clarify.
5. Humility
Service and support leaders should recognize their bots' limits and know when to guide customers from self-service to assisted service. The goal should be to provide seamless handovers to human representatives rather than to trap customers in ineffective self-service loops.
By escalating issues to human agents when necessary, organizations can drive long-term customer engagement and loyalty.
Wondering where to start? When putting these five principles of good conversation design into practice, start with the principle that will have the greatest impact on intended self-service outcomes.
Kim Hedlin is a senior principal of research in the Gartner Customer Service and Support Practice, covering the customer service experience and how service organizations deliver differentiating customer and employee experiences through human-centric technology systems.