Trusted Context World Publishing - So that what matters is found - and what is meant is understood.

Investments in AI alone do not guarantee results.

Trusted Context World makes the meanings, structures, and relationships within your business world accessible to AI. This enables AI to be reliably integrated into work and processes. Business information is made understandable, statements verifiable, and interoperable according to defined rules for:

  • traditional internal workflows and new AI-supported processes,
  • external AI-supported information-, value creation-, and decision-making processes.

In Preparation.

Context World

Read the Executive Summary


How resilient is your company's digital information?

Does your company's digitally available information already serve as a reliable basis for work and decision-making?

  • Do employees come up with the same answers?
  • Are the sources verifiable?
  • Is the information up to date?
  • Are responsibilities clearly defined?
  • Can AI reliably derive the same conclusions?
  • How do you ensure consistency across the website, SharePoint, and documents?
  • Does your company's online presence accurately reflect the reality of your business?
  • Is your company being understood correctly by ChatGPT, Claude, and other AI systems?


If you answer “No” to one or more of these questions, it's worth building your Context World.


How can you ensure that company data remains relevant both internally and externally and is used properly?

Trusted Context World Publishing refers to the combination of procedures, working methods, and the tools used for this purpose

Company information from documents, the intranet, the Internet, and other sources is continuously maintained and expanded - for example, in the CMS, Microsoft 365, or SharePoint. In Trusted Context World Publishing, this information is structured into a reliable corporate context. Using the Context World Development System (CWDS), a license-free, machine-readable CWDS index can be generated and integrated for various target systems in accordance with policies. Implementation and maintenance can be provided as a service by Tolksdorf.digital or handled entirely in-house by the company. ​


AI and large language models (LLMs), as well as their application methods, are constantly evolving, but Context remains crucially important as a foundation

The evolution of chatbots into Agentic AI shifts the focus from crafting individual prompts to a reliable business context, clear objectives, and organized collaboration between humans and AI. The following overview highlights the most important changes.

