AI agent
A software-based AI capability that can carry out defined tasks or pursue a goal, often by interacting with systems, information and other tools.
AI in Plain English
A practical glossary for leaders working through enterprise AI.
The terminology changes quickly. These definitions focus on what the terms mean in practice, not just technically.
Why Elogica uses “AI-Enabled Transformation”
AI is the enabler. The business change is the point: how work is done, how decisions are made, what people own and where measurable value is created.
AI-first organisation
An organisation that deliberately designs work, decisions, products and operating practices around AI as a core capability rather than treating it as an add-on.
AI governance
The rules, responsibilities and controls that help an organisation use AI safely, responsibly and consistently.
AI-native company
A business built from the outset with AI central to its products, services or the way it operates, rather than adding AI later to an existing business model.
AI operating model
The practical way an organisation decides what AI to pursue, who owns it, how decisions are made, and how AI is managed across the business.
AI transformation
The process of changing how an organisation works, makes decisions, serves customers or creates value by embedding AI into the business, not just adding AI tools.
Agentic AI
AI that can work towards a goal by taking a series of actions, using tools and making decisions along the way rather than only producing a single response.
Copilot
An AI assistant that works alongside a person to help with tasks such as drafting, summarising, analysing, searching or completing parts of a workflow. The term is also used in specific Microsoft product names.
Frontier Firm
Microsoft’s term for an organisation designed around people and AI working together, rather than simply adding AI tools to existing ways of working.
A Microsoft termGrounding
Anchoring an AI response to relevant information or evidence that can be checked, rather than relying only on what the model already knows or generates.
Groundedness
How well an AI response is supported by the information it is meant to rely on. Groundedness can be tested or scored, with thresholds used to flag responses that are not sufficiently supported. In higher-risk or safety-critical uses, it is one control among several.
Human in the loop
A design where a person remains responsible for reviewing, approving or stepping in at important points rather than leaving the AI to act entirely on its own.
Physical AI
AI used in systems that interact with or influence the physical world, such as robotics, autonomous systems, machines, sensors and other real-world environments.
RAG (Retrieval-Augmented Generation)
A way of giving an AI system access to a selected body of information that is stored separately from the underlying model, such as company policies, manuals, product information or internal knowledge. When a question is asked, relevant material is retrieved and supplied to the model to help it answer. It is useful when an organisation wants AI to work with its own current or specialist information.
Value realisation
Making sure an initiative delivers the benefit it was meant to create, and being able to show that it did.
Not sure what these terms actually mean for your organisation?
Elogica can help you work out what matters, what does not, and where AI can create practical value.