Enrichment

Asset enrichment uses AI to automatically generate asset structures and populate property values. Instead of manually filling in each field, you provide a name, a type, and a context note or documents, and VSEC builds a detailed, structured asset for you.

What is Enrichment?

When you create an asset using the Generate tab, the asset itself is created immediately. If you attached documents, VSEC also submits a generation request to your active Intelligence provider: the provider reads the documents plus your workspace’s live asset-type hierarchy, including all custom types and their full depth, then returns a suggested structure of child assets, new types, and the relations between them.

That request is asynchronous. The asset shows a “Generating…” status in your list while it runs; when it completes, the status changes to Ready to review and you review the proposed types, assets, and relations before anything else is created.

Generating from context only (no attached documents) skips this step entirely — VSEC creates the asset and takes you straight to its Context tab, where curation fills in properties over time (see Asset Context below).

See Creating Assets → Generate for the full workflow.

Intelligence Providers

Enrichment is powered by the Intelligence layer — a workspace-wide setting that routes AI requests to the active provider.

Deep Harbor is not available. Its row appears under Settings → Intelligence but cannot be enabled.

Configure your provider under Settings → Intelligence.

Asset Context

Each asset has a Context tab for building a living AI understanding of that asset. The tab is always shown; curation runs when Claude or Gemini is the active Intelligence provider, and until one is connected the tab shows a notice asking you to connect a provider.

The Context tab gathers a corpus from the asset’s context entries — freeform notes, system-added information, and submissions from outside contributors via a context request — together with its attached files (wiring diagrams, spec sheets, datasheets, and other supporting documents). When you curate the context, the active provider reads the corpus and returns:

• An understanding — a narrative AI summary shown read-only in the Context tab and regenerated on each curation pass
• Suggested properties — property values to add or update
• Suggested links — new asset relationships to create
• Suggested removals — existing asset links to consider detaching, each rated by confidence
• Suggested child assets — sub-components found in the corpus but not yet in the hierarchy, accepted via Create selected

You review every suggestion before anything is saved. Additions are pre-selected for review; removals are opt-in and require explicit confirmation before a link is detached.

See Asset Details → Context for the full tab walkthrough, or Asset Context Curation for the deep dive on corpus limits and ingestion status.