Terminology

This page acts as a terminology guide, covering general AI and service operation terms as well as common Ebbot-specific expressions.

AI agent – An AI agent operates independently of the communication channel and can assist users via chat, email, voice, and more. It uses generative AI to understand context and execute complex tasks.

Chat agent – A specific type of AI agent configured to operate within the Chat channel.

Email agent – A specific type of AI agent configured to manage communications via Email or a ticketing system.

Chatbot – Traditionally, a chatbot provides intent-based answers to specific, predefined questions. The Ebbot platform supports both classic intent-based scenarios and modern AI functionality. In daily conversation, the terms "chatbot" and "AI agent" are often used interchangeably, as many solutions mix predefined flows (Scenarios) with generative AI features.

Persona – The system instructions (prompt) that define your AI agent's personality, tone of voice, and specific behavioral rules. This is added within EbbotGPT Configurations, which is available for when configuring Chat agents.

Prompt – Instructions given to an AI model. In Ebbot, you create a Persona within the EbbotGPT Configuration, which acts as the agent's core instruction. Within Automations, prompts are used in LLM nodes to direct the agent on exactly what to perform in a specific workflow.

Config – Short for EbbotGPT Configuration. This is where you define your chat agent’s persona, set up safeguards, enable tool calling, connect to the AI agent's knowledge, and adjust technical parameters like temperature.

Scrape – The process where Ebbot automatically crawls and extracts content from a website or other source to be used as knowledge for your AI agent.

Scenario – A predefined, intent-based conversation flow in the Chat platform that guides an end user through specific steps or questions without necessarily using generative AI.

Extension – A connection in the Chat platform to external systems or custom code that expands the platform's functionality, such as fetching real-time data during a conversation.

Workflow server – An integration built for workflows in the Automations platform, for example, to a ticketing system to create an Email agent workflow.

EbbotGPT – The umbrella term for Ebbot's generative AI technology, covering model configuration, tool calling, management of knowledge, and more.

Embedder – The technical component that converts text into vectors (numerical representations), allowing the AI agent to search and retrieve relevant information from your knowledge base.

Safe guards – Also known as guardrails, these are security settings and filters that restrict what the AI agent can discuss or how it should behave, ensuring safe and accurate interactions.

Tools – Functions or external integrations that the AI agent can "call" (tool calling) to perform specific actions, such as searching the web or checking an order status.

Knowledge – The collective information available to your AI agent. It serves as the agent's brain and consists of data gathered from documents, websites, and integrations.

Dataset – An organizational container within Knowledge used to group related information. A dataset can contain multiple different sources.

Document – The smallest unit of data. When a source is processed, it is broken down into searchable documents that the AI agent can reference in its answers.

Data sources / sources – The individual knowledge sources contained within a Dataset.