Terminology for Conversational AI you need to know in 2023
Mastering the Vocabulary of VUI Design
Chat GPT is getting a lot of well due attention right now and CES 2023 showcased the next generation of ambient experiences empowering customers with choice. So whether you’re a designer, developer, product manager, or simply someone who is curious about how products and services are designed, here’s important jargon that will help you better understand and speak the language of Conversational UX when designing for chatbots and voice assistants (also known as VUI Design).
1—VUI flow:
Stands for “Voice User Interface flow”, and refers to the flow of the conversation between the user and the system, including how the user interacts with the system and what responses the system provides
2 — Wake word:
The word or phrase used to activate the VUI. For example, “Alexa”, “Hey Google”, “Computer”
3 —Prompts:
A prompt is a pre-written statement or question that a chatbot or voice assistant uses to communicate with a user. Prompts are typically used to guide the user through an interaction with the technology, and they are used to provide information, ask questions, or give instructions to the user.
4 — Intent:
“Intents” and “entities” are terms used to describe the specific actions or pieces of information that a chatbot or voice assistant is able to understand and respond to.
5— Utterance (voice command or voice input):
Typically made in the form of a question, command, or request, utterances are used to initiate an interaction with the technology. Utterances are typically processed by the chatbot or voice assistant using natural language processing (NLP) and natural language understanding (NLU) techniques to understand the user’s intent and extract relevant information.
6 — Text command (or text input):
Refers to the way a user interacts with a chatbot or other text-based interface, such as a website or mobile app. It can take the form of a question, a request, or a command.
7— Slots:
A “slot” is like traditional a form fields in the sense that they can be optional or required, depending on what’s needed to complete the request. Example: when ordering a shirt size and color would be two required slots.
8— Sample Dialogue
A sample dialogue is a simulated conversation between a user and the chatbot or voice assistant. It is used to demonstrate how the technology is intended to work and what kind of interactions and responses the user can expect. A great way to ensure the experience feels natural and flows well.
9—SLU:
Short for “Spoken Language Understanding”, it’s the ability of the system to understand the user’s speech.
10—STT:
Stands for “Speech-to-Text” which is a technology that converts spoken language into written text. This technology is also known as “automatic speech recognition” (ASR). It enables a user to speak into a device, such as a smartphone or computer, and have their spoken words transcribed into written text in real-time.
11—TTS:
Stands for “Text-to-Speech.” It refers to the technology that converts written text into spoken words.
12—Voice user testing:
Part of the design process where you test the voice user interface with real users.
13—Error handling:
In the context of design, error handling refers to the consideration of potential errors and strategies to minimize or mitigate the effects of those errors. This can include guiding users through the process of correcting errors, or incorporating built-in error-checking mechanisms to detect and correct errors automatically.
14— Natural language understanding (NLU):
The ability of the system to understand and interpret human language. NLU is concerned with analyzing, understanding, and generating the meaning of text data in a way that is similar to how humans do it.
15— Natural Language Processing (NLP) :
The field of study which deals with the interaction between human language and computers. NLP is concerned with developing algorithms, methods, and models that enable computers to process, analyze, understand, and generate human language.
16— NLG (natural language generation):
The ability of the system to generate human-like language as a response. NLG systems use natural language processing techniques to convert structured data, such as database entries or spreadsheets, into written or spoken language that is easy for humans to understand.
What are other examples of jargon you’d add to the list? Leave a comment!