Yes. WooCommerce stores attributes in wp_postmeta, and SleekAI's data-source wizard maps any attribute you choose into the chatbot's context. Metal purity (14k, 18k, platinum), carat weight, certification number, and any custom field you have configured become things the bot can quote precisely. There is no separate sync; the bot reads what WooCommerce currently has every time a shopper opens a conversation.

If your products use size variations, the bot reads stock per size at request time and answers "size 7 is in stock, size 8 is made to order with a 10-day lead time" rather than "please contact us for sizing." For shoppers who need help measuring, you can include your sizing guide in the context so the bot walks them through the paper-strip method or refers them to a free ring sizer order. The widget can be scoped to engagement and wedding categories where sizing matters most.

Yes, as long as the distinction is captured somewhere in product data, either as an attribute ("origin: lab-grown") or in the long description. The bot can compare two specific SKUs on price, carat, and certification, which is the conversation that most engagement-ring shoppers want to have before booking a video appointment. Keep the comparison grounded in the policy and product copy so the bot does not editorialize on ethics or sourcing beyond what you have published.

Drop your engraving and resizing policy pages into the chatbot's context and the bot quotes lead times by metal type, including any rush options and their fees. Logs help you see which lead times are blocking conversions, which is useful around holidays and Valentine's Day. The bot can also remind shoppers that personalized items are typically non-returnable, which reduces dispute volume after the fact.

Add your appraisal sample, insurance partner FAQ, and warranty PDF to an OpenAI Files vector store. When a shopper asks "does this come with an appraisal for insurance," the bot quotes the relevant paragraph and links to the policy page. For higher-value pieces, this single answer often closes the gap between an interested browser and a paying customer who is comfortable buying online.

Yes. Configure the system prompt to lean on category and tag relationships from the catalog. A shopper asking for "a pair of stud earrings under 400 EUR in white gold" gets options drawn from real products that match those filters. The bot avoids aggressive cross-sell and stays factual: it surfaces options rather than pushing a single SKU, which is the tone most jewelry shoppers respond to.

In your WordPress database. Every message is logged with model name, token usage, and page URL, so you can audit the bot's answers and improve the system prompt over time. Nothing routes through Sleek-owned infrastructure: the data path is browser to WordPress to the model provider on your API key. For shops with GDPR commitments or AML-related data handling, this keeps the data-processor map simple.

Multibot lets you run several bots on one site with different system prompts, display rules, and even different model providers. An engagement-ring bot can speak slowly and ask qualifying questions, while a gift-finder bot can be brisker and focus on price and shipping. Display conditions cover post type, taxonomy term, user role, logged-in state, and URL pattern, so each bot only appears in its intended section of the catalog.