Glossary
The AI glossary
The key AI terms in plain language, explained from our own practice. Not a dictionary, but what a term means and why it matters for your business.
AI technique
Agentic AI Agentic AI is AI that carries out work by itself: it is given a goal, plans the steps, uses systems and tools and checks the result, rather than only answering a question. Agentic workflow An agentic workflow is a process in which an AI agent, not a human, plans the steps, calls tools, checks intermediate results and continues until the goal is reached. AI agent An AI agent is software that pursues a goal on its own: it plans the steps, uses tools or systems, checks the result and adjusts course, without a human directing every step. AI assistant An AI assistant helps a person with a task, such as writing, searching or summarising, but takes no actions by itself unless the user asks. AI engineering AI engineering is the discipline of turning AI models into reliable software that runs in production: connected to your data and systems, tested with evals and managed like any other business-critical system. API An API is a fixed entry point that software opens up to other software, so systems can request and send data without a person in between. Business intelligence (BI) Business intelligence is collecting, combining and visualising company data in reports and dashboards, so you can steer on figures. Data engineering Data engineering is the discipline of extracting data from your systems, cleaning it, combining it and making it reliably available, so reports, analyses and AI applications can run on it without manual work. Data lakehouse A data lakehouse combines a data lake, low-cost storage for every kind of data, with the structure and reliability of a data warehouse, so you can run reporting, analysis and AI in one place. Data platform A data platform is the combination of tools and storage with which an organisation collects, processes, secures and makes available data from all its systems for reporting, analysis and AI. Data quality Data quality is the degree to which your business data is correct, complete, current and consistent, and therefore usable for decisions and for AI. Data warehouse A data warehouse is a central database in which data from different systems is collected, cleaned and organised, so you can run reports and analyses on it. ERP system An ERP system is business software that supports core processes such as finance, purchasing, inventory, production, projects and HR in one coherent system with one database. Eval Evals are automated tests for AI features that measure how accurately, safely and stably they perform on a fixed set of representative examples. Fine-tuning Fine-tuning is training an existing language model further on your own examples, so it fits a specific task, tone or field better. Generative AI Generative AI is AI that creates new content, such as text, images, code or summaries, based on patterns it has learned from large amounts of data. Hallucination A hallucination is when a language model confidently gives information that is incorrect or made up, presented as if it were a fact. Human-in-the-loop Human-in-the-loop means a person reviews or approves an AI process at the right moments, so mistakes are caught before they do any harm. LLM (Large Language Model) An LLM is a large language model trained on enormous amounts of text, which lets it understand and generate language; it is the engine under tools like ChatGPT, Claude and Gemini. Master data management (MDM) Master data management is managing your core data, such as customers, products, suppliers and employees, so that it is the same, complete and current in every system. MCP (Model Context Protocol) MCP is an open standard that lets AI models connect to external tools, data and systems in a uniform way, like a USB port for AI. Microsoft Copilot Microsoft Copilot is the AI assistant Microsoft has built into Word, Excel, Outlook, Teams and other 365 apps, for writing, summarising and searching. Middleware Middleware is software that sits between your systems and translates, checks and passes on data, so applications can work together. Multi-agent system A multi-agent system is a setup in which several AI agents each have their own task and work together towards one goal, often directed by a coordinating agent that divides the work and combines the results. No-code automation No-code automation is building automated workflows between your apps with a visual editor instead of programming code, using tools such as Zapier, Make, n8n and Power Automate. Prompt engineering Prompt engineering is carefully formulating the instruction to an AI model, so it reliably produces the desired result. RAG (Retrieval-Augmented Generation) RAG lets a language model first retrieve the right documents from your own sources and base its answer on those, so it draws on your knowledge instead of only its training. RPA (Robotic Process Automation) RPA lets a software robot carry out fixed, repetitive actions in your systems the way a person would, clicking, copying and retyping, but automatically and without mistakes. Single source of truth A single source of truth means each piece of data is authoritative in one place, so people and systems work with the same, current version of customers, figures and agreements. Sovereign AI Sovereign AI is AI you keep control over: you choose where models and data run, who can access them and which supplier you depend on. System integration System integration is connecting software so that systems such as your CRM, accounting and ERP exchange data automatically instead of someone retyping it. Vector database A vector database stores content as series of numbers that capture meaning, so AI can quickly find the passages that match a question.
