Ivory Search is the go-to plugin when you need multiple, scoped search forms in WordPress: a global header search, a different one in a menu, a docs-only one in the sidebar, and an excluded-content version on landing pages. Each form is configured with included post types, taxonomies, and custom fields. It is excellent at scoping. It is still a list of results.

SleekAI uses those same forms as data sources. Pick an Ivory Search form ID, and the chatbot runs the visitor's question through that form, gets the included posts back, and reads their fields from wp_posts, wp_postmeta, and ACF. The model then writes one paragraph that picks the most relevant match and offers two alternatives. Visitors see an answer that respects every inclusion and exclusion rule the form already has.

Multiple forms become multiple chatbots. Under SleekAI multibot, a docs chatbot can be tied to the docs-only form, a marketing chatbot to the global form, and a member-only chatbot to a logged-in form. Each lives under its own display conditions, with the same single SleekAI install and a key you control.

Ivory Search exists because WordPress's default search is too coarse for any site with mixed content. A docs page, a product, a blog post, and a marketing landing all collapse into the same result list. Ivory Search fixes this with multiple search forms, each with its own inclusion and exclusion rules and its own front-end placement.

Site owners build a search experience that is genuinely scoped to context. The chatbot story should match. A docs chatbot should not pull in marketing posts.

A member assistant should not leak gated content to guests. A general header bot should respect global exclusions. Most generic chatbots flatten all of this back into one index.

SleekAI does not. It calls the Ivory Search form by ID on every turn, picks up exactly the rules the form enforces, and answers from inside that scope. Multibot then lets one site run several chatbots, each tied to a different form, each with its own display conditions and system message.

The visitor gets an answer that fits the page they are on, the editor keeps a single source of truth, and the conversation log tells the team where the scoped catalog still has gaps.