The way you communicate with the user, has always been a fundamental part of a great user experience. Marketers and copy writers have always strived to get the brand personality across, establish a tone of voice and aim to talk to their customers in engaging ways.
Retailers and technology firms are starting to turn up the dial on how they talk and interact with their customers - wanting the conversation to be two way. The increase of messaging apps, advances in machine learning and natural language processing are providing companies new ways to talk with their customers.
Companies see conversational user interfaces (CUI) as a way to reduce some of their business overheads, make users' tasks easier, provide customer service and sales support and create personal and tailored shopping experiences.
The increase of messaging apps, advances in machine learning and natural language processing are providing companies new ways to talk with their customers.
At Webcredible, a lot of us have been involved in creating user interfaces with a conversational style. A few of us got our heads together and collated some tips / principles that we've learnt through our experiences.
When you meet someone new it's common etiquette to introduce yourself. The same goes for a conversation whether it's in a wizard, messaging app or Siri. Keep it short and to the point, use the introduction to define some expectations of the conversation, the next steps and the eventual outcome.
Ideally, the conversation should make the task quicker and easier for the user than interacting with a more traditional user interface or doing the task offline. If you need the user to interact with a screen - keep the conversation flowing. Get them to select an answer with touch or with their voice. Reduce the need to type. Try and anticipate your users' needs.
Moodnotes lets users to track their mood over time. It allows for very quick inputs if users don't have time to add more detail.
Give the user the opportunity to continue the conversation at a place or on a device that's convenient for them. Aim to design a cross channel experience, that can take place anywhere. Allow the users to save the conversation and resume it at a later time or day.
If appropriate to the brand or service, consider adding an element of fun to the interaction and during the conversation. Gamification can increase the user's engagement with your brand and make mundane tasks like form filling, more delightful and memorable.
Brompton helps users find a bike, collecting information in a conversational and fun way.
Conversational interfaces are gaining traction in customer service and retail - because it's a great chance to use existing staff knowledge. Try to reflect the expertise of your in-store or telephone staff and compliment their knowledge with your conversation.
The tone of the language should feel conversational and genuine. Reflect the style of the person you are trying to emulate. If it's an online health triage service, the tone and style should be the same as a nurse would adopt when talking to a patient. Use plain un-jargonised language, it's easier for users to understand and take action.
Make sure you read the conversation dialogue out loud. Does it make sense from beginning to end, with a comprehensive order, rhythm and pace? Test it with users and tweak as necessary, adding or reducing questions if needed.
You may also need to emulate the writing style of your user. This is especially applicable to using a messaging app to talk to your users, especially if your target audience are millennials or tweenies. Your chat bot will need to be fluent in text speak and emojis.
Quartz' news app adapts its language
To make your conversation appear more natural, adding a small amount of friction or delay makes the interaction feel more real. Introduce a pause... maybe with additional words like, "um... just thinking" to help the conversation feel more genuine.
Give hints to the types of answers you are looking for. If you have a small amount of choices, present all the possible answers. Avoid open-ended questions and keep the conversation focused. When the answer is valid, repeat it to make sure it has been understood and then move on the next question. Reveal suggestions to the user that are relevant and useful to them. Base suggestions and replies on the user's previous actions. A great example of this is Lunchy, a Slack bot from TruthLabs.
The conversation will eventually flow down to a specific action, like go to store location or present a small selection of choices, e.g., three options of trainers (on an eCommerce website). If presenting the user with results based on their answers, explain the rationale. Why did these results appear for them? And, as with any conversation, give users the chance to steer it, if they want to. Give them the flexibility to amend their choices and to save or dismiss these results.
Like with all user experiences, the conversational user interface should be tailored for the user. It is important to understand the user through research and then be prepared to dive deeper to understand their language.
Many thanks to Maya, Elvis and Martha for their contribution to this article.