Complete AI Training
Sign inGet my AI kit

Your job's AI kit

Get your AI kit

Tell us who you are and what you do. We show you your kit right away and email you the link: skills, prompts, AI agents, MCP servers and courses for your job.

500+ jobs ready, and we make a kit for any other job. No payment needed to look.

Share

AI agent for online course creators

Drop-Off Diagnosis Agent

A fix that improves completion at the lesson where most students leave.

AnalysisDecisionsEducation
See how it worksGet access
Drop-Off Diagnosis Agent: what goes in, what the agent does and what you get

What it does

Completion drops from 80% in lesson 3 to 41% in lesson 6, and you guess why. This agent finds the lessons with the biggest drop in views or completion and compares length, difficulty, format and student comments around them. It forms hypotheses, for example that lesson 6 is 40 minutes long and has no practice task, and proposes a small fix to test. After the next cohort goes through, it checks whether the numbers improved. If not, it tries another hypothesis. Edge case: a sharp drop is due to a broken video link, so the agent reports the bug before any content change. The creator approves changes.

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.

Start and resultWhat it doesA check on its own workWaits for your OKGoes back and retries
Yes, continueApprovedYes, continueNoNo 1 STARTS WHEN Enough cohort data collected 2 USES A TOOL Pull view and completion data by lesson 3 DOES Find the lessons with the largest drop 4 USES A TOOL Compare length, format, difficulty and comments forthose lessons 5 CHECKS THE RESULT Is the drop caused by a technical fault such as abroken video? If not: report the fault and wait for the fix beforeanalyzing content. Back to step 2. 6 DOES Form hypotheses and rank them 7 DOES Propose one small change to test 8 YOU APPROVE Creator approves the change 9 USES A TOOL Check the next cohort's numbers for the lesson 10 CHECKS THE RESULT Did completion improve by at least 10 points? If not: try the next hypothesis. Back to step 4. 11 RESULT Drop-off findings and test record
Read the steps as a list
  1. Enough cohort data collected
  2. Pull view and completion data by lesson
  3. Find the lessons with the largest drop
  4. Compare length, format, difficulty and comments for those lessons
  5. Is the drop caused by a technical fault such as a broken video?If not: report the fault and wait for the fix before analyzing content. Back to step 2.
  6. Form hypotheses and rank them
  7. Propose one small change to test
  8. Creator approves the changeThe agent waits here for your OK.
  9. Check the next cohort's numbers for the lesson
  10. Did completion improve by at least 10 points?If not: try the next hypothesis. Back to step 4.
  11. Drop-off findings and test record

How it decides

It picks the hypothesis best supported by the data and tests one change at a time so the effect can be seen.

  • Test one change per lesson at a time.
  • Count improvement only after at least 30 students pass the lesson.
  • Rule out technical faults first.
  • Do not change graded content without approval.

Make it yours

Every agent is a starting point. You choose these settings for your own situation.

  • Improvement target (default 10 points)
  • Minimum student count
  • Metrics watched
  • Test length
  • Review frequency

What keeps you in control

It always asks you first

  • Every proposed change
  • Any change to paid content

Hard limits

  • Do not edit course content itself.
  • Do not draw conclusions from fewer than 30 students.

It stops when

  • Done: completion improved or hypotheses are exhausted and logged
  • Stop: too few students to judge, so the agent recommends waiting

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.

10 minto set it up in your AI
5 AIsChatGPT, Claude, Copilot, Gemini, Grok
  • 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
Get access to this agent

An example run

What happensLesson 6 loses 39 points of completion. The video plays fine, so the fault check passes. The lesson runs 38 minutes with no exercise. The agent proposes splitting it into two 15-minute parts with a short task. Priya approves. After the next 45 students, completion rises 8 points, below target, so the agent suggests a recap quiz next.

More agents for online course creators

Learning · Education

Learning Misconception Practice Agent

Each learner overcomes their specific misconceptions, with evidence the instructor can see.

8 steps1 check13 jobs
Monitoring · Automation

Course Resource Link Repair Agent

Every course resource works, still matches the lesson and is accessible.

8 steps2 checks2 jobs
Creating · Quality control

Instructor Assessment Variant Agent

New practice questions that test the intended objective at the intended difficulty, each with a verified answer.

8 steps1 check5 jobs
Quality control · Education

Course Prerequisite Mapping Agent

Every lesson's prerequisites are taught earlier in the course.

7 steps2 checks6 jobs