What is Conversational Context (Context)?

Conversation context refers to the flow and context of a conversation. It encompasses various elements such as the situation in which the conversation takes place, the relationship between the speakers, and the purpose of the conversation. Understanding the conversation context is crucial for effective communication. This is because it allows you to better understand what the other person is saying and respond appropriately.

There are several ways to improve your understanding of conversation context.

Understanding the Other Person's Background

Understanding the other person's age, gender, occupation, and interests can help you better understand the context of the conversation.

Understanding the Situation

It is also important to carefully observe the situation in which the conversation takes place, such as the location, time, and atmosphere.

Paying Attention to Nonverbal Cues

Nonverbal elements such as the other person's facial expressions, tone of voice, and posture can also help you understand the conversation context.

Utilizing Feedback

Checking the other person's reaction to see if your message has been delivered correctly can also help with understanding the context.
A good understanding of conversation context leads to smoother communication and reduces the likelihood of misunderstandings and conflicts. Therefore, understanding conversation context is essential for effective communication.

So, what does conversation context mean in the context of chatbots?

Example 1

• User: What's the weather like today?
• Bot: The weather today is clear.
• User: How about tomorrow?
• Bot: Tomorrow's weather is clear.

Example 2

• User: How about tomorrow?
• Bot: What about tomorrow? (Note: Siri, for example, provides the user's schedule for the next day)
• User: What's the weather like?
• Bot: Tomorrow's weather is clear.

This ability to maintain continuity between previous and subsequent conversations is referred to as conversation context.

Another example would be going to a car repair shop and saying "It's broken." The bot wouldn't be able to determine what is broken. In this case, the bot would have to ask "What's broken?" or simply perform a general checkup.

Ultimately, when the initial statement is ambiguous, it's necessary to ask follow-up questions to guide the conversation towards the desired answer.

How can a chatbot understand conversation context?

Conversation History Analysis: Analyzes previous conversation content to infer the context.

Natural Language Understanding: Uses natural language processing techniques to accurately understand the meaning of sentences.

Utilizing Multimodal Input: Integrates and analyzes various inputs such as text, images, and voice. (This is why we mentioned paying attention to nonverbal cuesin the context of human conversation, as it's not possible to rely solely on text.)

Understanding conversation context is crucial for improving a chatbot's natural conversation abilities. This provides users with a better experience. As chatbot technology continues to advance, we expect the ability to understand conversation context to improve as well.

❗When creating a chatbot using a chatbot builder?

During the chatbot creation process, it's important to carefully consider how to handle users' comprehensive questions using context effectively. Chatbot builders generally offer context functionality. However, for comprehensive questions or those requiring follow-up questions, you need to manually create and connect the interactions within the builder. This ensures that the bot can provide the most relevant responses to the user's query.

❗When creating a chatbot using Generative AI?

The chatbot's capabilities are partly determined by the amount of conversation history it can store. Therefore, in certain scenarios, a separate database might be constructed to store specific information. For instance, information like hobbies (e.g., skiing), travel preferences, and so on can be stored in a separate database while the conversation context itself is discarded. This is because storing a vast amount of conversation data can be challenging. Then, this stored database information can be retrieved and utilized as needed during the question-answering process or when the flow of conversation requires it.

Example 3

• User: I like winter, you know.
• Bot: Right! You mentioned that skiing is your hobby. Have you been to a ski resort this year?