The bots are coming!

Currently, there exist two types of CUIs: voice assistants and chatbots. Where the dialogue with a chatbot consists of text messages, a voice assistant wants to hear actual words. In return it gives feedback to the user in audio messages and visual feedback on the screen. The voice assistants Siri and OK Google are well-known and also other companies such as Amazon and Microsoft have developed their own talking assistant. Next to voice assistants, chatbots are becoming popular as well. From within services such as Facebook Messenger and Telegram it is possible to ask a chatbot what the weather is like outside, buy clothes or ask for a cab. And also Slack, the popular messaging app for office workplaces incorporates chatbots to increase productivity. For example, there is a bot that can arrange meetings or a bot that can assign tasks to team members. It is even possible to program your own bot to perform specific tasks. These chatbots provide endless possibilities and more and more companies are starting to use them. Therefore, it is interesting to take a look at the advantages and challenges of CUIs and which role a UX designer can have.

Using conversational UIs can have several advantages over conventional UIs. First of all, the language that is used to communicate with the user is natural language. In other words, a CUI speaks the language of its user. This also empowers a more personal tone of voice. In addition, the chat-style way of communicating can help in communicating the brand’s personality by giving the chatbot a specific character. Moreover, since the CUI consists of a very familiar interface which is used every day during contacting a friend or family, there is almost no learning curve. Another advantage of a CUI is the possibility of integrating a chatbot in popular social platforms, such as Facebook Messenger. This can lower the threshold for a user to make use of a certain service of a company. The user does not have to install an app or visit a website, but can simply start talking to the chatbot from within a messaging service. There are also companies that are developing their own chatbots. This can have advantages since there are no limitations from an existing platform and there is a more direct connection to the company. However, the user still should install an app or visit a certain website, which can result in less users. Therefore, it is recommended to investigate what is the best fit for the context and tasks of the chatbot.

Although using a CUI can have advantages over a GUI, it doesn’t mean that using a CUI is always a good idea. Just as in conversations with humans, it is easy to lose track of the original intention of the conversation and miscommunications can be made. A GUI can provide feedback through graphical elements, whereas a CUI can only give output in verbal or written text. Next to that, there is almost no discoverability in a CUI. How can users find what their looking for without graphical affordances? The user is confronted with an empty text field and can enter an endless amount of possibilities without graphical guides. However, it is possible to provide the user with pre-programmed answers to steer the input of the user in a certain direction. But even if the discoverability is improved, performing a task through a CUI can be slower and less intuitive than when the task is performed through a GUI. This can cause frustration for the user. It is therefore important to test whether a CUI is a real improvement over a conventional GUI for the tasks that should be performed.

It may look that a conversational UI is only a matter of programming, but that is certainly not the case. The tone of voice, type of messages and pre-programmed answers (quick replies) to steer the conversation in the right direction, should all be designed to give users the best possible experience. Although the use of CUIs are relatively new, there are already a lot of guidelines to be found on the internet. In this article, Eunji Seo gives a good overview of best practices for designing chatbots. Additionally, this article by Matty Mariansky (designer of the chatbot Meekan that can schedule meetings) is very helpful for designers to improve the experience with a chatbot.

Also, voice assistants need a lot of refinement before it can give a trouble-free experience. In example, since voice assistants like Google Home and Alexa (by Amazon) can be used by different users in a household, this asks for new ways of validating users. This and other challenges with voice assistants are described in this article about Amazon’s voice assistant named Alexa. Although a lot depends of the maturity of the technology, there are possibilities for UX designers to make interactions with voice assistants less frustrating. The first job applications for Voice UX designers are already there, asking for strong pioneering skills.

As CUIs will become more and more visible in the digital world, UX designers should come up with ways to improve its experience and develop best practices. As the technology behind CUIs is improving, the better a user can use natural language to communicate with devices. It seems unlikely that there will be a complete shift from interacting with GUIs to interacting with CUIs. Some services will just work better with a GUI. However, that CUIs will have a great impact on the way we will communicate with the devices around us in the future is undeniable.