Terminology

This table shows the terms you must know before you begin using the GenAI Assistant platform:

Term	Description
Large Language Model (LLM)	A machine learning algorithm that is trained to simulate natural language processing tasks. These tasks may range from text generation to customer analysis.
Token	A unit of text that the LLM processes for every user message and response generation.
Prompt	A user response to the LLM that may include task-specific instructions.
Prompt Engineering	A new science that explores various techniques for extracting value from an LLM.
Prompt Catalog	Platform to compare LLM responses from up to three models.
Tool	Functional modules that the GenAI Chat Service can leverage to address the user’s request.
Tool Description	A description of a Tool that details what the Tool’s API must be used for.
Agent Instructions	Detailed instructions for why a particular Tool must be selected, including any no- table dependencies.
Chat	Testing platform for custom Tool executions.
Agent	System which independently accomplishes specific user tasks, streamlining the workflow
Retrieval Augmented Generation (RAG)	The ability for the LLM to reference a dedicated knowledge base outside of the LLM’s training data before generating a response.
API endpoint	API endpoints deployed on an API Gateway are used by skills to communicate with the applications to perform a transaction or get specific information.