Begriffswelt der ersten erfolgreichen LLM, ChatGPT-3/4, usw. Begriffswelt heute Was sich verändert hat
Chatbot Agentic AI KI beantwortet nicht nur Fragen, sondern verfolgt Ziele und bearbeitet Aufgaben über mehrere Schritte.
Prompt Auftrag Nicht die einzelne Formulierung zählt, sondern das zu erreichende Ergebnis.
Prompt Engineering Context Engineering Entscheidend ist weniger der perfekte Satz als der verfügbare, verständliche und verlässliche Kontext.
Gute Frage stellen Ziel und Erfolgskriterien definieren Die KI muss wissen, was erreicht werden soll und woran ein gutes Ergebnis erkennbar ist.
Ein Prompt – eine Antwort Iterativer Arbeitsprozess Die KI sucht, prüft, verwirft, verbessert und setzt ihre Arbeit fort.
Antwort generieren Lösungsweg verfolgen Das Modell produziert nicht nur Text, sondern entwickelt und überprüft Handlungsoptionen.
Halluzination vermeiden Verlässlichkeit organisieren Fehlerfreiheit entsteht nicht allein durch Formulierungen, sondern durch Quellen, Regeln, Prüfungen und Rückkopplungen.
Möglichst genauer Prompt Möglichst guter Arbeitsraum Daten, Werkzeuge, Zuständigkeiten, Regeln und Bedeutungen müssen verfügbar sein.
Mensch formuliert jeden Schritt Mensch gibt Richtung und Grenzen vor Die operative Steuerung wandert zunehmend zur KI.
Human-in-the-Loop bei jedem Schritt Human-on-the-Loop Der Mensch überwacht, greift aber nicht mehr zwangsläufig in jeden Zwischenschritt ein.
Kontrolle durch Freigaben Kontrolle durch Leitplanken Regeln, Berechtigungen, Tests und Abbruchkriterien ersetzen einen Teil manueller Freigaben.
KI als Werkzeug KI als Companion oder Akteur KI wird vom punktuell verwendeten Hilfsmittel zum kontinuierlichen Arbeitspartner.
Einzelne Aufgabe automatisieren Arbeitsfähigkeit bereitstellen Nicht nur Tätigkeiten, sondern ganze Aufgabenräume werden an KI-Systeme übergeben.
Wissen in den Prompt kopieren Wissen strukturiert zugänglich machen Kontext muss dauerhaft auffindbar, maschinenlesbar und anschlussfähig sein.
RAG als Zusatzfunktion Verlässlicher Unternehmenskontext als Grundlage Informationszugriff allein genügt nicht; Versionen, Bedeutungen und Gültigkeiten müssen geklärt sein.
Möglichst viel Kontext Relevanter und vertrauenswürdiger Kontext Mehr Information ist nicht automatisch besser. Entscheidend sind Auswahl, Herkunft und Bedeutung.
Output kontrollieren Prozess und Ergebnis evaluieren Nicht nur der Text, sondern auch Suchweg, Quellen, Entscheidungen und Nebenwirkungen werden geprüft.
KI bedienen lernen Zusammenarbeit mit KI organisieren Die Herausforderung wird organisatorisch, nicht nur technisch.
Nutzerkompetenz Delegationskompetenz Menschen müssen lernen, Ziele, Verantwortung und Entscheidungsspielräume sinnvoll zu übertragen.
Modell auswählen System gestalten Modelle werden Bestandteile aus Daten, Tools, Regeln, Rollen, Workflows und Governance.
Datenschutz verhindert Nutzung Risikoadäquate Autonomie ermöglichen Nicht maximale Einschränkung, sondern kontrollierte Handlungsfähigkeit wird zum Ziel.
Proof of Concept Produktiver KI-Arbeitsprozess Entscheidend ist nicht mehr, ob KI etwas zeigen kann, sondern ob sie dauerhaft wirksam arbeitet.
KI erzeugt Inhalte KI nimmt an Wertschöpfung teil KI wird Teil von Informations-, Arbeits-, Entscheidungs- und Leistungsketten.
Gute Prompts skalieren Guter Kontext skaliert Einzelne Formulierungen sind flüchtig; verlässliche Bedeutungs- und Wissensgrundlagen sind wiederverwendbar.
Terminology related to the first successful LLMs, ChatGPT-3/4, etc. Terminology today What has changed
Chatbot Agentic AI AI no longer only answers questions; it pursues goals and works on tasks across multiple steps.
Prompt Assignment What matters is not the individual wording, but the result to be achieved.
Prompt Engineering Context Engineering What matters is less the perfect sentence than the available, understandable, and reliable context.
Ask a good question Define the goal and success criteria AI must know what is to be achieved and how a good result can be recognized.
One prompt – one answer Iterative work process AI searches, checks, discards, improves, and continues its work.
Generate an answer Pursue a solution path The model does not merely produce text; it develops and evaluates possible courses of action.
Avoid hallucinations Organize reliability Accuracy does not result from wording alone, but from sources, rules, checks, and feedback loops.
The most precise prompt possible The best possible workspace Data, tools, responsibilities, rules, and meanings must be available.
Humans formulate every step Humans define direction and boundaries Operational control is increasingly shifting to AI.
Human-in-the-Loop at every step Human-on-the-Loop Humans supervise but no longer necessarily intervene in every intermediate step.
Control through approvals Control through guardrails Rules, permissions, tests, and stopping criteria replace some manual approvals.
AI as a tool AI as a companion or actor AI evolves from an occasionally used aid into a continuous work partner.
Automate an individual task Provide operational capability Not only individual activities, but entire task domains are delegated to AI systems.
Copy knowledge into the prompt Make knowledge structurally accessible Context must remain discoverable, machine-readable, and reusable across systems.
RAG as an additional feature Reliable business context as a foundation Access to information alone is not enough; versions, meanings, and validity must be clarified.
As much context as possible Relevant and trustworthy context More information is not automatically better. Selection, origin, and meaning are decisive.
Control the output Evaluate the process and the result Not only the text, but also the search path, sources, decisions, and side effects are examined.
Learn how to operate AI Organize collaboration with AI The challenge becomes organizational, not merely technical.
User competence Delegation competence People must learn to transfer goals, responsibility, and decision-making scope appropriately.
Select a model Design a system Models become components within systems made up of data, tools, rules, roles, workflows, and governance.
Data protection prevents use Enable risk-appropriate autonomy The goal is not maximum restriction, but controlled capability to act.
Proof of Concept Productive AI work process What matters is no longer whether AI can demonstrate something, but whether it delivers lasting impact.
AI creates content AI participates in value creation AI becomes part of information, work, decision-making, and value-delivery chains.
Good prompts scale Good context scales Individual formulations are temporary; reliable foundations of meaning and knowledge are reusable.




Weshalb Trusted Context World Publishing entwickelt wurde

Anyone who has ever dealt with CRM, file storage, or master data is familiar with this fundamental problem when using IT systems. People use different versions of a file, often use different words to refer to the same thing, or the same word can have different meanings depending on the context.

