Conversational UX Characteristics
Ease of application learning for users
With conversational UX, users no longer need to memorize the rules of complex UIs. For example, in applications with many types of operations, such as creative tools, users can benefit from operating with natural language.
Input via natural language
Applications that adopt conversational UX must assume messages (prompts) in any natural language as user input.
It is difficult to perform validation at the time of user input because there are countless patterns of natural language messages. Unlike traditional application inputs, these inputs are not limited to words or keywords, and are not always selectable.
In addition, it is not known whether the application can process the input until the language model interprets the prompt.
Therefore, it can be assumed that the pattern of user input in conversational UX is almost infinite.
Difficulty of application learning for users
Users have fewer opportunities to learn how to operate conversational UX applications than with traditional ones. This is because the basic structure of the UI is limited to the input field for prompts, and the use of buttons and menus is limited.
Also, since the history of conversational UX is not yet long enough, users are not accustomed to this experience. This means that if learning is not done carefully, users may get lost.
Bing AI presents a candidate prompt immediately after the previous assistant's message to assist users in learning how to use the application.
We will explain these learning assistance methods later, but for now, we will only mention that users have limited opportunities to learn about the application and are not yet fully accustomed to it.
User expectations for AI
Users may not yet fully understand the characteristics of language models. Applications that are centered on language models may make users believe that they are capable of understanding human language and intentions perfectly, and that they are able to grasp real-time situations in the real world.
Challenges of conversational UX
Users have too much freedom
It has been some time since the introduction of conversational UX, but there still seems to be a wide range of differences in the amount and method of use among users. While some people use it well, there are still many who do not know how to use it. This is an unavoidable problem due to the high degree of freedom in input in conversational UX.
As mentioned earlier, input in conversational UX is in natural language, so it is possible for users to get lost without knowing what they can and cannot do.
Possibility of disappointing user expectations
Given the characteristics of conversational UX, users always run the risk of dropping out because they say "Why can't it do this?" For example, even with ChatGPT, it is common to see cases where users ask about the current weather or information about acquaintances, but receive unexpected responses.
Last updated
