100 terms
KI-Beratung AI glossary
Overview
This editorial collection explains 100 AI terms the way they are used in real projects: with definition, use fields, industries and an honest decision guide. Every term links the relevant services and is published only after evidence, internal links, FAQ and bilingual parity are checked.
Fundamentals
- AI MVPAn AI MVP is the smallest usable version of an AI solution that real users can apply in daily work.
- AI prototypeAn AI prototype is a deliberately small, quickly built pre-version that tests an idea against real data.
- Artificial intelligenceArtificial intelligence describes methods by which software takes on tasks that would require perception, language or judgement in a person.
- Deep learningDeep learning is a branch of machine learning that uses multi-layer neural networks to learn patterns directly from large amounts of data.
- Generative AIGenerative AI creates new content such as text, images, code or audio by extending patterns learned from large training data.
- Large language models (LLM)Large language models, or LLMs, are AI models trained on very large amounts of text to understand and generate language.
- Machine learningMachine learning is the approach of letting software learn patterns from data instead of hand-coding every rule.
- Multimodal AIMultimodal AI processes several kinds of data in one model, such as text, image, sound and sometimes video together.
- Narrow vs. general AINarrow AI solves tightly scoped tasks, such as summarising text or sorting images.
- Neural networksNeural networks are computational models loosely inspired by how nerve cells interact.
- Vibe codingVibe coding describes building software by describing to an AI assistant in plain language what should be created and largely adopting the generated code.
Agency and consulting
- AI adoptionAI adoption describes the extent to which an AI application is actually used and taken into daily work.
- AI agencyAn AI agency is a service provider that designs and builds AI solutions itself, from first idea through prototype to production.
- AI change managementAI change management covers all measures that make people and processes genuinely adopt a new AI application.
- AI consultingAI consulting clarifies where AI truly has impact in a company, prioritises use cases by value and effort, and delivers a de-risked roadmap from strategy to delivery.
- AI hackathonAn AI hackathon is a tightly time-boxed format in which mixed teams try out concrete AI ideas within a few days and build first working approaches.
- AI literacyAI literacy is people's ability to use AI systems knowledgeably, critically and responsibly.
- AI readinessAI readiness describes how well a company is positioned to use AI reliably.
- AI roadmapAn AI roadmap is a time-ordered plan that translates an AI strategy into concrete steps.
- AI strategyAn AI strategy defines what contribution AI should make to a company's goals and how the initiative is de-risked.
- AI transformationAI transformation describes the longer-term change of a company in which AI permanently reshapes processes, roles and decision paths.
- AI use caseAn AI use case is a concrete application in which AI solves a specific task in the company and creates measurable value.
- AI workshopAn AI workshop is a facilitated format in which a team develops the basics, opportunities and limits of AI together and applies them to its own company.
- Build vs. buyBuild vs.
- Proof of conceptA proof of concept, PoC for short, tests in a tightly bounded frame whether an AI idea works technically and functionally.
Technology and models
- Agent orchestrationAgent orchestration is the coordination of several AI agents that solve a task together.
- Agentic AIAgentic AI describes the approach of giving AI systems a goal and leaving the path to them, instead of prescribing every step.
- AI agentAn AI agent is a software assistant that does not just answer a goal but plans and executes steps on its own to reach it.
- Context windowThe context window is the amount of text a language model can consider at once per request, measured in tokens.
- EmbeddingsEmbeddings are number sequences that represent the meaning of text, images or other data so that similar things sit close together.
- Fine-tuningFine-tuning is the targeted retraining of a pre-trained language model with your own examples so it reliably hits a company's tone, format and terminology.
- Foundation modelA foundation model is a large AI model pre-trained on vast amounts of data that serves as a versatile base for many applications.
- Function callingFunction calling is a language model's ability to produce a structured call to a predefined function instead of plain text.
- HallucinationA hallucination is an answer from a language model that sounds fluent and plausible but is factually wrong or entirely made up.
- InferenceInference is the process by which a trained AI model produces an answer to an input.
- LLMOpsLLMOps refers to the practices and tools used to organise the ongoing operation of language models.
- Model Context Protocol (MCP)The Model Context Protocol (MCP) is an open standard through which a language model accesses external tools, data sources and services.
- Multi-agent systemA multi-agent system consists of several AI agents that divide a task among themselves and exchange results.
- Open-source modelsOpen-source models are AI language models whose weights are openly available and which can be run in your own data centre or with a freely chosen provider.
- Prompt engineeringPrompt engineering is the systematic wording of instructions to a language model so it reliably delivers usable results.
- RAG (Retrieval Augmented Generation)RAG stands for Retrieval Augmented Generation.
- TokensTokens are the small units into which a language model splits text to process it.
- Vector databaseA vector database is a store that holds content as embeddings and finds the most similar ones by content very fast.
Assistants and voice
- AI answering machineAn AI answering machine takes calls when nobody is available and turns the message into text automatically.
- AI assistantAn AI assistant is a software helper that responds to requests in natural language, gathers information and takes on simple tasks.
- AI phone assistantAn AI phone assistant answers inbound calls for a business and handles standard matters automatically.
- ChatbotA chatbot is a text-based dialogue system that takes questions in written language and answers them automatically.
- Company brainA company brain is a central, searchable knowledge layer that connects a firm's documents, data and systems so AI applications answer from verified company knowledge rather than generic training data.
- Conversational AIConversational AI is the umbrella term for systems that converse with people in natural language, by text or voice.
