Question 1

What are agents in SKUuz?

Accepted Answer

Agents are autonomous AI-powered automations that perform specific catalog operations without manual intervention. Each agent has a type (content, translation, pricing, validation, etc.), a trigger condition (on import, on schedule, on manual run), and an action configuration (which fields to generate, which pricing rules to apply, etc.). When triggered, agents process products through a queue and record their actions in the activity log. You configure agents once, then they run continuously as your catalog evolves. A cron-based dispatcher runs agents automatically, and a stuck-processing auto-reset ensures the pipeline never stalls.

Question 2

What agent types are available?

Accepted Answer

SKUuz deploys ten agent types: Content (generate descriptions, SEO titles, feature bullets using AI Field Prompts and BrandGuard rules), Translation (translate fields to target locales using glossary-aware AI, with batch processing for efficiency), Categorization (auto-assign categories from your hierarchy using AI, with configurable confidence thresholds, auto-create mode, and re-categorization support), Validation (enforce data quality via configurable rules including AI-powered semantic validation), Schema (detect unmapped attributes after import and suggest data model additions, with AI-powered near-duplicate attribute detection), Pricing (apply markup rules and psychological rounding from cost prices), Image (generate alt text and trigger Image Studio operations), Publish (run publish rules, assign to channel nodes, and trigger push to Shopify), Syndication (map and transform product data for external marketplaces and feeds), and Compliance (verify product data meets regulatory and channel-specific requirements). All agent types are available on every paid plan.

Question 3

Do agent runs consume credits?

Accepted Answer

Agent execution itself is free. Credits are consumed only when agents perform AI operations — generating a description (5 credits), translating a field (2–3 credits), categorizing a product (2 credits), or generating alt text (2 credits). Validation agents checking rules without AI consume zero credits. Pricing agents applying formula-based rules consume zero credits. This means you can run validation and pricing agents freely without worrying about credit burn.

Question 4

What happens when I run out of credits?

Accepted Answer

Agents that require AI operations pause until credits renew at the start of your next billing cycle. Agents that don't consume credits (validation, pricing with formulaic rules) continue running normally. You can purchase a Launch Pack at any time to resume immediately, or upgrade your plan for more included monthly credits.

Question 5

Can I control whether agents auto-apply changes or require approval?

Accepted Answer

Yes. Each agent has a run mode: 'auto-approve' applies changes immediately, 'queue for review' places changes in a review queue for manual approval. You can configure this per agent, so your Content Agent might auto-approve while your Pricing Agent requires review. When queue-for-review is active, the agent's proposed changes appear in the Review Queue and on the Command Center as pending items.

Question 6

How does the Categorization Agent work?

Accepted Answer

The Categorization Agent v2 auto-assigns products to categories from your hierarchy using AI analysis of product attributes (title, vendor, product_type, and configurable additional attributes). It includes several advanced features: configurable confidence thresholds (low, medium, high) to control how aggressive the matching is; a hallucination catch that rejects AI-suggested category names not in your actual hierarchy; and an escalation system for products that don't match any category. With auto_create_categories enabled, the agent can create new categories when no match is found (in 'auto' mode it creates immediately, in 'suggest' mode it creates an attention item for your review). You can also trigger re-categorization of specific products, run a confidence audit to find low-confidence assignments, and merge categories — all from the Command Center Agent.

Question 7

What is the Command Center Agent?

Accepted Answer

The Command Center Agent is an AI-powered chat assistant accessible via the green icon in the bottom-right corner of the PIM. It has full context of your catalog — products, categories, attributes, agents, locales, credits, publish rules, AI content config, and BrandGuard terms. It can execute 24 different actions including: managing attributes, categories, and list values; configuring publish rules and diagnosing publish issues; enabling and configuring AI content generation; managing BrandGuard terms; re-categorizing products and running confidence audits; merging categories; updating categorization settings; searching products by attribute filters or by ID; and importing files directly via chat upload. Search actions execute immediately (they're read-only), while data-modifying actions require confirmation first.

Agents