AgenticCart
Docs Product Collections

Collections

Collections decide which products an AI sales agent can recommend. Use them to keep recommendations focused, on-brand, and intentional for every buying journey.

An AI agent product collection is the curated set of products an AgenticCart agent is allowed to recommend. Collections are the merchandising layer between your full catalog and your AI shopping assistant — they decide which products the agent can suggest, which stay hidden, and how a shopper's buying journey is scoped.

Why AI agent product collections matter

Your full catalog is rarely the right recommendation pool. A supplements brand does not want its gift-buying assistant suggesting bulk whey; a clothing store does not want its denim specialist surfacing swimwear in January. Collections are how AgenticCart keeps AI shopping recommendations focused, on-brand, and intentional.

  • They scope each agent to one buying journey — gifting, a category advisor, comparisons, or a campaign — which is what gives shoppers a clear experience, whether the underlying collection holds fifty products or five thousand.
  • They protect merchandising strategy by excluding products you do not want promoted right now.
  • They make each agent commercially intentional — a gift-finder agent should recommend gifts, not bulk protein tubs, even if both live in the same catalog.
  • They make testing easier because you know exactly which curated product sets the agent can consider when you review conversations.

What a good first collection looks like

AgenticCart is built for catalogs of any size — from small boutiques to stores with thousands of SKUs. What matters for a first collection is not how many products it holds but how clearly it maps to one real buying journey. Four patterns that work across verticals:

  • Bestsellers — your strongest SKUs, whatever that count is for your store. Works because these products already have strong descriptions, reviews, and availability, so the agent has plenty of signal to explain and compare.
  • Gift guide — products grouped by recipient, occasion, or budget. A gifts store can launch a "Under 100 for friends who travel" collection and let the agent do the recipient-to-product translation.
  • Category advisor — the complete set for one product category. A clothing store's "Denim" collection or a supplements brand's "Plant-based protein" collection gives the agent a clear area of expertise, however many SKUs sit under that category.
  • Campaign set — products tied to one launch, ad campaign, or seasonal offer. Perfect for a homeware brand running a spring sale: the agent stays glued to the promoted products and does not wander into evergreen SKUs.

TIP

A collection should match a buying journey, not a SKU cap. AgenticCart handles large catalogs — the question to answer is "which products belong to this conversation?", not "how small can I make this list?"

How to structure collections for different buying journeys

Different shopper intents deserve different collections, even if the underlying catalog is the same. A single clothing store can run three separate AI shopping collections at once: a denim advisor, a workwear collection, and a weekly sale campaign set. Each one becomes the recommendation pool for a different AI sales agent.

  • By customer segment — gifts versus self-purchase, new customers versus returning, budget-conscious versus premium.
  • By page context — one collection for your main hosted chat domain, another for a campaign hosted chat page, another for a seasonal gift-finder domain.
  • By seasonality — rotate campaign collections as offers change without touching your core bestsellers collection.

When to update a collection

  • A campaign starts or ends, and the featured product set changes.
  • New products launch and should be promoted to shoppers immediately.
  • Inventory or availability changes materially — see removing unavailable products.
  • Recommendation quality feels too broad in real conversations.
  • You want to test a different buying journey with the same catalog.

How collections relate to your catalog

Your catalog is the full product data AgenticCart pulls from your WooCommerce store or product feed. A collection is a subset of that catalog chosen for one specific agent. The catalog changes as your storefront changes — prices, stock, new products — via catalog sync. Collections only change when you edit them.

This split matters: sync keeps data fresh, and collections keep merchandising intentional. When a product is removed from your store it drops out of your catalog automatically, and any collection that referenced it will show it as unavailable until you remove it from the collection view.

Frequently asked questions

How many products should a collection contain?
As many as genuinely belong to that buying journey. AgenticCart works with catalogs of any size, including stores with thousands of SKUs — a "Running shoes" collection should contain all your running shoes, and a "Under $50 gifts" collection should contain everything that qualifies. The rule is relevance to the shopping intent, not a SKU cap.
Can a product belong to more than one collection?
Yes. A single SKU can sit in your bestsellers collection, a seasonal gift guide, and a category-specific AI shopping collection at the same time. Use overlap freely — collections are lightweight merchandising filters, not physical inventory.
Do collections affect my site's actual storefront?
No. AgenticCart collections exist inside AgenticCart only. They control what your AI sales agent can recommend inside the hosted chat page. Your WooCommerce categories, storefront pages, and checkout are untouched.
What happens if I delete a collection that's in use?
If a live agent is assigned to the collection, you will be asked to reassign the agent to another collection first. This prevents an agent from going dark mid-conversation because its recommendation pool disappeared.

Next steps