- CopilotA copilot is an AI assistant that supports a person directly inside their work process rather than completing a task on its own.
- Inbound automationInbound automation handles incoming enquiries automatically, whether by phone, email, form or chat.
- Outbound qualificationOutbound qualification actively reaches out to contacts to check their fit for an offer.
- Speech recognition (STT)Speech recognition, also speech-to-text or STT, turns spoken language into written text automatically.
- Text-to-speechText-to-speech, or TTS, turns written text into spoken language automatically.
- Voice agentA voice agent is an AI system that runs a spoken conversation on its own and completes tasks along the way.
- VoicebotA voicebot is a voice-based dialogue system that answers spoken requests along bounded paths.
Automation
- AI automationAI automation combines classic automation with models that process language, images or patterns, so tasks with judgement and unstructured data can run too.
- API integrationAPI integration connects two systems through their designated interfaces so they exchange data in a structured way.
- Exception handlingException handling is the planned way of dealing with errors and special cases in an automated flow.
- HyperautomationHyperautomation describes the systematic interplay of several automation techniques to run whole process chains largely without manual effort.
- Intelligent document processing (IDP)Intelligent document processing, or IDP, reads content out of documents and turns it into structured, usable data.
- MakeMake is a cloud platform for building automations visually.
- n8nn8n is an automation platform that lets you assemble flows from individual building blocks.
- Process automationProcess automation hands whole business flows to software so they run without constant intervention.
- RPARPA stands for robotic process automation.
- WebhookA webhook is a message one system sends automatically to another as soon as a certain event occurs.
- Workflow automationWorkflow automation chains individual work steps into one continuous flow that software runs without manual effort.
- ZapierZapier is a cloud platform that connects web services through simple rules.
Governance and law
- AI auditAn AI audit is a structured review of whether a company's AI systems meet legal requirements, internal rules and quality criteria.
- AI complianceAI compliance is meeting all legal and internal requirements for the use of AI systems, and being able to prove it.
- AI governanceAI governance is the frame that defines who decides on AI in a company, who approves it, and how its use is controlled and documented.
- AI literacy obligation (Art. 4)The AI literacy obligation in Article 4 of the EU AI Act requires providers and deployers of AI to ensure a sufficient level of AI competence among the people who work with the systems.
- AI risk classesAI risk classes are the categories into which the EU AI Act sorts AI systems by their potential for harm.
- Bias in AIBias in AI is a systematic distortion through which an AI system disadvantages or favours certain groups or cases.
- Data protection and AIData protection and AI describes the handling of personal data when training and operating AI systems under the GDPR.
- Data sovereigntyData sovereignty is a company's actual control over its own data, meaning storage location, access, use and disclosure.
- EU AI ActThe EU AI Act is the European regulation that governs the use of AI systems according to their risk.
- EU hostingEU hosting means that data and AI services run on servers within the European Union and thus fall under EU law.
- Explainable AIExplainable AI describes methods that make it traceable how an AI system arrives at a result.
- GuardrailsGuardrails are defined limits that steer what an AI system may accept and output.
- Human in the loopHuman in the loop describes an approach where a person reviews, approves or corrects AI results before they take effect.
- Prompt injectionPrompt injection is an attack on AI systems in which manipulated inputs get a language model to disregard its actual instructions.
Data and knowledge
- Data pipelineA data pipeline is the automated path along which data is pulled from sources, cleaned, reshaped and delivered to a target.
- Data qualityData quality describes how well data fits a given purpose.
- Document extractionDocument extraction reads structured information out of documents, such as invoice number, amount or contract date from PDFs, scans or emails.
- Knowledge baseA knowledge base is an ordered collection of a company's knowledge from which an AI application can draw reliable answers.
- Knowledge graphA knowledge graph is a structured representation of knowledge as a network of entities and their relationships.
- RetrievalRetrieval is the targeted fetching of matching information from a knowledge source before a language model answers.
- Synthetic dataSynthetic data is artificially generated data that mirrors the statistical properties of real data without directly representing real people or events.
- Unstructured dataUnstructured data is information without a fixed format, such as emails, PDF documents, contracts, images or call notes.
Process and operations
- AI implementationAI implementation is the practical delivery of an AI solution from a validated idea to a running system.
- AI monitoringAI monitoring is the ongoing observation of an AI system in operation.
- AI operationAI operation covers everything needed for an AI system to run reliably over time after go-live.
- AI pilotAn AI pilot is a deliberately limited real-world test that checks whether a use case works for real and delivers value.
- AI ROIAI ROI describes the ratio of value to cost of an AI investment.
- BaselineA baseline is the measured state before a change.
- Business process managementBusiness process management (BPM) is the discipline of deliberately designing, steering and continuously improving business processes.
- Process miningProcess mining reconstructs real workflows from the data traces that IT systems leave behind, such as timestamps in ERP or CRM.
- Scaling AIScaling AI means taking a proven solution from pilot into broad, reliable operation.
- Total cost of ownership (AI)Total cost of ownership (TCO) for AI captures all costs of a solution across its entire lifecycle, not just acquisition.
Quality Standard
Every detail page has an extractable short definition, industries and use cases, a decision table, an example, related terms, evidence logic, one clear next step (book a call) and three questions with answers.
Book a strategy call
Book your strategy call directly in the embedded calendar: Book a strategy call. Alternatively: info@ki-beratung.co.
Or send an email to info@ki-beratung.co