What is Dialog Studio?
Dialog Studio is a workbench for building custom conversation flows. It abstracts conversational concepts into three basic elements: entry (trigger), function, and response. By combining these elements, you can build almost any conversation flow. This approach supports classic slot-filling mechanisms, such as parallel and dependent slots. It also allows for conditional branching based on parameter values collected during slot filling for tasks such as duplicate name checks or data validation.
The origin of Dialog Studio
Artificial intelligence (AI), powered by deep learning, is now used in applications worldwide. This widespread adoption has led to the creation of many types of chatbots. Some are for daily life, such as smart speakers. Others are for work, such as efficient customer service bots. In the chatbot realm, user queries are typically classified into three main types:
Informational: "What is our company's personal income tax rate?"
Chit-chat: "Hello," "Thanks," or "You're great."
Task-oriented: "Help me schedule a meeting with John," "I need to request time off," or "Help me request a return."
Example
Among these query types, task-oriented scenarios usually require two or more rounds of conversation to resolve the user's issue. In the following example, U represents the user and B represents the bot:
U: "Help me book a train ticket to Beijing for tomorrow."
B: "Okay, which city are you departing from?"
U: "Hangzhou."
B: "OK. I have found the following train tickets for you. Please make a selection."
…
Analysis
In this ticket booking scenario, the user's first sentence clearly states the intent to book a ticket. It also provides two key details: the departure date ("tomorrow") and the destination ("Beijing"). However, the bot's logic requires one more piece of information: the departure city. The bot then asks a follow-up question to obtain the missing detail. The text for these follow-up questions is user-configurable. After collecting all the necessary information, the bot performs the task. This could involve searching for matching train tickets or even buying a ticket for a window seat based on the user's known preferences.
Refinement
As more conversation scenarios are developed and become popular, the demand for better tools grows. To accelerate chatbot development, we need tools and platforms that adhere to industry standards, increase efficiency, and help developers structure their logic. These tools should meet the following requirements:
Abstract the core elements of complex conversations to serve as building blocks for all conversation flows.
Define clear rules for how these elements connect. This helps developers and conversation designers structure their logic and master conversation design.
Be simple, efficient, and easy to use to reduce the learning curve for new users. Provide as many common, pre-configured components as possible for out-of-the-box use.
Welcome to the new world of chatbots.