Validation: Testing the Hypotheses in Practice

Working Question

No theory of design systems for the age of AI can emerge from reasoning alone. Its true value can only be assessed through long-term use in a real product.

These documents should therefore not be understood as a collection of finished conclusions, but rather as a research framework. They formulate a set of working hypotheses that will be continuously tested throughout the design, implementation, governance, and everyday use of a newly developed design system.

Main Working Thesis

The purpose of this work is to establish a framework for systematically testing—and, where necessary, disproving—the proposed theories.

Objectives of the Practical Validation

To determine whether the proposed principles actually lead to:

  • a more understandable design system,
  • better collaboration across different roles,
  • more effective collaboration with AI,
  • fewer ambiguities,
  • simpler governance and evolution of the design system,
  • greater long-term product consistency.

Scope of the Validation

The practical implementation will focus primarily on the following areas:

  • knowledge architecture,
  • the knowledge node model,
  • the representation of design intent,
  • multi-level documentation,
  • AI working with the knowledge base,
  • design system governance,
  • everyday use across different roles.

Method

Each thesis will be tested during the real-world design, implementation, operation, and evolution of the design system.

I expect the proposed model to evolve continuously as new experience is gained. Some hypotheses may be confirmed, others refined, and some abandoned altogether.

Success Criteria

The hypotheses will not be considered validated merely because the proposed system works.

More important will be the answers to questions such as:

  • Does the proposed architecture help people in their daily work?
  • Does it help AI interpret knowledge more accurately?
  • Is it sustainable over the long term?
  • Does it reduce duplication, inconsistencies, and ambiguity?
  • Can it be extended without fundamental changes to its information model?

Future Research

The practical implementation represents only the first experimental validation of this work.

If the proposed principles prove successful, they may provide a foundation for further research into the architecture of design systems, the representation of design knowledge, and collaboration between humans and autonomous systems in the design of digital products.