Education centres commonly use WhatsApp AI chatbots in three different ways. The first is an AI knowledge base, which answers general questions from approved course information, fee schedules and centre policies. The second is an AI connected to the education-centre management system. After identity verification, it can look up a particular student's lessons, fee status, attendance and remaining lesson balance. The third is an AI agent connected to the centre's systems. In addition to reading data, it can use approved tools to submit a leave request, create a make-up lesson request or start a follow-up workflow.
All three can appear as a conversation inside WhatsApp, but they do not have the same data access or permission to act. If a parent asks, "My son needs to miss his Saturday English class. Can he attend a make-up lesson?", a knowledge-base bot can explain the general policy. A system-connected, read-only bot can verify the student's class and check existing records, but it does not change them. An AI agent can submit the request and show eligible make-up options when the centre's rules permit it. Refunds, exceptional discounts, complaint decisions and other sensitive cases should still be handled by staff.
This article introduces the three operating models directly through common parent and student conversations. It does not rank them.
1. How does the first setup—an AI knowledge base—answer general questions?
An AI knowledge base answers from material supplied by the education centre. That material may include timetables, suitable grade levels, fees, tutor information, addresses, trial-lesson arrangements, leave policies and bad-weather procedures.
It is useful for questions that almost any parent might ask.
1.1 Example 1: Checking lesson times
- Parent: Which day is the Primary 5 English reading class?
- AI: The Primary 5 English reading class runs on Wednesdays from 5:00 pm to 6:30 pm. A Saturday morning class is also available, covering the same syllabus.
1.2 Example 2: Finding a suitable course
- Parent: My child's English grammar is weak. Which class should we consider?
- AI: The primary English consolidation programme covers grammar, reading comprehension and writing. Tell me the student's grade and I can list the relevant classes and times.
1.3 Example 3: Asking about fees
- Parent: How is the Form 1 mathematics class charged?
- AI: The Form 1 mathematics class is charged by term. The number of lessons and total amount are shown in the current course information. I can send you the latest course and fee sheet.
1.4 Example 4: Finding the centre
- Student: Which MTR exit is closest to the centre?
- AI: The centre is a short walk from Mong Kok Station Exit E2. Here are the address and map link.
1.5 Example 5: Checking bad-weather arrangements
- Parent: What happens if Signal No. 8 is issued two hours before class?
- AI: Under the centre's bad-weather policy, face-to-face lessons are suspended while the No. 8 tropical cyclone warning is in force. The centre will send the day's arrangement through WhatsApp.
1.6 Example 6: Asking about leave and make-up rules
- Parent: Is a make-up lesson always available after taking leave?
- AI: Make-up eligibility depends on the notice period and available places under the centre's policy. I can explain the general rules, but the centre must check the student's record to confirm an individual case.
1.7 Example 7: Asking about a trial lesson
- Parent: Can my child try a summer course before enrolling?
- AI: Trial lessons are available for selected summer courses. Tell me the student's grade, subject and preferred time, and the centre can confirm the suitable options.
A knowledge-base bot knows the centre's published information. It does not know who the person in the chat is. If a parent asks, "Have I paid this term's fees?" or "When is my daughter's next lesson?", the bot should not guess. It should explain that identity verification and a system or staff lookup are required.
2. What can the second setup find after connecting to the management system?
This setup combines the centre's knowledge base with the education-centre management system. After the parent's identity is verified, the AI can read authorised student records and answer personalised questions.
In this article, the second setup is deliberately defined as read-only. It can look up data, but it cannot change records or carry out a parent's request.
2.1 Example 1: Checking the next lesson
- Parent: When is Chan Siu Ming's next English lesson?
- AI: After identity verification, I found that Chan Siu Ming's next lesson is the Primary 6 English class on Saturday at 2:00 pm, taught by Miss Lee in Room 3.
2.2 Example 2: Checking two siblings' schedules
- Parent: Are both of my children attending lessons on the same day this week?
- AI: Chan Siu Ming's English class is on Saturday at 2:00 pm. Chan Siu Mei's mathematics class is on Sunday at 11:00 am, so they are on different days.
