Glossary Updates16 new terms added to the glossaries · 12 November 2026, 21:23 CET
AI TechDocKnowledge
  • English (US)
  • English (UK)
  • Deutsch
Luminous signal lines flowing across a deep navy ground in Brand Blue and Ice Blue

AI TechDoc Knowledge | Glossaries and Frameworks

A reference space for the terms, norms and frameworks behind technical documentation in the age of AI.

A multi‑domain reference space for terminology, regulatory frameworks and information‑architecture structures used in technical documentation. Glossaries are cross‑linked through SKOS relations; context cards provide concise, self‑contained explanations of the technical background behind AI‑related documentation tasks, regulatory changes and industry news.

A standalone architecture model for system coordination. Its eight elements, SC1 to SC8, describe how software, control, communication, motion, safety, diagnostics, people and models work together across defined layers and interfaces. The framework is a conceptual model in its own right, separate from the glossaries, and serves as a reference for analysing, designing and discussing coordinated systems.

Saina Veigel's personal space for notes and reflections on clarity, safety and semantics, and on technology that empowers the people who use it. The notes follow the intersection of regulation, technical documentation and product communication, where a precise sentence decides whether a system is understood, trusted and used safely.

NewKnowledge vault Explore the graph (opens in a new tab)

New white paper · free download

From knowledge graphs to adaptive semantics

Evolving DITA and iiRDS digital twins for AI

By Michael Iantosca, Senior Director of Content Platforms and Knowledge Engineering.

AI regulationTechnical documentationSemantic systemsKnowledge graphsMetadataEU AI ActSafety SystemsSecurity SystemsAI regulationTechnical documentationSemantic systemsKnowledge graphsMetadataEU AI ActSafety SystemsSecurity SystemsAI regulationTechnical documentationSemantic systemsKnowledge graphsMetadataEU AI ActSafety SystemsSecurity SystemsAI regulationTechnical documentationSemantic systemsKnowledge graphsMetadataEU AI ActSafety SystemsSecurity Systems

What I cover

Nine fields, three questions, one editorial standard.

Every field answers one of three questions a reader brings to technical knowledge — and serves one of its three readers. The three readers are lenses, not single people: behind each of them stand several roles, from the people who write and check the source to the systems that compile and answer from it. Read a row from left to right, or a column from top to bottom.

The person behind this publication works in four core areas. Each of the nine fields below carries a badge with the core area it belongs to.

The four core areas on sainaveigel.com
  • 01Information architecture and semantics
  • 02Regulated documentation and safety communication
  • 03AI-ready content and AI regulation
  • 04Learning and communication
  1. The source

    Not a reader in itself: the decided content, checked by the people who write, supply, review and release it.

    Is it decided, consistent and traceable?

  2. The person

    Everyone who reads and acts on it — each with a different qualification, authority and situation: the instructed operator, the trained operator, the skilled person, the auditor.

    Can I understand it and act on it safely, here and now?

  3. The third reader

    Every system between the source and the person: integrations, content systems, translations, portals, answer functions and agents. Some read, many compile, a few write back.

    Does the next system know which value counts?

How the three readers break down into roles — writing, deciding, compiling, using and writing back — is described in the blueprint on the role grammar of technical information. The role grammar: readers, writers and compilers

Meaning

What does it mean?

01Core area 01: Information architecture and semantics

The source · the decided content and the people who write and check it

Information architecture

How content is divided, named and related before a single sentence is written.

02Core area 01: Information architecture and semantics

The person · who reads and acts on it — from the instructed operator to the skilled person

Semantic systems

One term, one meaning — across documents, languages and teams.

Evidence

What must it prove?

05Core area 02: Regulated documentation and safety communication

The person · who reads and acts on it — from the instructed operator to the skilled person

Safety communication and evidence

From the hazard in the risk assessment to the warning in the manual — and the proof in between.

Communication

How does it reach people?

08Core area 04: Learning and communication

The person · who reads and acts on it — from the instructed operator to the skilled person

Visual and multimedia communication

What a diagram, a map or a short film explains better than a paragraph.

09Core area 04: Learning and communication

The third reader · systems that pass it on, compile it or answer from it — AI is one of them

Technical and product marketing

Product knowledge turned into messages that match the documentation.

From thinking model to check method

Three readers find the questions. Check rules need roles.

The three readers above are a thinking model: enough to design a check method. To build your own check processes and check rules, each placeholder has to become a role — and each role a system, subsystem, function or subfunction a rule can name.

  1. 1

    Thinking model

    Three placeholders — the source, the person, the third reader — find the questions a check has to ask.

