Yes. SleekAI reads your WordPress and WooCommerce data live, so picks stay inside your real cellar. The bot can mention adjacent wines for context (the Crozes-Hermitage that the recommended Cornas drinks like) but the actual recommendation - the bottle the shopper should buy - stays in stock. Out-of-stock or sold-out variations get filtered before the prompt is built.

Yes. Drop your buyer's tasting notes and shop newsletter copy into the OpenAI Files vector store and the bot quotes them naturally. The shopper hears your house voice - the buyer's specific take on a 2020 Janasse - rather than generic 'notes of dark fruit and oak' language. Tasting note voice is part of the shop's identity, and the bot inherits it.

Yes. Multibot plus display conditions lets each section have its own bot and prompt. The natural bot knows your low-intervention biases, your producers (Shobbrook, Foradori, Frick), and your tonal register. The fine bot leans classical, knows your en primeur policy, and references your fine-wine consultant for shoppers spending over a threshold.

Yes. As long as those live in custom fields or post content, they make it into the prompt. A shopper asking 'do you have any 2019 Burgundy under 50 EUR' gets a real answer because vintage is a structured field. Sub-regional detail (Cote de Beaune Villages vs Bourgogne Rouge) also comes through if the catalog tracks appellation.

Conversation logs flag interest signals you can email or list-add later. When the natural-wine truck arrives Thursday, the bot can ask 'want me to flag your email when they hit the shelf?' and the buyer's morning starts with a list of named allocations rather than a first-come-first-served drop. Small allocations especially benefit - the buyer routes the 12 bottles to the customers who asked first.

Yours. SleekAI is bring-your-own-key for OpenAI. For wine shops with thousands of monthly conversations, GPT-4o-mini handles most pairing questions cheaply, and you can reserve GPT-4o for nuanced producer-history questions or by-the-case advice where the chemistry of the conversation matters more than the recommendation count.

Yes. If your site publishes case discounts (10% off mixed cases, 15% off solid cases, trade pricing for restaurants), the bot quotes them and routes the conversation to the right path. A restaurant buyer asking about wholesale gets a different reply than a retail customer - the bot recognizes the signal and pivots without the customer having to find a separate trade form.

If your event pages live in WordPress with date, ticket price, and theme fields, the bot pulls from them. A shopper asking 'are you doing any natural-wine tastings this month' gets a real answer with date, price, and signup link - rather than a 'check our events page' deflection that loses the booking momentum.