The Luddy Course Coach is an AI consultant that walks through your course with you — sorts every graded assignment by how AI should (and shouldn't) touch it, stress-tests your big projects, and hands you a concrete redesign blueprint. One question at a time, at your pace.
Every graded element gets sorted: Bedrock work students must own unassisted, or Signature work where AI is a collaborator — each rated for how easily AI could substitute for the student.
Pick one foundation and secure it — including AI-powered ways to help students prepare: unlimited practice items, Socratic tutors, deliberately buggy code to diagnose.
Take one big project apart. Green Zones: parts a working professional hands to AI. Red Zones: parts the human does and answers for. Then stress-test both against a lazy AI and a savvy student.
A written blueprint: triage table, redesign pivots, paste-ready AI-status language for each assignment, and an action list you can take straight to a workshop.
Anything a student can type, a model can generate. Every recommendation the coach makes pushes graded effort toward three capacities the model can't fake:
Stating a problem precisely enough that a correct solution can be recognized — requirements, constraints, edge cases, data assumptions, and what counts as done.
Showing the output is correct rather than plausible — reading code you didn't write, testing, debugging, statistical validation, checking that a cited source exists and says what's claimed.
Owning what the system does to people once it runs — security, privacy, licensing, accessibility, bias, and defending a design choice to someone who isn't technical.
IU's data rules are treated as hard constraints; campus teaching guidance as strong defaults. The coach flags both as it goes.
Not a transcript of the conversation — a blueprint. Six sections, each with consultant insight: a risk you didn't raise, a hidden tradeoff, or a comparison to how the work is done in practice.
Build yours →| GRADED ELEMENT | CATEGORY | AI SUB. RISK |
|---|---|---|
| Weekly SQL problem sets | Bedrock | High |
| Client discovery interviews | Signature | Low — live client |
| Capstone build & repo | Signature | High |
| Final design defense | Signature | Low — oral, live |
Red Zones kept: data-model decisions, deciding what needs testing, the accessibility review. Pivot: replace the final-state grade with a commit narrative plus a 10-minute defense where one requirement changes live.
"AI status: allowed. You may use IU-approved AI tools for scaffolding and boilerplate. You are graded on your audit memo: what you changed, why, and one thing the model got wrong."
Name the course you want to audit — course number, term, and year. The coach confirms before anything else happens.
Syllabus, assignment prompts, rubrics — plus the things that carry the real requirements: starter repos, autograder specs, project briefs. No student work or grades.
The coach works at your pace and pushes back when an answer wouldn't translate into a concrete course decision. You end with the report.