Product intelligence
Product intelligence is the layer that turns raw catalog rows into decisions: what to fix, what to publish, what will fail on a channel, and what AI should work on next.
Lead with agent findings from the Agent Control Center; use completeness scores as publish gates — they measure different things. See Catalog quality — findings not fill rates.
Signals
| Signal | What it tells you |
|---|---|
| Agent insights | Open findings from specialist agents — SEO, taxonomy, compliance, syndication |
| Catalog Quality score | Rollup of open insights + pending approvals (ACC) |
| Approval queue | Agent-proposed writes awaiting human review |
| Completeness score | % of required attributes filled per family/channel — publish gate |
| Enrichment queue | SKUs below threshold needing AI or manual work |
| Export job items | Per-SKU syndication success/failure |
| Task backlog | Workflow approvals blocking go-live |
| Dashboard KPIs | Volume, activity, assignments (MCP: get_dashboard_stats) |
Intelligence in action
APIs and tools
| Operation | REST | MCP |
|---|---|---|
| Product completeness | Product API | get_product_completeness, analyze_product_completeness |
| Find gaps | Search / reports | find_products_needing_enrichment |
| Compare SKUs | — | compare_products |
| Recalculate | Product API | recalculate_product_completeness |
Why this beats spreadsheets
Spreadsheets do not know channel rules or family schemas. CataZenta intelligence is computed from the same model that powers Amazon, Shopify, and your agents — one source of truth, many outputs.