Labs Tool
Enterprise AI Glossary
Plain-English definitions of enterprise AI terminology.
A
Agent
An AI system that can perceive its environment, make decisions, and take actions to achieve goals. Enterprise agents can call tools, retrieve information, and execute multi-step workflows.
Agentic Workflow
A workflow where AI agents autonomously plan, execute, and iterate on multi-step tasks, potentially calling external tools and APIs.
C
Chunking
The process of splitting large documents into smaller segments before embedding them into a vector database. Chunk size affects retrieval quality and context window usage.
Context Window
The maximum amount of text an LLM can process in a single request. Larger context windows allow more retrieved documents to be included in the prompt.
E
Embedding
A numerical vector representation of text that captures semantic meaning. Similar texts have embeddings that are close together in vector space.
F
Fine-tuning
The process of further training a pre-trained model on domain-specific data to improve performance on specific tasks.
G
Grounding
The practice of connecting AI responses to verified, factual sources — typically your enterprise knowledge base — to reduce hallucinations.
H
Hallucination
When an LLM generates plausible-sounding but factually incorrect information. RAG and grounding techniques are used to reduce hallucinations in enterprise deployments.
Hybrid Search
A retrieval approach that combines dense vector search (semantic) with sparse keyword search (BM25) to improve recall and precision.
K
Knowledge Graph
A structured representation of entities and their relationships, used to enhance AI retrieval with contextual connections between concepts.
L
LangChain
An open-source framework for building LLM-powered applications, providing abstractions for chains, agents, memory, and tool use.
LangGraph
A library for building stateful, multi-actor AI applications using graph-based workflows. Used for complex agent orchestration.
LLM
Large Language Model. A neural network trained on vast amounts of text data, capable of generating, summarising, translating, and reasoning about text.
M
Multi-model Routing
The practice of directing different types of queries to different AI models based on cost, capability, latency, or compliance requirements.
P
pgvector
A PostgreSQL extension that adds vector similarity search capabilities, enabling hybrid relational + vector storage in a single database.
Prompt Engineering
The practice of designing and optimising input prompts to elicit better, more accurate, or more useful responses from LLMs.
R
RAG
Retrieval-Augmented Generation. A technique that enhances LLM responses by first retrieving relevant documents from a knowledge base and including them in the prompt context.
RBAC
Role-Based Access Control. A security model where access to resources is determined by a user's role within an organisation.
Re-ranking
A post-retrieval step that re-orders search results using a more powerful model to improve relevance before passing them to the LLM.
S
Semantic Search
Search that understands the meaning and intent behind a query, rather than just matching keywords. Powered by vector embeddings.
SSO
Single Sign-On. An authentication scheme that allows users to log in once and access multiple systems. Essential for enterprise AI deployments.
System Prompt
Instructions given to an LLM at the start of a conversation that define its persona, constraints, and behaviour. Critical for enterprise AI assistants.
T
Token
The basic unit of text processed by an LLM. Roughly 4 characters or 0.75 words. LLM pricing and context limits are measured in tokens.
V
Vector Database
A database optimised for storing and querying high-dimensional vector embeddings. Examples include Pinecone, Weaviate, Qdrant, and pgvector.
Z
Zero-shot
The ability of an LLM to perform a task without any task-specific examples in the prompt. Contrast with few-shot prompting.