By: Tiziana Nannavecchia, Aryeh Bookbinder, and Maggie Pitt
Writing responses for a chatbot shouldn’t feel like a guessing game. There’s an entire discipline focused on this task: it’s called conversation design, and we have a team of conversational designers at Glia! They’ve developed best practices and principles for response-writing to make your deployment easier and your members interactions all-the-more frictionless.
Though chatbots are traditionally designed specifically for the channels they live in (phone, web, mobile, etc), we at Glia champion an approach of Unified Interaction Management (UIM). With UIM, our Glia Virtual Assistants (GVAs) are designed to be channel-adaptive and our best practices are relevant across them all.
Want a sneak peek into the world of conversational design? Our own expert designers have shared 10 of their best practices for ensuring your chatbot’s pre-written responses are effective, informative, and leave users with a positive experience.
- Develop a distinct tone of voice
Your chatbot is your brand ambassador: having a defined voice that aligns with or complements your brand is crucial, as users often infer personality from virtual assistants.
- Make it clear what users are talking to
Chatbots should talk just like a human does for a natural and engaging interaction, but be sure to clarify to users that they are conversing with a virtual assistant. Manage expectations transparently and introduce your virtual assistant as such in the welcome message.
- Be strategic in how you begin and end the interaction
Your welcome message is your chance to create a great first impression, and set expectations on what your chatbot can handle and how to engage with it. Wrap up your responses by asking if the answer was helpful, or ask if there’s any further questions.
- Encourage engagement with your chatbot
Though users should be aware of the option to speak with an agent, you should avoid presenting this information up front. Embedding agent offers in nested responses allows you to introduce the opportunity at the right moment.
- Acknowledge users’ queries
In digital interactions, opt for a more relaxed approach. Seamlessly weave landmark statements into responses by using phrases like “When it comes to setting up a direct deposit…”. In phone interactions, choose a more direct approach by using phrases such as, “Sure, I can help you find your transaction history.”
- Include alternative phrasing
For responses likely to be repeated in a session (i.e. main menu, fallback, wrap up messages), try to avoid having your chatbot continuously repeat itself. This makes the interaction feel less robotic and more engaging.
- State your information in order of importance
If a piece of information might be helpful to a smaller number of users, you should share it after sharing the information that is more relevant to the majority of users. This way, the bulk of your users will receive the information they need sooner.
- Lean on the side of brevity
Research has shown that the average human attention span is approximately nine seconds, so present only useful information in a digestible manner. In your chatbot, for example, separating ideas into as many as three chat bubbles makes for easier scanning compared to having one long response.
- Avoid confusion at all costs
In your phone chatbot, maintain clarity with straightforward language while avoiding complexity, jargon, and superfluous information. Limit menus to a maximum of five unambiguous choices to prevent frustration and ensure easy navigation, especially in noisy environments where audio might be compromised and callers easily distracted.
- Provide value
Offer easy-to-follow instructions and practical information without unnecessary ‘business-speak’ or redundancy. Regularly update your responses and be open to improvements based on feedback: Always ensure your content meets your users’ needs and expectations, today and over time.
Implementing these best practices will help you more easily and effectively write responses for your business’s chatbot—and save you time in deploying across channels.
