this post. one anonymized result. an online education company running 4-month live cohorts cut grading time per cohort by 70%. the ai scores each assignment on the client's own rubric. a person still handles the calls that need judgment.
the buyer here is a head of learning at an online course company. the cohorts are live and 4 months long. every assignment gets graded by hand. that work eats instructor hours, and instructor hours are the biggest line in the cohort's cost. margin leaks one submission at a time.
the fear was quality. drop a tool on the pile and grading gets faster and worse. the instructors said so out loud. so we did not replace them. we shipped an lms extension that scores assignments on their rubric and shows every score in a sample-audit dashboard.
#does ai grading make online courses cheaper without making them worse?
does ai grading for online courses lower cost without hurting quality?
in this engagement, yes. an online course company cut grading time 70% per cohort by scoring assignments on its own rubric with ai, then escalating the marginal 15% to instructors. instructor-quality nps held steady. the rubric stayed the client's. a person still owned every borderline call.
here is what kept quality flat. the rubric stayed the client's. the ai applied it consistently across the straightforward 85% of submissions. it escalated the marginal cases to a human. so instructors spent their hours on the 15% that actually needed judgment, not on the whole pile.
#where does the 70% actually come from?
not from grading faster on every paper. from grading fewer of them by hand. about 85% of submissions in these cohorts were clear passes or clear misses against the rubric. those are the ones the ai scored. the 15% that sat near a rubric boundary went to an instructor with the rubric line and the student's work side by side.
- the lms extension reads each submission and scores it against the client's existing rubric
- a confidence threshold splits the pile: clear cases scored, borderline cases escalated
- a sample-audit dashboard lets instructors pull any batch and override any score
- every score is traceable to the rubric line it came from
#how do instructors know the scores are right?
they audit. every batch shows up in a dashboard where an instructor can pull a sample, read the ai's score next to the rubric, and override it. that is why instructor-quality nps held steady. the people who feared the drop are the ones checking the work.
we are
we ship an lms extension that scores on your rubric and hands instructors a dashboard to audit and override, and we name the number: 70% less grading time per cohort.
we aren't
we are not a general grading bot that swaps your standards for its own and grades the whole pile unread.
this build sits in our platforms work: software that runs inside your product. the rubric design and the human-in-the-loop policy came out of the consultancy side. we do not ship a model and leave. we ship the extension, the dashboard, and the escalation rule together.
the client asked not to be named yet. the number is real and the build is real. we publish the outcome and one receipt because a case study with no receipt is just a claim. more of these are in the journal.
the ai grades the ones that are obvious. the instructor keeps the ones that are hard.
if you run cohorts and grading is eating your margin, the shape above repeats. same rubric, same instructors, fewer hours on the obvious pile. tell us what you're building and we will show you where the 70% would come from in your courses.