    The nine fields
  2. 2

    Roles

    Each placeholder opens into roles: who writes, decides, compiles, uses or writes back a value.

    Who decides which value counts
  3. 3

    Check rule

    Each role gets a unit, a value, a condition, evidence, an owner and a trigger.

    The check rule grammar
One check rule

CCMS › publishing subsystem › assembly function › revision lookup — the revision of every assembly in the order manual equals the revision released in PLM — evidence: the build record.

A placeholder finds the question. A role makes it checkable.

From the newsletter

Recent articles on AI TechDoc Press

All articles on Substack
Cover image of the article “Was die Maschinen-Risikobeurteilung von der Medizintechnik lernen kann”

11 November 2026

Was die Maschinen-Risikobeurteilung von der Medizintechnik lernen kann

Gefährdungsszenarien verbinden reale Handlungsabläufe mit Schutzmaßnahmen und Sicherheitsnachweisen

Read on Substack
Cover image of the article “What Machinery Risk Assessment can learn from MedTech”

11 November 2026

What Machinery Risk Assessment can learn from MedTech

Hazard-related scenarios connect real-world activities with protective measures and safety evidence

Read on Substack
Cover image of the article “AI TechDoc Friday: Cyberattacks on refrigeration systems in Europe”

6 November 2026

AI TechDoc Friday: Cyberattacks on refrigeration systems in Europe

How manipulated controllers put cold chains at risk, and which safeguards help

Read on Substack
Cover image of the article “AI TechDoc Freitag: Cyberangriffe auf Kühlanlagen in Europa”

6 November 2026

AI TechDoc Freitag: Cyberangriffe auf Kühlanlagen in Europa

Wie manipulierte Steuerungen Kühlketten gefährden und welche Schutzmaßnahmen helfen

Read on Substack
Cover image of the article “Who can judge an AI model output? Competence as the weak point of an AI-enabled QMS”

30 October 2026

Who can judge an AI model output? Competence as the weak point of an AI-enabled QMS

An AI tool in use is not a competence

Read on Substack
Cover image of the article “KI im QMS: Der Prüfmaßstab verschiebt sich von der Wirkung auf die Beherrschung”

29 October 2026

KI im QMS: Der Prüfmaßstab verschiebt sich von der Wirkung auf die Beherrschung

Zwei Blickrichtungen auf denselben Gegenstand

Read on Substack
Cover image of the article “EN 18286 certifies the system, not the company: An AI QMS that holds up in an audit”

29 October 2026

EN 18286 certifies the system, not the company: An AI QMS that holds up in an audit

The first harmonized AI standard covers a very narrowly defined area

Read on Substack
Cover image of the article “Die erste harmonisierte KI-Norm ist veröffentlicht – die Aufsicht fehlt noch in mehreren Ländern”

28 October 2026

Die erste harmonisierte KI-Norm ist veröffentlicht – die Aufsicht fehlt noch in mehreren Ländern

KW43: Was ein KI-Qualitätsmanagementsystem nach Artikel 17 dokumentieren muss, damit es einer Prüfung standhält.

Read on Substack
Cover image of the article “AI deployment is becoming an institutional design problem”

28 October 2026

AI deployment is becoming an institutional design problem

AI adoption depends on the institutions around it — a global challenge

Read on Substack
Cover image of the article “AI deals and deployments across Southeast Asia, Europe, and the US”

28 October 2026

AI deals and deployments across Southeast Asia, Europe, and the US

A new AI lab and fresh deals mark the fortnight

Read on Substack
Cover image of the article “Before an AI recommendation becomes a machine command”

24 October 2026

Before an AI recommendation becomes a machine command

Industrial AI needs a documented handover from interpretation to action

Read on Substack
Cover image of the article “Wenn KI-Agenten Grenzen überschreiten – 12. Oktober – 17. Oktober 2026”

21 October 2026

Wenn KI-Agenten Grenzen überschreiten – 12. Oktober – 17. Oktober 2026

KW42: Agenten erreichen reale Systeme – entscheidend sind Zugriff, Eindämmung und Wiederanlauf.

Read on Substack

AI TechDoc Press on Substack

Stay ahead of what AI means for your documentation.

New articles land several times a week — analysis of regulation as it takes shape, practical guidance for documentation teams and a weekly look at international AI news. Free to read, straight to your inbox.

  • Timely analysis of the EU AI Act and machinery regulation
  • Practical guidance for documentation and product teams
  • Weekly roundup of international AI news
  • Premium deep dives for subscribers