Why DecisionPen
Architecture knowledge that matches the pace of change
AI coding and agentic systems multiply how fast organisations can change. Decision systems that still run on empty repositories and stale documents cannot keep up.
Why classic EA tools die
Enterprise architecture initiatives stall on the Discovery Tax: empty repositories, months of workshops and factsheet entry, then drift the moment the baseline is declared done. Architects become data clerks. The tool becomes a historical snapshot — not a living guide for decisions.
In an agentic world, a quarterly inventory is already obsolete. Portfolio systems of record still matter for governance programmes; they are the wrong primary frame for day-to-day decision velocity.
Why manual documentation is no longer enough
HLDs, Confluence pages, decks, and chat threads still hold the real IT DNA. But unstructured knowledge cannot support consistent blast-radius questions, ownership clarity, or due diligence before AI-assisted change ships. Writing more documents does not create a shared decision system.
The missing blend
IT knowledge management must unite structured EA entities — applications, capabilities, technology, integrations, initiatives, and their relationships — with the underlying documentation context that evidences them.
Structure without evidence is brittle. Documents without structure are unsearchable noise. DecisionPen is the blend: AI proposes map updates from reality; humans curate; the organisation interrogates a living map grounded in citations. Modelling is metamodel-aware and can align to LeanIX-style structures when that standard matters — as rigor, not as identity.
A workspace for anyone who decides
This is not an architect-only drawing board. It is an Architecture Decision Workspace for CTOs, CIOs, chief architects, solution architects, engineering leads, programme leaders — anyone who must decide under change pressure.
Workflow: Add knowledge → Review → Map → Ask → Artifacts. Decisions keep pace with AI coding because the knowledge layer finally does.
Not another EA suite. Not a chatbot. Not a wiki.
- vs EA suites — you own the decision loop, not the full catalogue feature war.
- vs generic GenAI — metamodel-aware and map-grounded, with citations.
- vs wikis — structured, relationship-aware truth with human-in-the-loop control.