Corpus operations

The corpus is the moat. Corpus operations keep it accurate, source-backed, and commercially usable.

Work queues

• Catalogue records.
• User suggestions.
• Crawl and enrichment results.
• Country showcase candidates.
• Brand corrections.
• Note image review.

Confidence tiers

• Curated: checked against strong sources.
• Brand supplied: provided by the house or distributor.
• Community: useful but needs review.
• Limited: partial data only.

Rules

• Keep aliases separate from canonical names.
• Keep flankers and concentrations distinct.
• Record source URLs.
• Never rely on blocked scraping paths.
• Show confidence to agents and API users.

Commercial outputs

• API search.
• Fuzzy matching.
• Shopify product mapping.
• Catalogue cleanup reports.
• Brand onboarding.
• Country packs.
• Trend reporting.