AI Glossary
All key terms around AI, automation, and data infrastructure — explained for decision-makers, not computer scientists.
80 Terms
AI Fundamentals
Core concepts and technologies behind modern artificial intelligence.
Retrieval-Augmented Generation
Retrieval-Augmented Generation (RAG) is an AI architecture that connects a Large Language Model with an external knowledge base.
→AI FundamentalsLarge Language Model
A Large Language Model (LLM) is a neural network with billions of parameters trained on vast text corpora.
→AI FundamentalsNatural Language Processing
Natural Language Processing (NLP) is the AI subfield that enables computers to understand, analyze, and generate human language.
→AI FundamentalsTransformer Architecture
The Transformer architecture is the neural network design behind all modern Large Language Models.
→AI FundamentalsEmbedding
An embedding is the mathematical representation of text, images, or other data as a numerical vector in a high-dimensional space.
→AI FundamentalsFine-Tuning
Fine-tuning is the targeted retraining of a pre-trained AI model on domain-specific data.
→AI FundamentalsPrompt Engineering
Prompt engineering is the systematic design of instructions (prompts) to an AI model to achieve precise and reliable results.
→AI FundamentalsHallucination
An AI hallucination occurs when a Large Language Model generates plausible-sounding but factually incorrect information.
→AI FundamentalsGenerative AI
Generative AI (GenAI) refers to AI systems that can create new content — text, images, code, audio, or video.
→AI FundamentalsNeural Network
A neural network is an AI model that — inspired by the human brain — consists of interconnected layers of artificial neurons.
→AI FundamentalsDeep Learning
Deep learning is a subset of machine learning that uses neural networks with many layers (hence "deep") to recognize complex patterns in large datasets.
→AI FundamentalsMachine Learning
Machine Learning (ML) is a subfield of artificial intelligence where algorithms learn from data to recognize patterns and make predictions — without being explicitly programmed.
→AI FundamentalsReinforcement Learning
Reinforcement Learning (RL) is a learning paradigm where an AI agent learns through trial and error to make optimal decisions in an environment.
→AI FundamentalsTransfer Learning
Transfer learning is an ML technique where a model pre-trained on large datasets is transferred to a new, more specific task.
→AI FundamentalsFoundation Model
A Foundation Model is a large AI model pre-trained on broad datasets that serves as a base for diverse downstream tasks.
→Agentic AI
Autonomous AI systems that plan, decide, and act independently.
AI Agent
An AI agent is an autonomous software system that independently plans tasks, makes decisions, and executes actions — without step-by-step human guidance.
→Agentic AIAgentic Workflow
An agentic workflow is a multi-step, AI-driven work process where one or more AI agents autonomously plan, execute, and decide their next step based on intermediate results.
→Agentic AIMulti-Agent System
A multi-agent system (MAS) is an AI architecture where multiple specialized agents collaborate to solve complex tasks.
→Agentic AITool Calling
Tool calling is the ability of an AI model to autonomously invoke external tools — querying databases, calling APIs, or processing files.
→Agentic AIChain-of-Thought
Chain-of-Thought (CoT) is a prompting technique that instructs an AI model to reveal its reasoning step by step.
→Agentic AIFunction Calling
Function calling is a standardized protocol through which an LLM generates structured function calls instead of free text.
→Agentic AIAutonomous Agent
An autonomous agent is an AI system that plans, executes, and monitors task progress without continuous human oversight.
→Agentic AIHuman-in-the-Loop
Human-in-the-Loop (HITL) is an architecture pattern where an AI agent requests human approval at defined decision points before proceeding.
→Agentic AIAgent Orchestration
Agent orchestration is the architecture pattern that defines how multiple AI agents are coordinated, prioritized, and managed.
→Agentic AIReflection Pattern
The Reflection Pattern is an agentic AI design pattern where an AI agent critically evaluates its own outputs and iteratively improves them.
→Agentic AIPlanning & Reasoning
Planning & Reasoning describes an AI agent's ability to decompose complex tasks into sub-steps, identify dependencies, and determine an optimal execution sequence.
→Agentic AIMemory Management
Memory Management in Agentic AI describes the techniques by which AI agents store, retrieve, and prioritize information beyond individual interactions.
