KI-Glossar für den Mittelstand.
KI-Fachbegriffe klar und praxisnah erklärt — damit Sie informierte Entscheidungen treffen können.
KI-Wissensgraph
Semantische Datenstruktur die Unternehmenswissen aus ERP, CRM und Dokumenten verknüpft und für KI nutzbar macht.
KI-Betriebssystem
Zentrale Infrastrukturschicht die Wissensgraph, Agenten und Systemintegrationen zu einer unternehmensweiten KI-Plattform verbindet.
DSGVO-konforme KI
KI-Systeme mit geprüfter Rechtsgrundlage, Zweckbindung, Zugriffsschutz, dokumentierter Verarbeitung und wirksamen Betroffenenrechten.
RAG (Retrieval-Augmented Generation)
Retrieval-Augmented Generation (RAG) is a method in which a language model first retrieves relevant passages from a defined knowledge base and then generates its answer on the basis of those passages – instead of relying only on what it learned during training.
LLM (Large Language Model)
A large language model (LLM) is an AI model trained on very large amounts of text that predicts the most likely continuation of a text piece by piece and can therefore answer questions, summarise, translate and write code. Well-known examples are the model families behind ChatGPT, Claude, Gemini or Mistral.
AI agent
An AI agent is a software system in which a language model not only answers but pursues a goal independently: it plans steps, uses tools such as search, e-mail or file access, evaluates the results and continues until the task is done or a human has to decide.
Hallucination (AI)
A hallucination is an answer from an AI model that sounds plausible and confident but is factually wrong or not supported by any source – for example invented figures, quotes, sources or legal provisions.
Embeddings & vector database
Embeddings are numerical vectors into which an AI model translates texts, images or other content so that similar meanings lie close together; a vector database stores these vectors and quickly finds the entries most similar to a query.
Prompt engineering
Prompt engineering is the practice of formulating instructions for AI models (prompts) so that they deliver reliable, usable and verifiable results – through clear tasks, relevant context, examples and a defined output format.
MCP (Model Context Protocol)
The Model Context Protocol (MCP) is an open standard that defines how AI applications connect to external tools and data sources – similar to a universal plug through which an AI assistant can use a CRM, a file storage or a database without a separate integration for each combination.
BYOK (Bring Your Own Key)
Bring Your Own Key (BYOK) means that a company stores its own API key with an AI provider in a software platform, so that model requests from that platform run via the company's own contract and account with the provider – and are billed there.
AVV (data processing agreement, Art. 28 GDPR)
A data processing agreement (in German: Auftragsverarbeitungsvertrag, AVV) is the contract required by Art. 28 GDPR between a controller and a service provider that processes personal data on its behalf and according to its instructions – for example a cloud or AI provider.
On-premise AI
On-premise AI refers to AI systems – in particular language models – that a company operates on its own hardware in its own data centre or server room, instead of using them as a cloud service from an external provider.
How ready is your company for AI?
Three questions about usage, data protection and knowledge. Then you see which concept matters most for you next.