GoodLayers LMS powers many course sites built on GoodLayers themes. It registers a stack of custom post types: gdlr_lms_course for courses, gdlr_lms_lesson for lessons, gdlr_lms_quiz for quizzes, gdlr_lms_question for the bank, and gdlr_lms_certificate for credentials. Learner progress and quiz results are stored in postmeta and a usermeta map keyed by user_id and course_id. The dashboard surfaces course progress as a percentage and a checkbox list, then stops.

SleekAI maps those post types and postmeta into variables. The bot reads which lessons a student has marked complete, which quiz attempts passed or failed (with score and timestamp), and which certificate is eligible. A display condition scoped to the GoodLayers course post types or the student dashboard URL ensures the bot only appears where it can help. Multibot can run separate course-specific bots, each grounded in its own course content via course_id.

Generic chatbots see the course title and lesson list in the page HTML but nothing else. They cannot say which lesson is next in the configured order, why a quiz attempt failed, or how many points short the student was for the certificate. SleekAI quotes the actual lesson title, attempt score, and certificate rule, and the system instruction keeps it from spoiling future quiz answers stored in gdlr_lms_question.

GoodLayers LMS does an honest job of structuring courses and showing progress, but a progress percentage on a dashboard is a low-resolution signal for a struggling student. A learner who failed a chapter 3 quiz has many possible reasons: a missed lesson, a misremembered formula, a confusing question. The dashboard cannot tell them which.

The default fix is to re-watch the chapter, which is slow and demotivating. A coach who can say you missed two questions on meta tags, lesson 3.2 covers it, you stopped at the 6 minute mark, the meta description section starts at minute 8 is dramatically more effective. SleekAI provides exactly that coaching by reading the same data GoodLayers LMS already stores.

The same idea scales to certificate unlock rules, retake policies, and pre-enrolment questions, all of which currently route through email support or remain unanswered. Over a semester the conversation log becomes a curriculum-improvement dataset: lessons that drive heavy questions are candidates for re-recording, quizzes that fail at unusual rates are candidates for re-wording. The plugin gives you the data.

The bot turns it into outcomes.