What is automated enterprise architecture discovery?

What do people mean by automated EA discovery?

Usually: “stop typing factsheets — populate the landscape from what we already have.” Search phrases sound like auto populate EA repository from documents, automated IT landscape mapping, EA discovery vs CMDB. The phrase is overloaded. Some vendors mean a scrape. Some mean chat over PDFs. Useful discovery means something stricter.

What automated EA discovery is (and is not)

It is: software that proposes metamodel-aware entities and relationships from documents, meetings, transcripts, and related signals — then humans curate those proposals into a living map you can ask and act on.

It is not: a one-shot CMDB dump that lists assets without How/Why; or plain document RAG that retrieves passages without governed structure (why not Copilot or RAG alone).

Comparison of CMDB scrape, document RAG dump, and metamodel-aware discovery with curation — Useful discovery proposes structure you can curate — it does not pretend a scrape or chat over PDFs is a landscape.

Why discovery has to beat the Discovery Tax

Manual baselines are why architects hate modelling and why repositories stay empty — see why architects hate manual EA modelling. Automation that still produces thin factsheets without evidence just accelerates hollowness (factsheets vs How and Why).

How does DecisionPen solve this?

DecisionPen implements discovery as incremental modelling — continuous synthesis, not a mythical complete scan.

Ingest — HLDs, PDFs, wikis, transcripts, natural-language domain descriptions.
Propose — metamodel-compliant applications, capabilities, technologies, integrations, and relationships (LeanIX-aligned where that is your language).
Curate — architects accept, edit, reject, or defer; trusted updates beat “100% inventory.”
Bind evidence — How/Why passages stay attached so the map stays explainable.
Ask & act — Enterprise Ask and generated artefacts use the curated result.

Trial DecisionPen on one domain’s documents — prove a useful discovered slice in days, not a year of data entry.

When is DecisionPen not the right tool?

If you only need freeform diagramming/BPMN (Visio, Lucidchart) or a pure ops CMDB. DecisionPen is for curated living architecture context and generated views — not a drawing canvas or asset database.

What to read next

Frequently asked questions

What is automated enterprise architecture discovery?

Using software to propose metamodel-aware entities and relationships from documents, meetings, transcripts, and related signals — then having humans curate those proposals into a living, askable map.

Is automated EA discovery the same as a CMDB import?

No. A CMDB import can list assets; it rarely captures How/Why design context or strategic capability framing. Discovery for decisions needs structure plus evidence, not only inventory rows.

How does DecisionPen automate EA discovery?

DecisionPen ingests documents, meetings, and natural-language descriptions; proposes metamodel-compliant applications, capabilities, technologies, and links; binds evidence; and lets experts curate. Enterprise Ask then uses the curated map — incremental modelling, not a one-shot dump.

When is DecisionPen the wrong tool?

If you only need freeform diagramming or BPMN drawing (Visio, Lucidchart), or a pure CMDB/ops asset database. DecisionPen is for curated living architecture context and generated views.