JobBoardWP registers a jb-job custom post type for every job listing, with taxonomies jb-job-category and jb-job-type for category and full-time/part-time/remote classification. Listing details live in wp_postmeta with keys like jb-location, jb-salary, jb-company-name, jb-company-website, jb-application, jb-expiry-date, and jb-is-featured. The plugin also adds Google structured data for job-search rich results.

SleekAI maps those meta keys and taxonomies into the chatbot's system message. A visitor asking "any remote senior dev roles paying over $120k" gets a filtered answer pulled from real listings, not a vague placeholder. The bot also knows the application URL or email stored in jb-application, so it can hand off cleanly to the apply form without sending the visitor on a click hunt.

Generic chatbots can't see the jb-job post type as anything other than "some custom posts" and miss the salary, location, and expiry meta entirely. They invent role descriptions or refuse to answer. SleekAI's display conditions can scope a different bot to single job pages (where it knows the current listing's data automatically) versus the jobs archive (where it queries across all active listings).

JobBoardWP is a clean, lightweight job board plugin, and the default JBP UI does a good job at filtering open roles. The problem is that visitors don't always know how to filter. They type "remote senior dev" into a generic chatbot and either get a hallucinated list of companies that don't exist on the board, or they get a polite refusal because the bot can't see the jb-job custom post type at all.

SleekAI fixes that by reading the actual jb-job posts, the jb-job-type and jb-job-category taxonomies, and the salary, location, and expiry meta keys directly. A seeker asks in plain language and gets the same filtered result the search form would produce, with the bonus of knowing how to apply. On single job pages, the bot auto-injects the current listing's data so visitor questions about that specific role get answered without retyping the title.

Employer-side, a different bot scoped to the dashboard URL via display conditions answers questions about featured-boost expiry and submission status, without exposing internal data to the public bot. Token usage stays low because seekers ask short questions and the bot returns short answers. You bring your own OpenAI, Anthropic, Google, or OpenRouter key, so per-seat pricing doesn't get in the way of running a free job board.

Conversation logs surface where seekers drop off, which informs future board content.