Enhancing the Conversational Assistant Experience: Three Pitfalls to Avoid

Conversational assistants, such as text-based chatbots and voice assistants, have been commonplace for a number of years now. And with the advent of generative AI, even more users are engaging with conversational interfaces as part of their digital interactions with retail brands, healthcare providers, banks and more.

Yet, despite their ubiquity, conversational assistants have often struggled to meet user expectations and have not (yet) won the hearts of the public. Much like their predecessors, touch-tone IVR systems, users often find them cumbersome and frustrating, reporting issues with comprehension and task fulfillment.

These flaws haven’t, however, deterred businesses from continuing to implement conversational assistants. The allure of automation and its associated cost savings, especially in customer service, is simply too attractive. And as much as users may resent conversational assistants, they still expect companies to offer them as part of their customer service options.

Though it may seem counterintuitive, this represents an opportunity for companies. In a world where users are constantly disappointed by poor-performing bots, it stands out when a conversational assistant exceeds their expectations — so much so that it often fosters strong brand affinity. Research shows that 90% of respondents are more likely to make another purchase after a good interaction.

The primary objective of customer service conversational assistants is to help users complete their tasks easily and efficiently. On a basic level, the system needs to understand user intent. Repair flows such as “I’m sorry, could you please repeat that” or “I’m sorry, I didn’t get that” are incredibly frustrating for users, especially if the system simply forges ahead with the wrong intent. Avoiding situations like these is the bread and butter of conversational assistants.

Beyond this, there are three critical areas companies should focus on to enhance user satisfaction: clarity of purpose, access to human agents and frictionless interaction.

1. Defining the assistant’s purpose

Users need guidance on the sort of requests they can make to a conversational assistant for the interaction to be successful. Especially if this is the first time a user has encountered a conversational assistant, they may have no idea what to ask for and opt to speak to a human instead. That’s why companies must incorporate an opening prompt that gives the user concrete examples of the sort of task or inquiry the assistant can support them with. 

Example:

• Effective: “Hi, I’m your AI assistant. I can help with payments, order status, or disputes.”
• Less Effective: “Hi, I’m your AI assistant. How can I help you?”

2. Providing access to human agents

Users often want the option to speak to a customer service agent. Many people still have a lingering aversion to automated self-service processes ever since the days of “phone tree” IVR systems that made it near impossible to get through to a human. Thoughtful conversation design should address this concern while also promoting the advantages of AI assistance.

Example:

• “An AI agent can help you now, or you can provide your information for a human agent to contact you later.”

The underlying message should convey the brand’s genuine care for the customer. Increasing containment rates doesn’t necessitate concealing the option to speak with a human.

3. Delivering a frictionless experience

The majority of today’s conversational assistants are not as integrated with backend systems as they could be. This inhibits personalized experiences because assistants end up asking users for information they should already have access to, such as order numbers. The best conversational assistants reduce the number of conversational turns needed to resolve an inquiry by instantly retrieving relevant user information.

Example:

• Frictionless: “Your order number 123 is expected to arrive in 2 days. Any other orders you’d like to get an update on?”
• Less Frictionless: “What’s the number of the order you would like an update for?”

Another additional friction point, particularly in voice assistants, happens when the system takes too long to respond to a user. Dead silence, whether it’s because the assistant is looking something up or simply didn’t hear the user, can lead users to abandon the call. This can easily be rectified by using audio, such as the sound of someone typing, to indicate the system is busy. A robotic or unnatural voice, as well as a conversational style that doesn’t align with the brand’s identity, add more unwanted friction.

The Path Forward

AI automation through chatbots and voice assistants is rapidly becoming an indispensable component of every business’s customer experience strategy. It’s clear that they are here to stay. By prioritizing the user experience, businesses can not only cultivate greater customer satisfaction but also reap the benefits of improved brand loyalty and increased revenue.