{"id":743,"date":"2026-10-29T20:59:25","date_gmt":"2026-10-29T19:59:25","guid":{"rendered":"https:\/\/stefan-quisdorf.de\/?page_id=743"},"modified":"2026-10-29T21:25:13","modified_gmt":"2026-10-29T20:25:13","slug":"ai-glossary","status":"publish","type":"page","link":"https:\/\/stefan-quisdorf.de\/en\/ai-glossary\/","title":{"rendered":"AI glossary"},"content":{"rendered":"\n
\n

AI glossary<\/p>\n\n\n\n

AI terms, explained simply.<\/mark><\/h1>\n\n\n\n

The terms that come up most often in my conversations with managing directors of small and mid-sized companies \u2013 briefly explained, with what they mean for your day-to-day work.<\/p>\n\n\n\n

Jump to:<\/strong> Generative AI<\/a> \u00b7 Large Language Model (LLM)<\/a> \u00b7 Prompt and prompting<\/a> \u00b7 Hallucination<\/a> \u00b7 Shadow AI<\/a> \u00b7 AI policy<\/a> \u00b7 AI literacy (Art. 4 AI Act)<\/a> \u00b7 EU AI Act<\/a> \u00b7 Data protection with AI tools<\/a> \u00b7 AI agent<\/a> \u00b7 RAG (retrieval augmented generation)<\/a> \u00b7 AI maturity<\/a><\/p>\n\n\n\n

As of October 2026 \u00b7 Author: Stefan Quisdorf, AI Professional (IHK)<\/p>\n<\/section>\n\n\n\n

\n
\n

Basics<\/p>\n\n\n\n

Generative AI<\/h2>\n\n\n\n

Generative AI creates new content \u2013 text, images, tables or program code \u2013 based on patterns it has learned from large amounts of data. Well-known examples are ChatGPT, Microsoft Copilot, Google Gemini and Claude.<\/p>\n\n\n\n

In practice:<\/strong> In small and mid-sized companies, generative AI helps above all with drafts, summaries and routine texts. The decision and the review remain with people.<\/p>\n<\/div>\n\n\n\n

\n

Basics<\/p>\n\n\n\n

Large Language Model (LLM)<\/h2>\n\n\n\n

A large language model is a language model trained on very large amounts of text. It predicts, word by word, what is most likely to come next \u2013 which is why the result sounds fluent but is not automatically correct.<\/p>\n\n\n\n

In practice:<\/strong> What matters in practice: an LLM only \u201cknows\u201d what was in its training data or what you give it. It does not know current prices or internal processes on its own.<\/p>\n<\/div>\n\n\n\n

\n

Application<\/p>\n\n\n\n

Prompt and prompting<\/h2>\n\n\n\n

A prompt is the instruction you give an AI. Prompting means phrasing that instruction so that a useful result comes out: with a goal, context, examples and the desired format.<\/p>\n\n\n\n

In practice:<\/strong> Good prompts are not secret knowledge. If you say who the text is for, how long it should be and what must not be included, you get noticeably better results.<\/p>\n\n\n\n

More: AI workshop for teams<\/a><\/p>\n<\/div>\n\n\n\n

\n

Risks<\/p>\n\n\n\n

Hallucination<\/h2>\n\n\n\n

Hallucination is when an AI invents false facts, figures or sources and presents them convincingly. This is not a rare glitch but a consequence of how language models work.<\/p>\n\n\n\n

In practice:<\/strong> That is why results must be checked before they reach the customer \u2013 especially figures, legal questions and quotes.<\/p>\n\n\n\n

More: AI literacy in the team<\/a><\/p>\n<\/div>\n\n\n\n

\n

Risks<\/p>\n\n\n\n

Shadow AI<\/h2>\n\n\n\n

Shadow AI is the use of AI tools without the company’s knowledge or approval \u2013 often with private accounts. It usually does not stem from bad intent but from the lack of rules and approved tools.<\/p>\n\n\n\n

In practice:<\/strong> Bans rarely help. Approved tools, simple ground rules and training that shows how to do it right work better.<\/p>\n\n\n\n