2.3 Example 3: Checking fee status
- Parent: I made a bank transfer last week. Has the fee been received?
- AI: The system shows this term's fee as "received", with a receipt date of 13 July and no outstanding balance.
2.4 Example 4: Checking remaining lessons
- Parent: How many mathematics lessons are left?
- AI: The course record shows four lessons remaining. The next lesson is on Sunday at 11:00 am.
2.5 Example 5: Checking attendance
- Parent: Was my son absent last week?
- AI: The system shows him as absent from Saturday's English class and present at Sunday's mathematics class.
2.6 Example 6: Checking a make-up request
- Parent: Has the make-up lesson I requested been confirmed?
- AI: The request status is "confirmed". The make-up lesson is scheduled for 25 July at 4:30 pm in Room 2.
2.7 Example 7: Checking materials and test dates
- Student: Which workbook do I need next lesson, and when is the test?
- AI: Please bring English Reading Practice Book 3. The class notice in the system shows the quiz scheduled for 29 July.
2.8 Example 8: Checking a waiting-list position
- Parent: Has a place opened in the Saturday class my daughter is waiting for?
- AI: The system still shows her on the waiting list, with no status update. The centre will contact parents in waiting-list order if a place becomes available.
2.9 What a management-system-connected AI can do
- Verify the parent or student before answering personalised questions.
- Read the latest class, lesson time, teacher and classroom records.
- Check fee status, due dates, balances and remaining lessons.
- Check attendance, absence, make-up requests and waiting-list status.
- Combine general policy information with an authorised individual record.
- Stop disclosure and hand over to staff if identity verification fails or information is incomplete.
2.10 What a management-system-connected AI cannot do
- It cannot formally submit leave, cancel a lesson or change a class.
- It cannot create a make-up booking, modify a request or hold a place.
- It cannot alter student, parent, class, attendance or payment records.
- It cannot approve refunds, discounts or exceptional arrangements.
- It cannot access records for a student the user is not authorised to view.
- It cannot invent an answer when the system has no data.
If a parent says, "Please change my phone number to 9123 4567," a read-only AI can explain the update procedure or hand the request to staff, but it does not write the new number into the system.
3. What can the third setup—an AI agent connected to the system—help process?
An AI agent can also read authorised education-centre data. The difference is that it can use configured system tools to carry out a defined action within approved identity, permission and policy boundaries.
Being able to act does not mean the agent can decide everything by itself. Each action requires a tool, a permission scope, workflow conditions and a human handover rule to be configured in advance.
For the wider distinction between conversation and controlled workflow execution, see AI Agent vs Chatbot: What Is the Difference?.
3.1 Example 1: Submitting a leave request
- Parent: Chan Siu Ming needs to miss this Saturday's English class.
- AI Agent: I have verified the student record. Please confirm that the leave date is 25 July.
- Parent: Yes.
- AI Agent: The leave request has been submitted. The reference is L-0725-018. The centre will confirm the make-up arrangement under its policy.
3.2 Example 2: Finding eligible make-up times
- Parent: Which make-up times are available after the leave request?
- AI Agent: The system shows that this request is eligible for a make-up lesson. The current options are Sunday at 3:00 pm or next Tuesday at 5:30 pm. Which time would you like to request?
3.3 Example 3: Creating a make-up request
- Parent: Sunday at 3:00 pm.
- AI Agent: The make-up request has been created and is currently "pending confirmation". The centre will send another WhatsApp message after staff confirm it.
3.4 Example 4: Uploading a receipt for review
- Parent: I have just transferred the fee. Here is the receipt.
- AI Agent: The receipt has been attached to a payment-review item for the student. Its status is "pending staff verification". I will not mark the fee as received until a staff member confirms it.
3.5 Example 5: Updating authorised contact information
- Parent: I want to change the phone number used for notifications.
- AI Agent: Please complete identity verification first. After verification, I can submit a phone-number update under the centre's approved process. A change involving the primary guardian may require staff confirmation.
