Student questions sent by email or posted in the course community get an answer within a minute from an AI agent that reads the creator's own lessons, quotes the relevant explanation and links to the exact lesson and timestamp.
Typical results
Under 1 min
Typical reply time, day or night
≈8 hrs/wk
Creator time no longer spent on repeat questions
Weekly
Ranked list of unanswered topics for new lessons
Figures reflect typical results for this kind of automation.
The challenge
Students in a self-paced program ask questions around the clock, and the creator answers them alone between recording and marketing. Most questions are already covered somewhere in a lesson, but finding the right minute of video takes time. Replies slip to the next day, students stall or ask for refunds, and the same questions keep coming back.
Before and after
Creator hours per week on student questions
−80%Minutes to a first reply
−99%The automation
An AI agent is grounded in the program's lesson transcripts, worksheets and FAQ. When a question arrives by email or in the community chat, it finds the passages that answer it and replies within a minute, quoting the explanation with a link to the exact lesson and timestamp. Billing, refund and out-of-scope questions go to the creator's Slack with the student's history attached. Every week the creator gets a ranked list of topics the agent could not answer.
How it works
- 01
A question arrives
A student emails the course address or posts in the course community chat, at any hour.
- 02
Course material searched
The agent searches lesson transcripts, worksheets and the FAQ for the passages that answer the question.
- 03
Reply with the source
Within a minute it replies with the explanation and a link to the exact lesson and timestamp.
- 04
Escalations to the creator
Billing, refund and out-of-scope questions go to the creator's Slack with the student's purchase and question history.
- 05
Weekly content gaps
Every Monday the creator gets a list of topics the agent could not answer, ranked by how often they came up.
Starts from
- Student emails
- Course community chat
- Transcripts and FAQ
Delivers
- Answer with lesson timestamp
- Escalations to Slack
- Weekly content-gap list
The impact
About 70% of student questions get an accurate answer within a minute, at any hour, and the creator spends roughly 8 fewer hours a week on repeat questions. Students keep moving through the program, and the weekly list of unanswered topics turns into a clear backlog for new lessons.
Questions about this automation
How does it avoid making up answers?
The agent answers only from the course material it has been given: lesson transcripts, worksheets and the FAQ. Every answer quotes the passage it relies on and links to the lesson. When nothing in the material matches well enough, it says so, tells the student the creator will follow up and escalates to Slack.
Which channels and tools does it work with?
It reads a support email inbox and the course community chat, with Make connecting each channel and OpenAI writing the answers. Transcripts are pulled from the course platform or uploaded as text files. Escalations go to the creator's Slack with the student's purchase details and earlier questions.
What happens when a new lesson is added?
New lessons are transcribed and added to the agent's material on a schedule, so answers reflect the current version of the program. Topics on the weekly unanswered list drop off once a lesson covers them. This is a multi-step build, usually rolled out in stages, starting with email.