チャットボット＆AI 用語集

A chatbot glossary is a comprehensive reference of terms used in chatbot development, conversational AI, and customer experience automation. It helps developers, marketers, and business owners speak the same language when building and optimizing chatbot solutions. Understanding these terms ensures you make informed decisions about your chatbot strategy.

A chatbot is a software program that simulates human conversation, often using rule-based logic or predefined scripts. Conversational AI is broader — it uses natural language processing (NLP), machine learning, and large language models to understand context, intent, and nuance. All conversational AI systems include chatbots, but not all chatbots use conversational AI.

NLP stands for Natural Language Processing, a branch of artificial intelligence that enables machines to read, understand, and generate human language. In chatbot development, NLP powers intent recognition, entity extraction, sentiment analysis, and natural-sounding responses. Modern chatbots use transformer-based NLP models for significantly better understanding.

Intent recognition is the AI capability that identifies what a user wants to accomplish from their message. For example, when a user types 'I need to reset my password,' the chatbot recognizes the intent as 'password_reset.' Accurate intent recognition is critical because it determines which response or workflow the chatbot triggers, directly impacting user satisfaction.

Large language models (LLMs) like GPT-4, Claude, and Gemini are deep learning models trained on massive text datasets. Chatbots use LLMs to generate human-quality responses, understand complex queries, summarize information, and handle conversations that rule-based systems cannot. LLMs enable chatbots to be more flexible and natural in their interactions.

Rule-based chatbots follow predefined decision trees and keyword matching — they can only handle scenarios the developer anticipated. AI-powered chatbots use machine learning and NLP to understand user intent, learn from conversations, and handle unexpected queries. AI chatbots are more flexible but require more training data and computational resources.

Consider these factors: your use case (customer support, lead generation, internal helpdesk), required integrations (CRM, helpdesk, e-commerce), AI capabilities (NLP, LLM support, multilingual), deployment channels (website, WhatsApp, Slack), and budget. Platforms like Conferbot offer no-code builders with AI capabilities, making it easy to start without technical expertise.

Human handoff (also called agent escalation) is the process of transferring a chatbot conversation to a live human agent when the bot cannot resolve the query. Effective handoff preserves the full conversation context so the agent can continue without making the customer repeat themselves. Most enterprise chatbots implement handoff for complex or sensitive issues.