Building an AI Learning Coach Inside a PM Training LMS
A RAG assistant living inside a PM course platform, answering questions with citations to actual slides
✳The problem they went after
Learners working through a PM curriculum in an LMS get stuck in three predictable places: finding the right content at the right moment, questions that go unanswered mid-lesson, and passive consumption that never turns into reflective mastery. None of these are content problems. Progress slows, and what does get learned stays fragmented.
✳What they actually did
PM Coach is a RAG-based assistant sitting in a lightweight drawer over the existing LMS, leaving the platform layout untouched. Learners ask anything and get an answer citing the specific slides, lessons and sessions it came from, closing with a single actionable next step. The roadmap moves from chat with slide retrieval, to a version adding quizzes and spaced repetition, to a full AI training layer with mock interviews, personalised pathways and assignment feedback.
✳What you can take from it
Worth stealing
- RAG answers cite specific slides, lessons and sessions
- Every response ends with one actionable next step
- Drawer overlay avoids touching existing LMS layout
- Roadmap staged from retrieval to practice to feedback
Read this if
Designing an AI layer over content the company already owns
The transferable bitLearners need direction more than they need additional content. That claim reframes the AI from a content generator into a navigator, which is why citations and one clear next step matter more here than fluent prose. Delivering it in a drawer rather than a redesign is the other lesson, keeping integration cost close to zero.
✳The submission itself
Open the PDF →Reading on a phone? The embedded viewer is cramped — open the PDF in a new tab instead.
