AI Glossary: Key Terms Explained
This AI glossary explains the terms that appear in vendor pitches and board discussions, from LLMs and tokens to RAG, agents and fine-tuning. Each definition is short and says why the term matters when you are buying or building.
About this ai glossary
Terms are arranged alphabetically with short, plain definitions. Where a term is often misused, we say so. Many vendors call a simple chatbot an agent, for example, and fine-tuning is often proposed where retrieval would be cheaper and more accurate. Entries cover models and how they work (LLM, token, context window, temperature), building with AI (prompt, RAG, embeddings, vector database, tool use, agent) and risks (hallucination, prompt injection, data leakage, evaluation).
Use it alongside proposals and demos. If a vendor uses a term you cannot pin down, check it here and ask them to explain how it applies to your use case. Several entries link to our AI services and the guide on running a first AI project.
AI agent #
Software that uses an AI model to plan and carry out multi-step tasks, often using tools and APIs.
Related: AI Agent Development Services
Fine-tuning #
Further training a pre-trained model on specific data to improve performance on a task.
Generative AI #
AI that creates new text, images, audio or code based on patterns learned from data.
Related: Generative AI Solutions
Guardrails #
Rules and checks that keep an AI system’s outputs safe, accurate and on-policy.
Hallucination #
When an AI model produces information that sounds plausible but is false.
LLM (large language model) #
A model trained on large volumes of text to understand and generate language, such as Claude or GPT.
Multi-agent system #
Several AI agents with different roles that collaborate to complete a larger task.
Prompt #
The instructions and context given to an AI model.
RAG (retrieval-augmented generation) #
Giving an AI model relevant documents at query time so answers are grounded in your data.
Related: AI Knowledge Base Services
Token #
The unit of text a language model reads and writes, roughly three-quarters of a word.
Case studies & resources
Real-world results and practical reading to help you plan your next step.
Client stories
How similar work has played out for other businesses.
Helpful resources
Guides, tools and articles our team recommends on this topic.
Ready to put this into practice?
A free 30-minute session with a senior Nians strategist. You leave with a prioritised plan, whether or not you work with us.