Yes. Size charts can be stored as product meta, ACF fields, or as referenced documents in the OpenAI Files vector store. A shopper asking "what's my size in the Falcon cleat" gets the brand's actual chart, not a generic table. For stores carrying many brands, this is the difference between a chatbot that helps and one that confuses. The bot can also flag known fit notes, like "runs half a size small per brand guidance," if you publish that on the product page.

If your products use WooCommerce variations for size and color, the bot reads stock per variation at request time. A shopper asking "do you have men's 10.5 in black/red" sees the answer for exactly that variation, with a restock ETA if you publish one. There is no separate stock feed to maintain: WooCommerce is the source of truth and SleekAI reads it directly from postmeta when a question arrives.

Yes, if surface compatibility is stored as an attribute or in product copy. The bot can point a shopper away from FG cleats if they mentioned artificial turf and surface the AG or TF model instead. The same pattern works for snowboards by terrain, climbing shoes by route style, and tennis shoes by court surface. The catalog has to be tagged with surface or condition data; once it is, the bot has the information it needs to make accurate recommendations.

Yes. If you publish restock ETAs or seasonal cutoffs ("ski season collection ships through April 15"), the bot quotes them. For pre-orders, the bot can describe the expected ship window honestly, including the policy that you charge at shipment rather than at order. Setting clear expectations is the single biggest factor in reducing pre-order cancellation, and the chatbot is the most consistent place to set those expectations.

If you publish bundle or team pricing rules in product copy or a dedicated page, the bot quotes them. For a club uniform order, the bot can walk the buyer through team-pack discounts and minimum quantities, then route them to a contact form for the final order if your bundle plugin does not support it natively. For shops on WooCommerce Bundles or Composite Products, the bot can quote bundle prices from the actual configuration.

Sizing is one of the areas where catalog grounding matters most. The bot is given the brand-specific chart relevant to the product the shopper is on, which means it does not have to remember every brand's quirks from training data. It quotes the chart, references the fit note if one exists, and refrains from making up specifics. Logs from these conversations are useful to fine-tune the system prompt and to spot products that need a clearer size chart on the listing itself.

In your WordPress database. Conversations are logged with model name, token usage, and page URL, so you can audit answers, evaluate model choice, and improve the system prompt over time. Nothing routes through Sleek infrastructure: the data path is browser to WordPress to your model provider on your API key. For shops with strict data-residency commitments, this keeps the data-processor map predictable.

Multibot lets you run several bots with separate system prompts and display rules. An outdoor-gear bot can be detailed about temperature ratings and waterproofing; a team-sports bot can be brisker and focus on stock and shipping for tournament timing; an apparel bot can lean into fit and care. Display conditions scope each bot to the right section. For a single-store frontend, this gives you the precision of multiple specialist assistants without juggling multiple chatbot vendors.