AI Chatbot & Customer Support Glossary
Key terms and definitions used in AI-powered customer support, conversational AI, and chatbot technology.
- AI Chatbot
- An automated conversational agent powered by artificial intelligence that can understand and respond to user questions in natural language. AI chatbots use large language models (LLMs) and retrieval-augmented generation (RAG) to provide accurate, contextual answers from a business knowledge base.
- RAG (Retrieval-Augmented Generation)
- A technique that combines information retrieval with AI text generation. The system first searches a knowledge base for relevant documents, then uses those documents as context for the AI to generate accurate, grounded responses. This reduces hallucinations and keeps answers factual.
- CSAT (Customer Satisfaction Score)
- A metric measuring customer satisfaction with a support interaction, typically collected as a 1-5 star rating after a conversation ends. CSAT helps businesses track support quality and identify areas for improvement.
- Handover
- The process of seamlessly transferring a conversation from an AI chatbot to a human support agent. Smart handover systems evaluate confidence scores and visitor sentiment to determine when human assistance is needed, passing full conversation context to the agent.
- Knowledge Base
- A structured collection of information that powers AI chatbot responses. Sources include crawled website pages, uploaded PDF documents, custom Q&A pairs, and product catalogs. The knowledge base is indexed using vector embeddings for semantic search.
- Embeddings
- Mathematical vector representations of text that capture semantic meaning. Embeddings allow AI systems to find conceptually similar content even when the exact words differ. They power semantic search in knowledge bases and enable the AI to match visitor questions to relevant answers.
- Semantic Search
- A search technique that understands the meaning and context of a query rather than just matching keywords. Semantic search uses vector embeddings to find conceptually relevant results, enabling an AI chatbot to understand questions like "What are your hours?" and "When are you open?" as equivalent.
- Chat Widget
- An embeddable user interface component that adds live chat or AI chat functionality to a website. Chat widgets typically appear as a floating button that expands into a chat window, allowing visitors to communicate without leaving the page.
- Tenant
- An isolated workspace in a multi-tenant SaaS application. In WebDialogAI, each tenant represents one business or website with its own knowledge base, widget configuration, agent team, and conversation history. Tenant isolation ensures data privacy between businesses.
- Pre-Chat Form
- A form displayed to website visitors before starting a chat conversation. Pre-chat forms collect contact information such as name, email, and phone number, enabling businesses to follow up and qualify leads from chat interactions.
- Topic Guard
- A feature that restricts AI chatbot responses to relevant business topics. Topic guard uses similarity matching or LLM-based classification to detect and block off-topic questions, ensuring the chatbot stays focused on customer support rather than engaging in unrelated conversations.
- SSE (Server-Sent Events)
- A web technology that enables a server to push real-time updates to a client over a single HTTP connection. In AI chatbots, SSE is used to stream AI-generated responses token by token, creating the appearance of the AI "typing" its response in real time.
- Queue Management
- A system for organizing and prioritizing visitor conversations when all human agents are busy. Queue management provides live position tracking, estimated wait times, and automatic failover if an agent disconnects, ensuring no visitor is left waiting indefinitely.
- Canned Response
- A pre-written reply template that support agents can quickly insert using keyboard shortcuts. Canned responses save time on frequently asked questions and ensure consistent messaging across the support team.
- Vector Search
- A search method that finds similar items by comparing their vector embeddings in a high-dimensional space. Vector search enables semantic similarity matching, allowing an AI chatbot to find relevant knowledge base content even when the visitor uses different phrasing than the source material.
- pgvector
- An open-source extension for PostgreSQL that adds support for storing and querying vector embeddings. pgvector enables efficient similarity search directly in the database, making it a popular choice for RAG-based AI applications that need fast, scalable vector search.