Yes. Stock per flavour lives in postmeta and SleekAI reads it on every turn. A 7 pm visitor gets a real count, including warnings when a flavour is down to a single tub. When a flavour sells out the next reply reflects it and suggests the closest in-stock alternative from the same dairy or sorbet family.

Map dairy-free, vegan, and nut-free flags per product. The bot only confirms a flavour is vegan when the data says so. For ambiguous mixes it points the visitor to the allergen filter page or to a counter conversation, rather than guessing and risking a customer with a serious allergy.

Yes. The cake form usually stores size, flavours, swirl, message, and lead time in postmeta. SleekAI reads those rows and can quote a 9 inch round at $52 with two flavours and a 72 hour lead time. It refuses orders that fall inside the lead time and suggests the next eligible pickup date.

Yes. Sundaes are usually variable products with their own variation pricing for 2, 3, and 4 scoops, and add-ons like extra toppings, whipped cream, and waffle cones. The bot reads the right variation row directly and quotes a 3 scoop hot fudge at $14 with each add-on priced from the postmeta.

Map a rotation taxonomy where each flavour is tagged with the days it runs. The bot knows the lavender honey is Wednesday and Thursday only and the dairy-free pistachio is a weekend flavour, so a Monday request gets a polite heads up and an alternative from today's lineup.

Yes. Display conditions support post type, product category, tag, user role, and URL pattern. A menu bot focuses on scoop stock and sundae quotes, while a cake bot specialises in form fields and lead times. Both read the same WooCommerce data without any duplication of effort.

Pints and quarts are usually separate WooCommerce products. Map them and the bot can quote a hand-packed pint at $9.50 or a quart at $16 for any in-stock flavour. It can also confirm whether dry ice packing is needed for a long drive home in summer.

Yes. Every conversation is stored with the model name, token usage, and the page URL it ran on. Patterns like repeated questions about a missing dairy-free flag or an unclear cake lead time feed straight into the product fields and form copy, so the shop's online voice keeps closing the gap with the counter.