Guide
AI Glossary: Plain English Definitions
The AI industry produces jargon faster than anyone can absorb it. A lot of it is used loosely by people selling things. This glossary defines 52 of the terms you are most likely to meet as a UK business, in plain English. It is written by people who build AI systems rather than market them. Where a term maps to something we do, we have linked to it.
A
Agentic AI
AI systems built to pursue objectives across multiple steps. They do not answer once and stop. The defining trait is acting, as well as responding.
AI Agent
Software given a goal rather than a question. It plans the steps needed and carries them out using tools and systems. It checks its own results and adjusts. It can often complete a task without a person approving each step.
See AI Agents.
AI Governance
The policies, checks and accountability that keep a business's AI use safe, lawful and under control. It covers what the AI may do, what data it may touch, and who answers for it when something goes wrong.
See AI Governance.
AI Integration
Connecting AI to the systems a business already runs. It can then use real data and take action, rather than sitting in a chat window on its own.
See AI Integrations.
AI Literacy
A working understanding, across a workforce, of what AI can and cannot do. It also covers how to use AI sensibly and safely. It is usually the most valuable first training investment for a business.
See AI Governance.
Algorithmic Bias
When an AI system produces unfair outcomes for particular groups, time after time. This is usually because patterns in its training data reflected existing unfairness. It is a central concern in responsible AI.
API
Application Programming Interface. It is the standard door one piece of software offers so other software can talk to it. Most AI integrations work by connecting through an API.
See AI Integrations.
Artificial Intelligence (AI)
Software that performs tasks that normally need human intelligence, such as understanding language, spotting patterns or making judgements. In business use today it usually means machine learning systems, especially large language models.
Automation
Having software carry out a task instead of a person. Traditional automation follows fixed rules. AI automation can also handle steps that need some judgement.
See AI Automation.
B
Business Process Automation (BPA)
Automating a whole business process from start to finish, rather than a single task. It usually spans several systems and several steps.
C
Chatbot
Software that holds a conversation with a person, answering questions and responding to requests. It differs from an AI agent because it talks rather than completes tasks.
See AI Chatbots.
Churn Prediction
Using historical data to spot which customers are likely to stop buying or cancel. The aim is to step in early enough to make a difference.
See Predictive Analytics.
Classification
A machine learning task where the system sorts inputs into set categories. Examples are routing an enquiry to the right department or flagging a suspicious transaction.
Computer Vision
AI that reads images and video. It can read documents from photographs, spot defects on a production line, or recognise objects in a scene.
Context Window
The amount of text an AI model can consider at once. A larger context window lets the model work with longer documents in one go without losing earlier detail.
D
Decision Intelligence
Turning data and forecasts into specific recommendations at the moment a decision is made. A person is not left to read a dashboard and work it out.
Deep Learning
A branch of machine learning that uses layered neural networks. It underpins most modern AI, including language models and computer vision.
Demand Forecasting
Predicting how much of something will be needed, and when. Stock, staff and capacity can then be planned around what is coming rather than what happened last year.
E
Embedding
A numerical representation of text or other content that captures its meaning. It lets software find related information rather than match keywords. It is the mechanism behind most AI search over company documents.
EU AI Act
The European Union's comprehensive law regulating artificial intelligence. It sorts AI uses into four risk tiers, with obligations that grow with the risk. It can apply to UK businesses whose AI reaches people in the EU.
Read the UK guide.
Explainability
How clearly the reasoning behind an AI system's output can be understood and explained. It matters wherever decisions affect people and must be justified to a regulator or a customer.
F
Fine tuning
Further training an existing AI model on your own data so it does better on your specific tasks. It is often unnecessary. Grounding a model in your documents usually achieves more for less.
G
General Purpose AI (GPAI)
AI models that can do a wide range of tasks rather than one narrow job. It is a specific regulated category under the EU AI Act, with transparency obligations for providers.
See the EU AI Act guide.
Generative AI
AI that produces new content, such as text, images or code, rather than only classifying or predicting. It is the technology behind tools like ChatGPT and Claude.
Grounding
Connecting an AI model to your own documents and data so it answers from your information rather than general knowledge. It is the biggest single factor in whether a business AI tool is useful.
Guardrails
Deliberate limits on what an AI system may do or say, built in during development. Policy is different. Policy governs how people use AI, while guardrails constrain the system itself.
H
Hallucination
When an AI model produces confident, plausible output that is simply wrong. It is the defining failure of generative AI. It is the reason grounding and human checks matter.
High Risk AI System
Under the EU AI Act, AI used in areas such as recruitment decisions, credit scoring, critical infrastructure or medical devices. It carries the heaviest compliance obligations.