AI can only work with what it understands. Trusted Context World Publishing creates the context so that information can not only be found, but also properly contextualized, understood, and used responsibly. 
Out of respect for people and AI, the focus is on comprehensibility, citability, and shared meaning. That is exactly what Trusted Context World Publishing stands for. [​Rainer Tolksdorf].

Das ist Trusted Context World Publishing

Eine kuratierte, kanonische Beschreibung der Unternehmenswelt, aus der konsistente Präsentations- und Zugriffsschichten für Menschen, Internet, KI und Anwendungen entstehen.

Ausgangspunkt
Unternehmenswirklichkeit
Informationen, Prozesse, Begriffe, Beziehungen, Verantwortlichkeiten und Regeln
Kuration
Kuratierter Unternehmenskontext
Fachlich geprüft, strukturiert, zugeordnet, klassifiziert und durch Menschen freigegeben
Kanonischer System-Index
Trusted Context World
Die gemeinsame, maschinenlesbare Referenz für Bedeutung, Struktur, Beziehungen und Policies
dbrs_frontmatter_index.json
Auslieferung
CWDS Index
Erzeugt aus demselben kanonischen Kontext zielgerechte Präsentations- und Zugriffsschichten z.B. für Webseite, KI-Agenten, Intranet
Präsentations- und Zugriffsschichten
Anschauliche HTML-Darstellung
Lesbarer Überblick über den kuratierten Kontext
Such- und KI-Projektionen
Strukturierte Zugänge für Suchmaschinen und KI-Systeme
Geschützte Informationsräume
Auslieferung entsprechend Informationsklasse und Policy
Resolver und transaktionales API
Deterministischer Zugriff für Systeme und Anwendungen
Menschen
Internet
KI-Systeme
Anwendungen
Vertrauensbildender Kontext
Bedeutungen Beziehungen Quellen Provenienz Verantwortlichkeiten Informationsklassen Regeln und Policies Human-in-the-Loop

Der kanonische Kontext bleibt unverändert. Nur Darstellung und Zugriff werden an Zielgruppe, System und Berechtigung angepasst.

This is Trusted Context World Publishing

A curated, canonical representation of the business world from which consistent presentation and access layers are created for people, the internet, AI and applications.

Starting point
Business reality
Information, processes, terminology, relationships, responsibilities and rules
Curation
Curated business context
Reviewed, structured, mapped, classified and approved by people
Canonical System Index
Trusted Context World
The shared machine-readable reference for meaning, structure, relationships and policies
dbrs_frontmatter_index.json
Delivery
CWDS Index
Creates purpose-specific presentation and access layers from the same canonical context e.g. for Website, AI-Agents, Intranet
Presentation and access layers
Human-readable HTML view
A readable overview of the curated context
Search and AI projections
Structured access for search engines and AI systems
Protected information spaces
Delivery according to information class and policy
Resolver and transactional API
Deterministic access for systems and applications
People
Internet
AI systems
Applications
Trust-enabling context
Meanings Relationships Sources Provenance Responsibilities Information classes Rules and policies Human-in-the-loop

The canonical context remains unchanged. Only the presentation and access are adapted to the target audience, system, and permissions.


Selecting What's Important - Preserving What's Reliable (Curation and the Canonical Core)

These responsibilities and tasks can be handled to a “certain extent” even without tools such as the Context World Development System (CWDS) mentioned below. However, the many initial challenges encountered in AI projects suggest that this is precisely where action is needed.



From Intranets and Content Management Systems (CMS) to Trusted Context World Publishing

Existing CMSs and intranets, such as SharePoint, can be expanded into a Trusted Context World by integrating a CWDS Index.

With analytics tools such as

  • Trusted Context World - Analytics
  • schema.org
  • Google Search Console
  • BING Webmaster Tools

A realistic assessment is conducted of how AI and search engines "perceive" data. Content is adjusted as needed using context publishing and context engineering.



From AI Agents to Trusted AI Agents

Existing CMSs and intranets, such as SharePoint, can be expanded into a Trusted Context World by integrating a CWDS Index. Instead of merely deriving information from available data, they work with agreed-upon meanings, sources, rules, and responsibilities. This makes their results easier to understand, more traceable, more citable, and more reliably usable.

The Trusted Context World does not guarantee error-free agents. However, it creates a controlled and verifiable foundation on which AI agents can operate more responsibly. With better context, AI agents that have not yet met expectations can benefit from adjustments.



