Feature Proposal
AI-Graded Exercise
Allow exercises to reach completion using AI Practice Coach scoring alone, without requiring a human grader.
Client: ZoomInfo (contracted)
Status: Decide
Scope: Exercise · Task Completion · Config
Problem
Today, an exercise task never reaches completion until a human grader creates a scorecard on the handed-in response. This creates two problems for clients:
- Practice exercises jam up the grading queue when the exercise is just for rep practice, not formal evaluation.
- Courses containing exercises can't be completed without human graders, blocking learning program progress.
Core change
AI Practice Coach already auto-generates scorecards using the exercise's scorecard template. The core change is narrow: task completion logic must accept an AI-created scorecard as satisfying the grading requirement when AI-Graded mode is enabled on the exercise.
Key workflow distinction
The learner workflow for AI-Graded exercises is fundamentally different from human-graded exercises. The learner sees the AI score before handing in — hand-in is a confirmation step, not a scoring trigger.
Human grading (today)
- Learner records response
- Learner hands in
- Waits for human grader
- Gets score
AI grading (new)
- Learner records response
- AI scores automatically
- Learner reviews score & feedback
- Hands in (or re-records)
Configuration
- Company config toggle — gates feature availability for controlled rollout and EU AI Act compliance.
- Exercise property — new option under "who will evaluate" enables AI-only grading. When selected, human grader selection (users, groups, respective relationships) is disabled. AI is the sole grader; no pooling settings needed.
- Legal acknowledgment — required switch with regulatory awareness text and link to help documentation.
Exercise completion logic
- No passing threshold → handed in + AI scorecard = finished (complete)
- Passing threshold met → pass (complete)
- Passing threshold not met → redo required (not complete; learner re-records)
- Course completion: no pass/fail at the course level — just checks all required items are complete.
Hand-in validation
For AI-Graded exercises, block hand-in unless an AI-generated scorecard exists on the response being handed in. Applies to both web and mobile. This prevents unscored responses from completing the task and naturally handles AI scoring errors — if AI Practice Coach fails, the learner can't hand in until it succeeds.
Toggle-off behavior
If the company config is disabled or the exercise is switched back to human grading:
- Admin must assign human graders to affected exercises.
- Existing completions remain valid (not retroactively invalidated).
- New responses going forward require human grading.
Channels
Channel designation still applies — channels make handed-in responses available for review (a sharing function, not a grading function). Manager completion notifications are out of scope (separate feature).
V1 scope
| Item | Status |
|---|
| AI scorecard satisfies task completion | In |
| Company config gate | In |
| Exercise property for AI-only grading | In |
| Legal acknowledgment switch | In |
| Hand-in validation (require AI scorecard) | In |
| Updated hand-in text (web + mobile) | In |
| Re-trigger scoring via transcript edit | In (existing, not blocked) |
| Document Response exercises | Out (no AI scoring) |
| Dispute / appeal mechanism | Out |
| AI draft, human confirm | Out |
Open items to flag
- Reporting Reports, dashboards, exports, and data feeds need to display AI grader identity and feedback. Reporting PM to address.
- Legal UX Authority gap: exercise creator may lack legal authority for the acknowledgment. CS-created exercises and copied exercises bypass the prompt. Needs design refinement — possibly admin-level consent + exercise-level informational notice.