Answers:

• Automated curation trade-off: the upside is low setup; the downside is less explicit control. Genie ranks and conflict-resolves snippets automatically, so if your sources are inconsistent or ambiguous, you still need to review and clean up guidance. Internal positioning also says it is not a full formal/queryable ontology today.

• Export / snapshot: No public export/snapshot feature yet, access is through Genie / Genie MCP.

• Can you inspect provenance? Partially yes. You can inspect which knowledge sources were used for an answer via citations in Genie responses, including Genie Ontology citations in the MCP app. No dedicated lineage view showing every source that contributed to a specific ontology relationship/definition.

• How much prep is needed? Genie Ontology can start automatically from tables, queries, dashboards, docs, and connected apps—so metric views are not required just to get started. But for good results, you need a governed semantic foundation: metric views / measures / joins / business definitions / synonyms / knowledge store.

• Recommendation: use auto-curation as the base layer, but add curated metric definitions, joins, and business terms anywhere accuracy really matters.