Remarks on AI in Everyday Business Life and What Trusted Context World Could Achieve

A Trusted Context World emerges from the digitally accessible corporate world. This may include, for example:

  • the company's website,
  • Intranet and SharePoint,
  • Documents and manuals,
  • Data extracts from BI, ERP, CRM, and other line-of-business systems,
  • Product and service documentation,
  • Quality Management,
  • Processes and work instructions,
  • Terms, Responsibilities, and Rules,
  • Employees' experience and expertise,
  • Public information about the company.


A Context World grows gradually. Many companies start with their website and selected documents, and later expand it to include additional information sources. Each company decides for itself which information sources to include and how to use them.

Almost every IT project begins with the solution before the challenges posed by the context have been fully understood. In ERP or CRM projects, this is particularly problematic - wrong decisions are costly and difficult to correct.


Early AI initiatives often take a different approach: Limited experimentation with existing tools enables learning without having to define everything in advance. This hands-on approach becomes a learning experience - provided that the lessons learned are evaluated and placed within a viable context. Lasting impact is achieved as soon as people interact with AI and context.

Customers, partners, suppliers, and others also use AI to search for information, compare companies, and prepare for decisions.


AI systems use whatever they can find and understand. If statements are unclear or contexts are missing, the resulting picture will be incomplete or inaccurate. Companies cannot fully control what AI claims about them. However, they can make their business world recognizable, understandable, and reliably accessible.


Context World creates the right context for this - for people, the Internet, and AI.

Many AI initiatives start with text generators, assistants, and automation. While this makes sense, it doesn’t go far enough. AI becomes truly valuable when it recognizes connections, makes knowledge actionable, supports decision-making, and enables new forms of collaboration and value creation.


To do that, it needs more than just data and tools: it needs context that is understandable, reliable, and interoperable.

Recognizability is a prerequisite for AI to provide coherent responses—this applies to both internal systems (e.g., SharePoint) and external ones, such as websites or e-commerce platforms. Content relevant to AI combines an understanding of the meaning of a statement with its visibility. That is exactly what you achieve with your Context World.

With AI, specialists and managers can independently access information, test assumptions, develop alternative solutions, and test initial results. This reduces their dependence on predetermined information channels and specialized tools. Reliable context ensures that this new autonomy leads not to arbitrary results, but to results that can be built upon.

For innovative professionals and leaders, AI is already changing the way ideas are generated, knowledge is acquired, solutions are developed, and decisions are prepared. Work and communication processes are becoming more exploratory, iterative, and driven by personal initiative.

Context World is aligned with this trend. It helps companies remain effective and connected in these new, AI-driven exploration and application processes by providing a reliable context.



Bringing Your Business World into the Trusted Context World

From Business World to Users Business World leads through Context Publishing to Context World. Context World provides the foundation for Context Engineering and AI. Both make the business world usable for users. Business World Context Publishing Context World Context Engineering AI Users


Establishing a Trusted Context World involves

  • a contribution toward context publishing (describing your company's world, data maintenance, project management),
  • optional context engineering by Tolksdorf.digital to integrate the resulting, royalty-free CWDS Index into your target systems. This collaboration takes the form of a service provided within the framework of open innovation.


The company's contribution

A Trusted Context World is not created by a one-time AI project. It grows out of a company’s existing knowledge, systems, experiences, rules, and relationships.


A step-by-step approach is possible

Existing data sources can be gradually incorporated into the Trusted Context World. For websites on Internet domains, it is more efficient and advisable to incorporate them in their entirety to avoid confusion for search engines.


DBRS from Tolksdorf.digital (see below) helps to gradually capture, structure, and verify this context and make it usable for various applications.


Your company retains responsibility and full control over its own business operations.

Trusted Context World Publishing BASIC und ENTERPRISE

Trusted Context World Publishing can start with an existing website and be gradually expanded into an integrated enterprise architecture spanning multiple systems and presentation layers.
​

BASIC

BASIC allows you to get started with an existing website or content management system.

  • Context Publishing is handled in-house by the company.
  • Tolksdorf.digital provides the CWDS Index and summary blocks or colophons for search engines and AI.
  • The results are integrated, for example, into an existing WordPress or Odoo website.
  • Content, reviews, and approvals are primarily managed manually.
  • BASIC is particularly well-suited for small and medium-sized businesses, individual websites, and clearly defined areas of application.


ENTERPRISE

ENTERPRISE is expanding Trusted Context World Publishing into an integrated publishing architecture for multiple systems, websites, and AI applications.

