Dokan's Store Support module stores each ticket as a custom post type called dokan_store_support with vendor and order context attached through wp_postmeta keys like store_id, order_id, and status. Customers open tickets from the My Account area or the product page, and vendors reply from their own dashboard. The problem is that buyers usually do not know which vendor owns which product, and they ask the marketplace bot first.

SleekAI maps Store Support post fields, the vendor's shop_name from wp_users and the Dokan vendor metadata, plus the related WooCommerce order from wc_orders into the system message. Display conditions scope the bot to the My Account ticket page, a specific vendor's store URL, or a logged-in customer with an open ticket. Multibot lets a customer-facing assistant and a vendor-side triage assistant share the same install with different prompts.

A generic chatbot has no idea who the vendor of a given product is, will invent refund policies that the vendor has not agreed to, and cannot tell a customer that ticket number 482 is already marked as answered. SleekAI reads the actual ticket thread, status, and last reply timestamp, then either confirms the status or asks for a specific clarification the vendor will need.

Marketplaces fail when buyers cannot tell who is responsible for their order. A generic chatbot does not know that Dokan splits the relationship between marketplace operator, vendor, and customer. It will accept a refund question that only the vendor can decide, repeat a policy that does not exist on this store, or quote a shipping window that the vendor never agreed to.

Each of those answers becomes a chargeback, a public review, or a vendor leaving the platform. Dokan already stores everything the assistant needs to behave correctly. The Store Support custom post type, the vendor user record, and the underlying WooCommerce order are all in the WordPress database.

SleekAI exposes them as mapped variables. Marketplace operators can run a triage bot that knows whose ticket is whose, vendors can run a coaching bot scoped to their own dashboard, and customers stop bouncing between vendor and platform looking for an answer. The result is faster first response, fewer escalations, and a clearer signal back to the operator about which vendors need help.