European Trustworthy AI Association

Body of Knowledge

The online, browsable version of the end-to-end W methodology — the reference framework for engineering critical systems that integrate artificial intelligence, from operational analysis through validation, qualification and deployment.

Scope

The BOK captures how systems, software and data engineering practices can be combined into a structured, coherent workflow to build AI-enhanced systems in a more trustworthy manner.

The methodology jointly articulates three engineering levels, unfolding as a "W" cycle:

  • System level — specification, architecture, integration, verification, validation and qualification of the system as a whole.
  • Model / software level — specification, development and evaluation of the AI model(s) and their complementary software items.
  • Data level — data lifecycle management, a concern spanning the whole engineering cycle, alongside the definition of the Operational Design Domain (ODD).
The W-cycle methodology diagram, with the System, Model/software and Data engineering levels highlighted

Today, browsing the full methodology graph — every activity's purpose, steps, responsible engineering role, linked tools and process diagram — is reserved to Association members. Log in to explore it.