AI agent for training instructors
Learning Misconception Practice Agent
Each learner overcomes their specific misconceptions, with evidence the instructor can see.
What it does
Adult learners often make the same specific mistake again and again, while they get generic revision exercises. This agent reads each assessment and maps wrong answers to known misconceptions from the curriculum. It starts with the misconception that appears most often and picks an approved exercise aimed at it. It scores both the answer and the learner's explanation. If the number is right but the reasoning is still wrong, it chooses a simpler diagnostic exercise that exposes the error. After two rounds with a correct answer and correct reasoning, it moves to transfer practice in a new context. If the same error remains after three targeted items, it sends the case to the instructor. Edge case: a learner who guesses right without any explanation is treated as not yet mastered.
How it works
Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.
Read the steps as a list
- Learner completes an assessment
- Map errors to curriculum misconceptions
- Pick an approved exercise for the top misconception
- Score the answer and the explanation
- Is both the answer and the reasoning correct?If not: choose a simpler diagnostic exercise. Back to step 3.
- Move to transfer practice after two correct rounds
- Instructor reviews unresolved confusionThe agent waits here for your OK.
- Personalized practice path with evidence
How it decides
It chooses the next exercise from the current misconception and earlier responses; correct reasoning plus correct answer twice means mastery, repeated errors mean a simpler diagnostic exercise.
- Which misconception first: the most frequent in the learner's errors.
- Mastery: correct answer and reasoning in two consecutive items.
- Escalate: same error after three targeted items.
Make it yours
Every agent is a starting point. You choose these settings for your own situation.
- Mastery rule (default correct answer and reasoning in 2 consecutive items)
- Attempts before instructor review (default 3 targeted items)
- Exercise bank the agent may draw from (default approved course items only)
- Session length (default 15 minutes)
- Who sees unresolved cases (default the course instructor)
What keeps you in control
It always asks you first
- High-stakes grading
- Progression decisions
- Changes to accommodations
Hard limits
- Only approved items.
- No grading decisions.
It stops when
- Done: misconceptions resolved.
- Escalated to instructor.
Set it up
We guide you through the set-up, step by step
Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.
- One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
- The agent then walks you through connecting your own data, one source at a time
- A downloadable copy with the flow chart, the rules and the full guide
