Practical Guides · 3 MIN READ
AI glossary in plain language
Short explanations of common terms, with their practical implications.
Prompt
The instructions or input you give a model for a task. A useful prompt describes the outcome, audience and constraints.
Context and context window
Context is the information available to the model for a response. The context window is the amount it can process at once. Being included in context does not guarantee a detail will be used correctly.
Tokens
Units used to represent text for a model, often words or parts of words. Token counts affect how much text fits into a request.
Hallucination
A generated claim that is false or unsupported, even when expressed confidently. Check consequential claims against original evidence.
Retrieval-augmented generation (RAG)
An approach that retrieves external material to inform generation. Retrieved content still needs checks for relevance, currency and support for the claim.
Workflow and agent
A workflow follows a defined sequence. An agent lets a model choose parts of the next action or tool use. In either case, permissions and review points need deliberate design.
Grounding
Connecting a response to supplied information or evidence. A grounded response still needs checking: a source can be misinterpreted or incomplete.
Explore the webinar for guided practice, or subscribe for new perspectives.
Explore The Webinar ↗Subscribe Newsletter ↗