BrightPath Academy lifted completion rates 31% and added $1.4M in retained tuition with AI engagement analytics
A 3,000-student online bootcamp was losing nearly half of every cohort to silent dropouts. An AI engagement platform flagged at-risk students the moment they disengaged — turning mentor time from guesswork into targeted intervention that recovered $1.4M in tuition that would have walked out the door.

This is an illustrative case study. Company names, identifying details, and testimonials have been changed to protect client confidentiality. The technical solution and target outcomes are representative of the work we do. We extend the same confidentiality to every client.
Where they started
BrightPath had 3,000 active students but a 52% completion rate — meaning roughly half of every cohort quietly disappeared before finishing. Instructors couldn't see who was disengaging until a student missed their third assignment, by which point re-engagement rarely worked. Engagement data was scattered across the LMS, the discussion tool, and the video platform, so no one had a single view. The finance team had tied the dropout number to $1.4M in refunded and forfeited tuition every year, and the marketing team was spending 40% more to backfill cohorts that kept leaking.
What we built
We built a custom student engagement platform on our AI & Machine Learning stack with a RAG knowledge base over all course content, real-time engagement analytics that fuse LMS activity, video watch-time and discussion participation into one engagement score, and automated early-warning alerts that trigger mentor intervention workflows the moment a student's score drops below cohort baseline. Mentors now open their day to a prioritized queue instead of a roster, and the RAG layer surfaces the exact lesson each struggling student got stuck on.
Target outcomes
More from this project
We can now see exactly which students need help, when they need it, and what kind of help. The first cohort after go-live had our highest completion rate ever — and the $1.4M we stopped refunding paid for the platform four times over.
Representative testimonial — name and role changed to protect client identity.
