DecisionPen

Architecture Decision Workspace

Architecture decisions at the speed of AI

DecisionPen blends structured enterprise architecture entities with the documents, decisions and expert knowledge that explain them — not a chatbot bolted onto a wiki, and not another empty EA repository waiting for months of factsheet entry. Build a living IT landscape from a CMDB, your docs, or both: AI proposes, humans curate, then ask evidence-backed questions and generate architecture artefacts from the accepted model.

See it live Watch 90 seconds

Start from a CMDB or inventory spreadsheet, existing documents, or a blank domain · No credit card required

Watch 90 seconds

Think ChatGPT or Gemini — but for your enterprise architecture. See how DecisionPen works as an Architecture Decision Workspace.

Ready to try it yourself?

See it live

Why the old model fails

Classic EA tools die empty, then stale. The Discovery Tax — months of workshops and factsheet entry — leaves repositories hollow. The week after the model is “done,” reality drifts. Architects become data clerks; the tool becomes a historical snapshot.

Manual documentation cannot keep up. HLDs, wikis, and chat still hold the IT DNA, but unstructured knowledge cannot answer blast-radius, ownership, or due-diligence questions at AI-coding pace. Writing more docs is not a decision system.

The missing blend

IT knowledge management must unite structured EA entities with the underlying documentation context. Structure without evidence is brittle. Documents without structure are noise.

Structured entities

Applications, capabilities, technology, integrations and initiatives — including a catalogue imported from a CMDB or inventory spreadsheet — with metamodel-aware relationships and LeanIX-aligned modelling where that standard matters.

Document context

Notes, PDFs, HLDs, and transcripts that evidence and enrich the map — citations you can trust when you decide.

DecisionPen is the blend: AI proposes; humans curate; the organisation interrogates a living map.

No blank repository

Start from the knowledge you already trust

Skeptical of building architecture only from unstructured text? So are we. DecisionPen gives you two governed on-ramps — and most enterprises use both.

Path 1 · Structured-first

Seed from your CMDB or inventory

Export the catalogue you already maintain and import it from a spreadsheet to establish canonical applications and technology components. Then enrich those thin factsheets with purpose, integrations, decisions, constraints and evidence from your documents and experts.

How factsheet enrichment works

Path 2 · Document-first

Discover from institutional memory

Where the inventory is missing or untrusted, start with HLDs, PDFs, wiki pages, diagrams, transcripts and natural-language descriptions. AI proposes entities and relationships; architects curate them into the same living model.

Automated landscape modelling

Both paths converge: structured identity plus How-and-Why evidence, reviewed by humans. A spreadsheet extract from the catalogue you already maintain is enough to seed the model today.

How it works

From what you already know to decisions you can defend — in one shared workspace.

1
Start from what you have
Import a CMDB or inventory spreadsheet — or bring HLDs, PDFs, transcripts and dialogue.
2
Review
AI proposes entities and relations. You approve, link, or skip — steward of record.
3
Map
A living, shared landscape of applications, capabilities, tech, and flows.
4
Ask
Interrogate in plain English — neighbours, ownership, blast radius, gaps.
5
Decide
Blueprints, ADRs, and diagrams as byproducts of truth — not documentation theatre.

For anyone who makes important technology decisions

Not an architect-only drawing board — a shared workspace for the agentic era.

CTO / CIO
Change velocity without flying blind
Enterprise architect
Living landscape, not a failed inventory project
Solution architect
Design and ADRs from shared truth
Engineering leads
Due diligence before AI-generated change ships
Transformation / M&A
Dependencies and carve-outs from messy reality
Wider organisation
Grounded answers and faster onboarding

Where teams start

The same workspace — three common jobs.

Rigorous enough for architecture. Fast enough for delivery.

Human-in-the-loop — nothing joins the live map until you say so.
Evidence-linked — answers grounded in the map and source documents.
Metamodel-aware — relationships that hold; LeanIX-aligned modelling when you need that standard.

Why DecisionPen

Try the Architecture Decision Workspace

See it live in your own workspace — start with the banking example, then bring a system of interest of your own.