JetSmartFilters stores each filter as a custom post type entry, typically jet-smart-filters, with postmeta for the filter type, the linked query variable, and the data source (taxonomy, post meta, JetEngine field, or a custom callback). The plugin does not store submissions in a table. Filters are applied at query time, with the active state encoded in the URL as query parameters that the AJAX provider reads on every change.

SleekAI's data sources map each filter's _filter_type, _query_var, and option list into the chatbot prompt. That means the bot understands that color is a checkbox filter writing ?color=black, that price is a range filter writing ?price=0,200, and that brand is a select filter writing ?brand=acme. A natural-language question becomes a single canonical filtered URL that the bot can return as a link.

Display conditions scope each bot by post type, taxonomy term, user role, logged-in state, and URL pattern. On a catalog page the bot helps narrow results. On a property listing page it speaks beds, baths, and price. Every chat is logged inside WordPress with the model name, token usage, and origin page, so you can see which filter combinations real visitors actually want.

Faceted archives are powerful and unforgiving. Visitors who know what they want race through the filter columns. Visitors who do not give up and leave.

On mobile, even a confident shopper bounces between four columns of checkboxes to combine a color, a size, a brand, and a price band. The result is high exit rates on the archives that should be doing the most work. A generic chatbot cannot help.

It does not know which filters exist on this page, which query vars they read, or which values are valid. It returns generic links that the AJAX provider ignores, and contradicts the filter labels next to it. JetSmartFilters already has the metadata that solves this.

Each filter is a CPT entry with a type, a query var, and a data source. SleekAI maps that into the chatbot prompt with point and click data sources, so the bot turns a sentence like "black backpacks under 200 with a laptop sleeve" into the exact filtered URL the provider expects. Per-bot display conditions then let you run a catalog bot, a real estate bot, and a directory bot from the same install, each scoped to its archive and running on a key the site owner controls.