More: AI literacy package<\/a> \u00b7 AI maturity check<\/a><\/p>\n<\/div>\n\n\n\n

\n

Rules<\/p>\n\n\n\n

AI policy<\/h2>\n\n\n\n

An AI policy defines which AI tools are allowed in the company, which data may go into them, who checks results and who the contact person is. One page is often enough to start with.<\/p>\n\n\n\n

In practice:<\/strong> The policy is not an end in itself: it gives the team confidence and turns many individual experiments into a shared approach.<\/p>\n\n\n\n

More: AI literacy package with written ground rules<\/a><\/p>\n<\/div>\n\n\n\n

\n

Rules<\/p>\n\n\n\n

AI literacy (Art. 4 AI Act)<\/h2>\n\n\n\n

AI literacy means the knowledge and skills to use AI systems competently and to assess their opportunities and risks. Since February 2025, Article 4 of the EU AI Act has required companies to take care of this; since the \u201cDigital Omnibus\u201d of 27 July 2026, they must promote their staff’s AI literacy but no longer guarantee a specific level.<\/p>\n\n\n\n

In practice:<\/strong> In practical terms: suitable training, simple ground rules and a short record of who took part and when. This is not legal advice \u2013 if in doubt, ask your legal adviser.<\/p>\n\n\n\n

More: Building AI literacy in the team<\/a><\/p>\n<\/div>\n\n\n\n

\n

Rules<\/p>\n\n\n\n

EU AI Act<\/h2>\n\n\n\n

The EU AI Act is the European regulation on artificial intelligence. It classifies AI systems by risk and sets strict requirements for high-risk applications \u2013 for example in recruitment or lending.<\/p>\n\n\n\n

In practice:<\/strong> For most office applications in small and mid-sized companies, such as texts, summaries or research, the AI literacy obligation is what matters most \u2013 no reason to panic.<\/p>\n\n\n\n

More: What the AI Act requires<\/a><\/p>\n<\/div>\n\n\n\n

\n

Rules<\/p>\n\n\n\n

Data protection with AI tools<\/h2>\n\n\n\n

As soon as personal data or trade secrets go into an AI tool, the GDPR and confidentiality apply. What matters is whether the provider uses the inputs for training, where the data is processed and whether a data processing agreement (DPA) is in place.<\/p>\n\n\n\n

In practice:<\/strong> Rule of thumb for everyday work: use business versions with a DPA, no customer data in private accounts, and anonymise if in doubt.<\/p>\n\n\n\n

More: AI workshop: working in line with the GDPR<\/a><\/p>\n<\/div>\n\n\n\n

\n

Application<\/p>\n\n\n\n

AI agent<\/h2>\n\n\n\n

An AI agent does not just give a single answer but carries out several steps on its own \u2013 such as researching, filling in a table and drafting an email. To do so, it is given access to tools like calendars, files or shop systems.<\/p>\n\n\n\n

In practice:<\/strong> Agents save a lot of time but need clear limits: what may the agent do on its own, and what does a person have to approve?<\/p>\n<\/div>\n\n\n\n

\n

Application<\/p>\n\n\n\n

RAG (retrieval augmented generation)<\/h2>\n\n\n\n

With RAG, the AI searches your own documents \u2013 such as manuals, price lists or FAQs \u2013 before answering and bases its answer on them. It answers from your knowledge rather than only from its training data.<\/p>\n\n\n\n

In practice:<\/strong> Typical uses are internal knowledge bases and customer service. The quality depends heavily on how up to date and clean the documents are.<\/p>\n<\/div>\n\n\n\n

\n

Strategy<\/p>\n\n\n\n

AI maturity<\/h2>\n\n\n\n

AI maturity describes how far a company has come in using AI. I distinguish four stages: experimenting, rules, applying and embedding.<\/p>\n\n\n\n

In practice:<\/strong> Most small and mid-sized companies are somewhere between the first and second stage \u2013 a good starting point for a roadmap.<\/p>\n\n\n\n

More: AI self-check (5 minutes, free)<\/a> \u00b7 AI maturity check<\/a><\/p>\n<\/div>\n<\/section>\n\n\n\n