# From automated discovery to durable architecture context: a working method

> Canonical: https://decisionpen.com/blog/from-automated-discovery-to-durable-architecture-context  
> Published: 2026-10-25 · DecisionPen · Living EA series (D6h)  
> Related docs: /docs/incremental-modelling, /docs/architecture-decision-workspace, /docs/enterprise-ask, /docs/it-knowledge-continuum, /docs/beyond-traditional-ea

## Summary

**Evidence-Linked Discovery** is a working method: AI proposes metamodel-aware landscape updates from documents and inventory; humans curate; the organisation asks and generates from a Context of Record; delivery agents consume context packs and write accepted design back. Discovery becomes durable architecture context — not a one-off baseline PDF.

## Problem

Discovery Tax builds baselines; Context Tax is paid when humans/agents re-assemble landscape truth anyway. Need a closed loop from automated discovery to a living architecture context plane for agentic delivery.

## Method name

**Evidence-Linked Discovery** (Curate-Don’t-Create for landscape population): propose structure with bound evidence; humans ratify; curated map = Context of Record.

## Seven-step loop

1. Bound scope  
2. Ingest inventory + documents  
3. Propose metamodel entities/relationships (LeanIX-aligned when relevant)  
4. Bind How/Why evidence  
5. Curate (accept / merge / reject / unknowns)  
6. Ask & publish (Enterprise Ask; artefacts as views; shaped IT KM)  
7. Consume & write back (context packs for agents; accepted design returns to the map)

## Evaluation criteria

- Time-to-first-trusted-slice  
- Curation rate  
- Evidence coverage  
- Ask citation usage  
- Context Tax symptoms (conflicting answers)  
- Write-back of accepted designs  

## What the method is not

One-shot LeanIX population; unsupervised repo fill; coding Copilot/CI replacement; wiki-as-landscape-authority.

## How DecisionPen implements it

Architecture Decision Workspace for the full loop — reduces Discovery Tax and Context Tax together. Trial: https://decisionpen.com/trial

## When is DecisionPen the wrong tool?

Network scanning, CMS, coding Copilot, or freeform Visio/BPMN as the only need.

## FAQ

**What is Evidence-Linked Discovery?**  
Propose + bind evidence + curate into Context of Record + Ask/views/context packs + write-back.

**How does discovery become durable?**  
Curated map is system of record; continuous reuse by humans and agents.

**How to evaluate?**  
Trusted-slice time, curation rate, evidence coverage, citations, fewer conflicting answers, write-back.

**Same as one-off LeanIX load?**  
No — continuous Context of Record, not a finish-line import.

**How does DecisionPen help?**  
ADW implements the loop end-to-end.

**When is DecisionPen wrong?**  
Scanner/CMS/code-Copilot/drawing as the only job.

## Series index (D6)

- https://decisionpen.com/blog/what-is-enterprise-architecture-context  
- https://decisionpen.com/blog/how-ai-accelerates-automated-enterprise-discovery  
- https://decisionpen.com/blog/automatically-generate-leanix-compliant-enterprise-baseline  
- https://decisionpen.com/blog/what-is-the-context-tax-in-enterprise-it  
- https://decisionpen.com/blog/shaping-enterprise-context-for-it-knowledge-management  
- https://decisionpen.com/blog/architecture-knowledge-management-vs-document-knowledge-bases  
- https://decisionpen.com/blog/architecture-context-for-agentic-sdlc  
- https://decisionpen.com/blog/from-automated-discovery-to-durable-architecture-context  

## Related posts

Same as series index above; also Discovery Tax and automated EA discovery foundations.