# Chatbots for Training Centers and Schools: What to Hand Over

**The same questions every week: “When does the next class start?”, “How much is the fee?” If those answers were automatic, what would you do with the time?**

> A training-center chatbot should answer repeated questions: class schedules, fixed fees, enrolment steps, trial lessons and class reminders. Marks, complaints, special discounts and a student’s private matters must go to a person, and it must not ask children for extra personal details. It should answer only from texts the center has approved.

Author: Saman Ghahreman · Published: 2026-10-11 · Updated: 2026-10-11 · Canonical page: https://solocodelab.com/en/blog/chatbot-for-training-center-and-school/

Every training center and online class answers the same questions each week: “When does the next class start?”, “How much is the fee?”, “Do you have a trial lesson?”, “What if I miss a session?” A [Telegram or WhatsApp chatbot](https://solocodelab.com/en/blog/telegram-whatsapp-chatbot-for-business/) can lighten this load and free the manager for more important work.

## What can a training-center chatbot do?

| Task | Should the chatbot answer? | Note |
| --- | --- | --- |
| Class schedules and times | Yes | From the center’s official timetable |
| Enrolment terms and steps | Yes | Documents, dates and payment method the center has approved |
| Fixed fees and announced discounts | Yes | Only what is officially announced; special discounts go to the manager |
| Booking a trial lesson | Yes, with the enrolment officer’s approval | The chatbot prepares the request |
| Class and homework reminders | Yes | A short, polite message |
| Marks, complaints and a student’s personal matters | No | Only a person |
| Study advice and level placement | No, only first help | The final answer comes from an adviser |

### The path of a parent’s message in a training-center chatbot
1. **Parent or student message**: About a class, enrolment, fees or times.
2. **Sort the message**: Is it a simple question, or marks, a complaint or a personal matter?
3. **Answer from official text**: Only from the timetable and texts the center has approved.
4. **Handoff to a person**: A sensitive matter reaches the enrolment officer or manager with a chat summary.

## Where should you not rely on a chatbot?

- Telling a student’s marks or academic status to someone whose identity is unknown.
- Handling a student’s personal, family or behaviour problems.
- Deciding on fee refunds, moving a class or making exceptions.
- Any private conversation with a child without a parent’s or the center’s supervision.

UNESCO’s guidance on generative AI in education stresses that its use should be human-centred, student data must be protected, and age limits and supervision must be respected.

## How do you keep student data safe?

1. For enrolment, collect only a name, a phone number and the class wanted.
2. Don’t ask children for personal details, photos or documents; parents or the enrolment officer provide these.
3. Give marks and academic status only after identity checks and through a person, not the chatbot.
4. Tell people from the start that they are talking to a bot and can reach a person whenever they want.
5. Before launch, test that [tricky instructions and misuse](https://solocodelab.com/en/blog/responsible-ai-2026-business-checklist/) cannot pull the bot off course.

Child data protection laws differ from country to country. Before storing student information, talk to a legal specialist or the center’s technical manager.

## When should a person step in?

If the chatbot fails to give a confident answer twice in a row, or the message is about marks, a complaint or a personal matter, it should pass the chat, with a summary, to the enrolment officer or manager. The full method is in [AI customer support for small business](https://solocodelab.com/en/blog/ai-customer-support-for-small-business/). For the limits in sensitive work, see also [chatbots for clinics](https://solocodelab.com/en/blog/chatbot-for-clinic-and-medical-office/).

## How do you build the first version?

1. List about twenty repeated questions from parents and students and prepare the official answer to each.
2. Keep the class timetable, fees and enrolment terms in one up-to-date place; the chatbot answers from that text with the [RAG method](https://solocodelab.com/en/blog/how-rag-chatbot-works/).
3. Build and test the handoff to a person before anything else.
4. Try it for two weeks with a few parents and the enrolment officer, and add the unanswered questions.
5. Measure the result: how many phone calls fell? How many enrolments started through the chatbot?

You can see a small sample on your own center’s real questions before paying; read [why we show a live sample first](https://solocodelab.com/en/blog/show-live-sample-before-you-pay/).

> Want fewer repeated phone calls at your training center? [Send your project brief](https://solocodelab.com/en/contact/?service=ai-automation) to get a written proposal and a live sample for [automation and chatbots](https://solocodelab.com/en/services/ai-automation/).

## Frequently asked questions

### Can a chatbot tell a student’s marks?

Better not. Marks and academic status are sensitive and should be given only after identity checks, by a person or the center’s official system.

### Can a chatbot replace the enrolment officer?

No. It answers repeated questions and prepares enrolment requests, but approval, special discounts and problems stay with a person.

### What should we do with children and teenagers?

Don’t collect extra personal details, keep the chat simple and limited, and pass sensitive matters to parents or the center’s staff.

### Is it suitable for online classes too?

Yes. It can answer about class links, times, homework and common connection problems, and send reminders.

## Sources

- [Guidance for generative AI in education and research](https://unesdoc.unesco.org/ark:/48223/pf0000386693) — UNESCO
- [OWASP Top 10 for Large Language Model Applications](https://owasp.org/www-project-top-10-for-large-language-model-applications/) — OWASP