Strategy & implementation
AI Act The AI Act is the European law that sets rules for developing and using AI, with stricter requirements the more high-risk an application is. AI audit An AI audit is a focused review of your processes and data that maps where AI saves you the most time or errors, and what it takes to get there. AI business case An AI business case weighs the expected return of an AI application against its costs, so you know in advance whether and when the investment pays off. AI governance AI governance is the set of agreements, roles and controls with which an organisation decides how AI is used, who is responsible and how risks are managed. AI implementation partner An AI implementation partner is an agency that does not just advise on AI but actually builds it, connects it to your systems, and keeps it running in your day-to-day processes. AI operating model An AI operating model is the way an organisation is set up when AI carries out a fixed part of the work: which tasks AI does, who decides and which data and agreements underpin it. AI readiness AI readiness is the degree to which an organisation is ready to use AI meaningfully: the right processes, accessible data, ownership and skills. AI strategy An AI strategy determines which business goals AI should serve, which applications deliver the most and in which order you introduce them. AI use case An AI use case is a concrete application of AI to one clearly defined process or problem, with a clear input, a desired outcome and a measurable return. AIOS (AI Operating System) An AIOS is an AI layer that sits on top of your existing systems, reads data from all of them at once, and runs processes that affect the whole organisation, not just the individual employee. Business process management (BPM) Business process management (BPM) is the structured practice of documenting, measuring and improving the processes in an organisation, so that work runs predictably, efficiently and in a controlled way. Data governance Data governance is the set of agreements on who owns which data, how it is defined, who may view or change it and how long you keep it. Digital transformation Digital transformation is changing how an organisation works, delivers and makes decisions using digital technology, not just digitising what it already did. Enterprise architecture Enterprise architecture is the coherent map of an organisation: which processes exist, which systems and data support them and how everything is connected. Hyperautomation Hyperautomation is automating business processes as broadly as possible by combining techniques such as RPA, workflow automation, integrations and AI, rather than automating individual tasks in isolation. Knowledge management Knowledge management is capturing, organising and making findable the knowledge in an organisation, so everyone can use the same, current information. Lead scoring Lead scoring assigns a score to every lead or prospect based on how well they fit your customer profile and how much interest they show, so sales follows up the most promising ones first. Operational excellence Operational excellence is a way of organising in which every team continuously improves its own processes, so that work runs reliably, efficiently and with as little waste as possible. Order-to-cash Order-to-cash is the process from a customer order to the received payment: order, delivery, invoicing, accounts receivable and matching the payment. Process optimisation Process optimisation is the targeted improvement of a business process so that it runs faster, cheaper or with fewer errors, based on measurement rather than gut feeling. Purchase-to-pay Purchase-to-pay is the purchasing chain from order to payment: purchase order, receipt, purchase invoice, approval and payment to the supplier. Workflow automation Workflow automation lets a series of connected steps, such as a request that arrives, gets checked, approved and processed, run automatically without manual handovers.
Approach & collaboration
AI literacy AI literacy is your employees’ ability to understand AI, use it sensibly, and assess the results critically; since the AI Act it is also a legal obligation. AI policy An AI policy records how an organisation uses AI: which tools are allowed, which data must never go into them, where a person checks and who owns it. AI workshop An AI workshop is a hands-on session in which your team learns what AI can and cannot do and applies it directly to their own work, so the knowledge sticks instead of evaporating. Chief AI Officer (CAIO) A Chief AI Officer is the executive responsible for the use of AI across the organisation: strategy, priorities, governance and results. Proof of Concept (PoC) A proof of concept is a small, fast version of an AI solution that proves the idea works on real data, before you invest in a full build. Shadow AI Shadow AI is employees using AI tools without the organisation knowing or approving it, often with company or client data.
No term found. Try a different search.
Where does most of your time go?
That is where we start.
Tell us which work comes back every week. We show you what AI prepares, and what you only still approve.
Not ready for a conversation yet? See what we built first →
30 minutes. Online or in Enschede. You decide.