Human in the Loop
A design where a person reviews or approves an AI system's output before it takes effect. It is standard practice wherever a mistake would be costly or hard to reverse.
I
Inference
The act of an AI model producing an output from an input. Most AI running costs are inference costs, charged per use. That is why usage volume drives ongoing spend.
ISO 42001
The international management standard for artificial intelligence. Clients and partners increasingly ask about it during due diligence.
L
Large Language Model (LLM)
An AI model trained on very large amounts of text, able to understand and produce language. It is the technology behind most current business AI tools.
M
Machine Learning (ML)
Software that improves at a task by finding patterns in data rather than following written rules. It is the foundation of nearly all modern AI.
Model Context Protocol (MCP)
An open standard for connecting AI models to outside tools and data sources in a consistent way. It cuts the custom work each new integration needs.
Multimodal AI
AI that works with more than one type of input, such as text and images together. It is useful for tasks like reading a scanned form or reading a photograph alongside a written description.
N
Natural Language Processing (NLP)
The field of software that understands and works with human language. Large language models are the main approach today.
O
Orchestration
Coordinating several AI steps, tools and systems into one reliable workflow. That includes handling what happens when one part fails.
P
Predictive Analytics
Using historical data to forecast what is likely to happen next, so decisions can be made with foresight rather than hindsight.
See Predictive Analytics.
Predictive Maintenance
Using equipment data to predict failures before they happen. Unplanned downtime becomes scheduled work.
Prompt
The instruction given to an AI model. How a prompt is written has a real effect on the quality of the result. That is why prompting is a teachable skill.
See AI Training.
Prompt Engineering
The practice of writing instructions that reliably get useful results from AI models. It includes structuring the request and supplying the right context.
See AI Training.
Prompt Injection
An attack where hidden instructions in content an AI reads make it behave in ways its owner did not intend. It is a security risk specific to AI systems, and traditional controls do not catch it.
See AI Governance.
R
Responsible AI
Building and using AI in ways that are fair, open and accountable. It goes beyond the minimum a regulation requires.
Retrieval Augmented Generation (RAG)
A technique where an AI model retrieves relevant information from your own documents before answering. Its responses are then based on your material rather than general knowledge.
Robotic Process Automation (RPA)
Software that copies how a person uses a computer screen, clicking and typing through a process. It is older than AI automation and more brittle. It is useful where no API exists.
S
Shadow AI
Staff using AI tools that the business has not approved or does not know about. It is common, and a frequent source of data governance risk.
Structured Data
Information organised in a consistent, machine readable format, such as a database or spreadsheet. The opposite is free text, emails or documents.
Supervised Learning
Training a model on examples where the correct answer is already known. It learns to produce that answer for new cases.
T
Token
The unit AI models process text in, roughly a word fragment. Pricing is usually per token. Longer documents cost more to process than short questions.
Training Data
The information a model learns from. Its quality and coverage largely decide what the model is good at. They also decide where its biases and blind spots lie.
U
Unstructured Data
Information with no consistent machine readable format, such as emails, documents, PDFs and recordings. Most business information is unstructured. That is why AI that can read it is valuable.
W
Workflow Automation
Automating a sequence of connected steps across systems and people, rather than a single isolated task.
See Workflow Automation.
Questions about AI terminology
What is the difference between AI and machine learning?
Machine learning is a branch of artificial intelligence. AI is the broad idea of software performing tasks that normally need human intelligence. Machine learning is one way of doing that. It finds patterns in data rather than following written rules. Nearly all AI in business use today is machine learning.
What is an LLM?
A large language model is an AI model trained on very large amounts of text, able to understand and produce language. LLMs are the technology behind tools like ChatGPT and Claude. They are also behind most business AI applications that read, write or answer questions.
What does it mean when AI "hallucinates"?
Hallucination is when an AI model produces output that is confident, plausible and wrong. It is the defining failure of generative AI. It is the main reason business systems need grounding in real data and a way for people to check important outputs.
What is RAG, and why does it matter?
Retrieval augmented generation is a technique where an AI model looks up relevant information from your own documents before answering. It matters because it is the difference between two kinds of tool. One answers from general internet knowledge. The other answers from your actual policies, records and documentation.
Why is AI charged per token?
Models process text in units called tokens, roughly word fragments. Providers charge by how many are processed. This is why running costs grow with usage, and why long documents cost more to handle than short questions. It is worth estimating before you commit to a build.
Rather ask someone?
Definitions only get you so far. Book a free consultation and ask whatever you want about AI in your business. You get plain English, no obligation and no jargon.
Ask us about a term
The first call is free. Bring any AI term or idea you want explained in plain English.