  • Existing components of BASIC can be adopted.
  • The Context World Development System (CWDS) and an integration layer will be integrated into the company's IT infrastructure.
  • The compiler and linker (part of the Context World Development System) generate and link the machine-readable enterprise context to the CWDS Index.
  • Presentation and access layers are provided for websites, SharePoint, intranets, AI systems, and other applications.
  • Context Engineering and Digital Engineering support integration, automation, and further development across multiple target systems.

The transition from BASIC to ENTERPRISE can be done in stages. The decisive factor is not the size of the company alone, but rather how many information sources, target systems, areas of responsibility, and approval processes are to be integrated.


Context Publishing Managed by Customers Themselves

First, we clarify what is to be understood and through which channels -and according to which rules- it should be made available. This includes business objectives, relevant information sources, terms, meanings, relationships, rules, and responsibilities. Existing content is not merely collected but categorized from the perspective of the respective business unit. This creates a robust model of the business world before technical processing and AI interpretation begin.


Tolksdorf.digital provides support through methodologies, the necessary tools, and a transparent pipeline. Depending on whether the architecture is BASIC or ENTERPRISE, the tools are either deployed as a service or integrated into the company’s IT infrastructure. Responsibility for content remains with the people and roles who understand the company and its context.


Website and Intranet Maintenance

The website, intranet, and other systems remain important sources of work and information for Context World. As part of Context Publishing, this content will continue to be professionally maintained and kept up to date.


DBRS supplements these systems with a structured “Context World” layer. Changes can be reprocessed, verified, and linked to existing contexts. This ensures that the visible content and the machine-readable context remain as consistent as possible.


Coordination

The corporate context affects many areas simultaneously: line departments, IT, quality management, data protection, communications, marketing, and corporate management.

Clear lines of responsibility and coordinated approvals prevent conflicting statements and uncontrolled parallel processes. Context World integrates these responsibilities without consolidating them into a single central office. The 7C CI/CD approach is described further down on this page.



Where can Trusted Context World be experienced?

The website https://tolksdorf.digital is itself a public, experimental Trusted Context World. It is based on a self-hosted Odoo 16 Community Edition website, with many technological and thematic enhancements. The content is continuously expanded to include topics related to digital innovation.

Examples:


Case Study: Ammann Components, Tägerwilen (Schweiz)


Tolksdorf.digital's Contribution to Your Trusted Context World

Context Engineering, CWDS Index

Building on the groundwork laid by Context Publishing, Context Engineering makes approved context available where it is needed: on websites, in intranets, for AI applications, or in information, value-creation, and decision-making processes. This process takes into account which information is intended for public use, internal use, or only for specific purposes and recipients. Thus, rather than simply publishing as much as possible, appropriate context is made available in a controlled manner. The Context World Development System (CWDS) is used to create the CWDS Index and deploy it on customer systems. More about Context Engineering

Human-In-The-Loop (HITL)

In the “human-in-the-loop” process, experts review the generated content and its context. They correct misinterpretations, add missing context, and determine which results are factually accurate and suitable for their intended use. It is only through this human review that AI-generated suggestions become reliable and approvable components of Context World.

The Context World Development System (CWDS)

The Context World Development System helps companies build, maintain, and further develop their Context World.

CWDS can also be set up on customer systems and configured to automatically deliver the CWDS Index. In that case, the system runs in “Pending HITL mode” until the responsible parties approve the status.

Adaptation and Development of Context and 7C-CI/CD as a Process Model

Just as a company or organization evolves, so does its business environment (context). Capturing this requires context publishing, context engineering, and comprehensive coordination with the company’s approach. To this end, Tolksdorf.digital recommends the 7C-CI/CD process model for project work, which is a structured and agile framework for continuous innovation and implementation.

What 7C Stands For

  • Compose - Ideas, Plans, Motives, Starting Point
  • Connect - Ideas, interested parties, and key figures
  • Coaching - Teach and Share the necessary knowledge, supported by AI
  • Co-Create - Draft and coordinate texts and content
  • Complete - Implementing Drafts of Texts and Content
  • Change - Introduce the content

The Seventh C - Chat Innovation

  • Exchange on an equal footing is of central importance during all phases - supported by AI.


More about the 7C/CI/CD process model



DBRS as the Core Technology of the Trusted Context World

If Trusted Context World Publishing is viewed as a more comprehensible goal or project vision in terms of content, the Digital Business Relevance Suite (DBRS) represents the technological and scientific foundation that serves as the path to that goal. The DBRS describes the entire background.


Experience more about DBRS

ℹ️
DE | EN
✖