→Agentic AIGuardrails
Guardrails are safety mechanisms that keep AI agent behavior within defined boundaries.
→Agentic AIAgent Evaluation
Agent evaluation encompasses the systematic assessment of AI agents regarding accuracy, reliability, cost, and business impact.
→Data & Infrastructure
Data storage, processing, and the technical foundation for AI systems.
Vector Database
A vector database is a specialized database system that stores high-dimensional numerical vectors (embeddings) and enables lightning-fast similarity searches.
→Data & InfrastructureETL Pipeline
An ETL pipeline (Extract, Transform, Load) is an automated data process that extracts raw data from various sources, transforms it into a unified format, and loads it into a target system.
→Data & InfrastructureData Lake
A data lake is a central storage system that ingests structured, semi-structured, and unstructured data in its raw format — without prior schema adaptation.
→Data & InfrastructureData Warehouse
A data warehouse (DWH) is a central, schema-based storage system for processed and structured business data.
→Data & InfrastructureKnowledge Graph
A knowledge graph is a data structure that represents knowledge as a network of entities (nodes) and their relationships (edges).
→Data & InfrastructureSemantic Search
Semantic search is a search technology that looks for the meaning of queries rather than exact keywords.
→Data & InfrastructureData Quality
Data quality describes the degree to which enterprise data is complete, correct, consistent, current, and usable.
→Data & InfrastructureData Governance
Data governance encompasses the policies, processes, and responsibilities that ensure enterprise data is managed correctly, securely, accessibly, and compliantly.
→Data & InfrastructureData Pipeline
A data pipeline is the overarching architecture that transports data from source to target system while orchestrating all necessary processing steps.
→Data & InfrastructureFeature Store
A feature store is a central system for managing, storing, and serving ML features — the processed data points that serve as input for machine learning models.
→Data & InfrastructureData Mesh
Data Mesh is a decentralized architecture principle where data responsibility shifts from central IT teams to business domains (e.
→Data & InfrastructureData Catalog
A data catalog is a central directory that inventories, describes, and makes searchable all available data assets in an organization.
→Data & InfrastructureStreaming Architecture
A streaming architecture processes data continuously in real time as it is created — unlike batch processing, which handles data periodically in batches.
→Enterprise AI
AI in enterprises: governance, compliance, and strategic implementation.
AI Strategy
An AI strategy is the structured plan through which a company purposefully deploys artificial intelligence for value creation.
→Enterprise AIAI Readiness
AI readiness describes a company's maturity level for successfully introducing and productively using artificial intelligence.
→Enterprise AIERP AI Integration
ERP AI integration refers to connecting AI systems with Enterprise Resource Planning software such as SAP, Microsoft Dynamics, or Odoo.
→Enterprise AIPredictive Analytics
Predictive Analytics uses statistical models and machine learning to derive predictions about future events from historical business data.
→Enterprise AIAI Governance
AI governance encompasses the organization-wide policies, processes, and responsibilities for the responsible use of artificial intelligence.
→Enterprise AIMLOps
MLOps (Machine Learning Operations) applies DevOps principles to the entire ML model lifecycle: training, validation, deployment, monitoring, and retraining.
→Enterprise AIIntelligent Document Processing
Intelligent Document Processing (IDP) combines OCR, NLP, and machine learning to automatically classify unstructured documents — invoices, contracts, delivery notes, emails — extract relevant information, and convert it into structured data.
→Enterprise AIConversational AI
Conversational AI refers to AI systems that can conduct natural conversations with humans — via text or voice.
→Enterprise AIAI ROI
AI ROI (Return on Investment) measures the economic return of an AI investment relative to total costs.
→Enterprise AIChange Management for AI
Change management for AI encompasses all organizational measures ensuring that AI technologies are accepted, understood, and productively used by employees.
→Enterprise AIAI Compliance
AI compliance encompasses adherence to all regulatory requirements when deploying artificial intelligence — from the EU AI Act through GDPR to industry-specific regulations.
→Enterprise AIEnterprise AI Platform
An Enterprise AI Platform is a centralized technical infrastructure that consolidates all AI activities of a company: model hosting, data pipelines, experiment tracking, deployment, monitoring, and governance.
