15 terms · plain English
The AI terms without the jargon.
If you've ever sat in an AI vendor demo wondering whether they were speaking English, this is for you. Plain-language definitions of the terms that come up when businesses start exploring AI, written for business owners who need a useful answer.
15 terms
Agentic AI
AI that takes actions on your behalf, not just answers questions. An agentic AI can read an email, decide what to do, and actually do it, book the meeting, send the reply, update the system.
AI Agent
A piece of software powered by AI that can carry out tasks autonomously, reading emails, updating records, sending follow-ups, connecting systems.
AI Document Extraction
Using AI (specifically vision and language models) to pull structured data out of documents (invoices, contracts, drawings, forms) automatically. Better than OCR because it understands meaning, not just text.
AI Orchestration
The discipline of running multiple AI components together reliably, choosing which model handles which step, sequencing them, handling failures, monitoring quality. The "ops" layer beneath the visible AI features.
Business Process Automation (BPA)
The broader category that includes workflow automation, RPA, and AI agents, using software to handle business processes that humans currently do manually. Modern BPA increasingly means "AI plus traditional automation."
Computer Vision for Quality Control
Using AI to inspect products visually (catching defects, missing components, wrong configurations) from photos taken with phones or fixed cameras. Faster and more consistent than human visual QC.
ERP System
Enterprise Resource Planning, the central business system that handles accounts, sales, purchasing, stock, and HR. Most SMEs use Sage, Xero, or similar. ERP is the commercial brain; MES is the production brain.
Fine-tuning vs Prompting
Two ways to make an AI model do your specific task. Prompting (giving the model good instructions) is cheap and fast, usually the right answer. Fine-tuning (retraining the model on your data) is expensive and slow, needed in a small number of cases.
Generative AI
AI that creates new content (text, images, code) rather than just classifying or predicting. The category includes ChatGPT, Midjourney, Claude, and the AI features showing up in every business tool.
Internal AI Assistant
A custom AI chat interface that talks to your live business systems, answering "where's job 1647?", "what's due this week?", or "what did Stephen quote that customer in March?" using your real data, not the public internet.
LLM (Large Language Model)
The technology that powers ChatGPT, Claude, and Gemini, AI trained on huge amounts of text that can read, write, summarise, and reason about language.
MES (Manufacturing Execution System)
Software that tracks production on the shop floor, what's being made, where it is in the process, what stage each job is at. The "production-floor brain" of a manufacturing business.
OCR vs AI Document Processing
OCR turns images of text into text. AI document processing turns documents into structured business data, the difference between "we can read the words" and "we know what those words mean for your business."
RAG (Retrieval-Augmented Generation)
A technique that lets AI answer questions about your specific business data (your jobs, your customers, your documents) without retraining the model. The AI looks up the relevant data and uses it to answer.
Workflow Automation
Software that handles the handovers between your systems and teams, so people don't have to copy data between email, CRM, accounts, and operations. Increasingly AI-augmented for the steps that need judgement.
No matching terms. Try a different word or clear the search.
See real use cases.
See how these ideas apply to everyday business work, from handling enquiries to finding information and processing documents.
What could work
better for you?
A new website. More customers. Less admin.
Tell us where you want to get to. We will reply within one working day.