AI-native Systems Engineering Learning Method
A knowledge-first, AI-partnered methodology I developed from 20+ years of experience in enterprise architecture and systems engineering.
ANSELM is a methodology for conceiving, reasoning about, and evolving complex systems. It treats the system's knowledge ecosystem—expressed primarily in natural language and structured data—as the primary model.
AI serves as a co-pilot for synthesis, analysis, and coherence maintenance. Formal diagrams and documents become disposable, on-demand views—not the foundation.
The paradigm: Complexity seeking clarity.
The true model is the living, interconnected ecosystem of knowledge about the system—its intents, constraints, decisions, and behaviors. Diagrams are transient views into this ecosystem, not its foundation.
The purpose of AI is to engage in the intellectual work of systems engineering: to synthesize, to challenge, to reason over trade-offs, and to maintain coherence. It is a collaborative intelligence, not a drawing accelerator.
The primary medium of engineering must be the medium of thought: natural language, captured as text. Formalisms should be derived from this understanding, not imposed upon it.
Reducing cognitive load requires making relationships and contradictions explicit and traceable—a task for which AI is uniquely suited. Syntactic compliance with a graphical standard does not equal conceptual integrity.
Traceability cannot be a forensic exercise; it must be the natural byproduct of a connected reasoning process. The rationale for every decision must be as accessible as the decision itself.
Begin with unstructured and semi-structured knowledge. Structure emerges through collaboration with AI, not as a prerequisite.
The systems engineering cycle is a series of structured dialogues—between stakeholders, between disciplines, and between the engineer and their AI co-pilot.
Diagrams, reports, and documents are generated on-demand from the underlying knowledge graph. They are consumable, disposable artifacts, not source artifacts.
Consistency, constraint satisfaction, and compliance are assessed continuously by AI across the growing knowledge base, not in batch-process review gates.
The method thrives on interoperable, human-readable formats (Markdown, YAML, plain text) and avoids proprietary data prisons. The intelligence is in the process, not the file format.
The future of systems engineering is not more sophisticated notation.
It is amplified reasoning.
Visit anselm.ingA declaration of knowledge-first, AI-partnered systems engineering principles.
Why we need a knowledge-first approach to systems engineering.
Building systems with AI as your co-pilot — a practical guide.
Scaling AI-native systems engineering across teams and organizations.