→Enterprise AIAI Maturity Model
An AI maturity model describes the development stages a company passes through on the path to full AI integration — from initial experiments through productive individual applications to enterprise-wide AI adoption.
→Automation
Process automation — from classic RPA to AI-powered workflows.
Workflow Automation
Workflow automation refers to the digital mapping and automatic execution of recurring business processes — from order entry to invoice approval.
→AutomationProcess Mining
Process mining is a data-driven analysis method that reconstructs and visualizes actual business process flows from event logs in your IT systems (ERP, CRM, ticketing).
→AutomationRobotic Process Automation
Robotic Process Automation (RPA) is a technology where software robots mimic human inputs on user interfaces — clicking, typing, copying, pasting.
→AutomationIntelligent Automation
Intelligent Automation (IA) combines classical automation with artificial intelligence — particularly NLP, computer vision, and machine learning.
→AutomationDocument Automation
Document automation encompasses AI-powered capture, classification, data extraction, and processing of business documents — invoices, contracts, delivery notes, forms.
→AutomationEmail Automation
Email automation is the AI-powered processing of incoming and outgoing business emails — from automatic classification and routing through data extraction to generating context-aware responses.
→AutomationChatbot
A chatbot is an AI-powered dialogue system that communicates with users via text or voice.
→AutomationVoice Agent
A voice agent is an AI system that can independently conduct phone conversations — with natural language processing, real-time speech synthesis, and the ability to trigger actions in connected systems.
→AutomationSales Automation
Sales automation refers to AI-powered automation of sales processes — from lead generation through qualification and outreach to follow-up.
→AutomationInvoice Processing
Automated invoice processing encompasses AI-powered capture, validation, account assignment, and approval of incoming invoices.
→AutomationNo-Code Automation
No-code automation enables business departments to automate processes through visual drag-and-drop interfaces — without programming skills.
→AutomationHyperautomation
Hyperautomation is a strategic approach that systematically identifies and end-to-end automates all automatable business processes — through orchestrated combination of AI, RPA, process mining, no-code tools, and classical software development.
→AutomationTask Mining
Task mining analyzes the actual work steps of individual employees at desktop level — clicks, keystrokes, application switches — to identify repetitive patterns and automation potential.
→Web & Software
Modern software architectures and web technologies.
API Integration
An API integration connects two or more software systems through standardized programming interfaces (Application Programming Interfaces).
→Web & SoftwareHeadless CMS
A Headless CMS is a content management system that delivers content exclusively through APIs — without a tightly coupled frontend.
→Web & SoftwareJamstack
Jamstack is a web architecture based on JavaScript, APIs, and Markup.
→Web & SoftwareProgressive Web App
A Progressive Web App (PWA) is a web application that behaves like a native app: it can be installed on the home screen, works offline, and sends push notifications.
→Web & SoftwareMicroservices
Microservices are a software architecture where an application consists of small, independently deployable services.
→Web & SoftwareServerless Computing
Serverless computing is a cloud delivery model where the provider fully manages the server infrastructure.
→Web & SoftwareCI/CD Pipeline
A CI/CD pipeline (Continuous Integration / Continuous Deployment) is an automated process that automatically tests, builds, and delivers code changes to the production environment.
→Web & SoftwareDesign System
A Design System is a collection of reusable UI components, design tokens, and design guidelines that ensure a consistent user interface across all digital products.
→Web & SoftwareWeb Performance
Web performance describes the speed and responsiveness of a website from the user perspective.
→Web & SoftwareWeb Accessibility (WCAG)
Web accessibility means websites and applications are usable by all people — including those with visual, auditory, motor, or cognitive disabilities.
→Web & SoftwareComposable Architecture
Composable Architecture is an architectural approach where digital systems are assembled from interchangeable, best-of-breed components — instead of relying on a monolithic all-in-one platform.
→Web & SoftwareEdge Computing
Edge Computing moves data processing from the central data center to the edge of the network — closer to the end user.
→Bereit für Ihr nächstes Projekt?
Lassen Sie uns gemeinsam Ihre Vision verwirklichen. Kontaktieren Sie uns für eine unverbindliche Beratung.