3.6 Example 6: Creating a trial-lesson follow-up
- Parent: I would like an English trial lesson for a Primary 4 student, preferably on Saturday morning.
- AI Agent: I have recorded the grade, subject and time preference and created a trial-lesson follow-up. The centre currently has two relevant classes. Staff will confirm the student's level and class availability before contacting you.
3.7 Example 7: Notifying the right staff
- Parent: The make-up lesson is confirmed. Do I need to inform the original teacher?
- AI Agent: The workflow has added the result to the student's record and created a notification for the designated front-desk staff and tutor. You do not need to submit the same information again.
3.8 Example 8: Handing over an exception
- Parent: I missed the leave deadline, but there was a special reason. Can a make-up lesson still be approved?
- AI Agent: This request falls outside the standard make-up policy. I will retain the information you provided and pass it to staff, but I will not approve the make-up lesson.
3.9 Example 9: Organising a complaint for staff follow-up
- Parent: I am unhappy with the last class-change arrangement and would like the centre to follow up.
- AI Agent: I can record the student, lesson and details, create a follow-up case and route it to the responsible staff member. The centre will decide the complaint outcome and any remedy.
3.10 What an AI agent connected to the system can do
- Submit leave, make-up, trial-lesson or follow-up requests within its permission scope.
- Read a record and execute the next approved step in a configured workflow.
- Create or update specifically authorised records.
- Return a reference number, processing status and next step to the parent.
- Notify designated staff and retain an action and conversation record.
- Stop and hand over when it detects an exception, inconsistent data or a risk condition.
3.11 What an AI agent connected to the system cannot do
- It cannot independently approve refunds, exceptional discounts or fee waivers.
- It cannot decide a complaint or make an unauthorised promise.
- It cannot guarantee a class place, make-up place or teacher assignment.
- It cannot access or alter another student's data without permission.
- It cannot delete important records or bypass centre rules.
- It cannot perform an action if the system has no corresponding tool or permission.
- It cannot replace staff judgement in exceptional cases involving people, risk or commercial decisions.
4. How would the three setups answer the same parent question?
Assume a parent writes:
- Parent: My son needs to miss his Saturday English class. Can he attend a make-up lesson?
4.1 The AI knowledge-base answer
It explains the centre's general leave and make-up policy, such as the required notice period, eligible circumstances and the fact that places are subject to availability. It does not know whether this student is enrolled in the Saturday class or whether this request is eligible.
4.2 The management-system-connected answer
After identity verification, it can find the student's Saturday class, remaining lesson balance, previous leave record and current make-up-request status. It may confirm that the student has an English class this Saturday, but it does not submit leave or create a make-up request.
4.3 The AI-agent-connected answer
After identity verification, it can confirm the lesson, check the centre's leave and make-up conditions, submit leave and list eligible make-up times. Once the parent chooses a time, it can create the request and return the status. If the notice deadline has passed or the parent asks for an exception, it stops the automated workflow and hands the case to staff.
The point is not that one setup is universally better. The same WhatsApp message produces a different result depending on the data and permissions behind the conversation.
5. What basic controls should apply before any of the three replies?
Every education-centre WhatsApp AI chatbot should have three basic controls:
1. Identity verification: before verification, answer only general questions and do not disclose a student's lessons, fees or attendance. 2. Permissions and records: define which data can be read and which actions can be performed, and retain records for important operations. 3. Human approval: staff decide refunds, discounts, complaints, sensitive-data cases and exceptional arrangements.
For a broader view of how enquiries, scheduling, fees and follow-up can sit in one controlled workflow, read the Education Centre AI Agent System Guide. For a practical approval boundary, see AI Automation for Tuition Centres: 3 Do's and 3 Don'ts.
An education centre can begin by listing its everyday WhatsApp enquiries, then dividing them into general questions, personal-data lookups and requests that require a system action. That list is more useful than simply asking whether the centre "wants a chatbot".
If the centre wants to bring WhatsApp, student records, lesson arrangements and internal workflows into one manageable system, oneflash can help map the data sources, permissions, staff approval points and AI action boundaries. You can also start with the oneflash education-centre system.