This guide follows the admissions chats an AI Agent can take on — who writes, what they ask, how the agent asks back and where the chat ends. All names, courses and fees in the examples are made up. For the same idea in another industry, see AI Agents for Real Estate Sales.
AI Agents for Education and Coaching: Quick Answer
The short version:
What it does: an AI Agent answers every enquiry on WhatsApp the moment it arrives, asks what the student needs, recommends a course from your own course information and moves the person to a demo class or a counsellor.
Where it helps most: evenings and weekends, the same fee and timetable questions asked all day, enquiries that arrive from ads, and the first round of qualification.
What stays with people: discounts and scholarships, admission decisions, refunds and anything the agent has no approved answer for.
In WABB: a knowledge database, an AI Agent with instructions and boundaries, and a fallback that hands the chat to your team. Booking into a calendar or CRM needs an integration.
Where Education Enquiries Get Lost
It is 9:40 pm. A marketing executive has finished work, taps a course ad and writes: “Which course is right for me?” The counsellors have gone home. The answer comes at 10 the next morning — and by then she has spoken to another institute. This is an illustration, not a case study, but every admissions team knows the scene.
Education businesses sell something that needs a conversation: which course, which batch, at what price, starting when. Five things go wrong in that conversation again and again.
Replies arrive hours late
Students and parents write in the evening, between lectures or on a Sunday. If the first answer comes the next morning, the student may already have written to another institute.
The same five questions all day
Fees, batch timings, syllabus, online or offline, can I try a class. Counsellors retype the same answers again and again.
Qualification is manual
Before recommending anything, someone has to learn the goal, the current level, the exam date and the budget — usually on a call.
Ad leads go cold
A tap on a Click-to-WhatsApp ad starts a chat. If nobody answers within minutes, the money spent on that tap is gone.
Context is lost between people
The next counsellor asks the same questions again, and the student repeats the story for the third time.
None of these is dramatic on its own. Together they are the gap between the number of enquiries and the number of enrolments.
What an AI Agent Can Actually Do for a Coaching Business
An AI Agent is not a menu of buttons. It reads a free-text message, uses your own course information and decides what to ask or do next. In an admissions chat that comes down to seven jobs:
| Job | The student or parent says | The agent |
|---|---|---|
| Recommend a course | “Which course is right for me?” | Asks the goal and current level, matches them to courses in your knowledge base and explains why. |
| Answer fees, schedule and format | “Is there a weekend batch? What is the fee?” | Answers from your fee sheet and timetable, and says what is included. If something is missing, it says so instead of guessing. |
| Qualify the lead | Nothing yet — the counsellor still needs the facts | Collects the three or four things a counsellor would ask — goal, level, date, preferred timing — one question at a time. |
| Follow a change of plan | “Actually it is for my nephew, he is in Class 9.” | Updates the picture and carries on, instead of restarting the form. |
| Request a demo class | “Can I try a class first?” | Offers the slots you allow, confirms the choice and records the name. Booking straight into a calendar needs an integration. |
| Hand over with context | “I would like to talk to someone.” | Moves the chat to a counsellor, who opens it and sees every answer already given. |
| Show what it could not answer | A question your materials do not cover | Falls back to a person and logs the question, so you can add the answer and the next student gets it. |
Notice what the list leaves out: negotiating a discount, promising a result, deciding who gets a seat. Those are decisions, not answers, and you decide who takes them — see How It Works in WABB.
Four Use Cases: From the First Message to the Handover
Each case starts with a problem the business already has, shows the chat that solves it and ends with what the team receives. The conversations are illustrations, not transcripts of real students. Read them for what the agent does between the lines: it asks one thing at a time, keeps what it learned, refuses to promise what it cannot and ends on a clear next step.
Course recommendation with a trial request
Online courses and skill programmes
The problem. People arrive from ads and social media without knowing which programme fits. A counsellor starts every call by asking the same background questions.
Admissions assistant
AI Agent · WhatsApp
Hi, I saw your ad. I want to move from marketing into data. Which course should I take?
Hi Ananya, happy to help. A couple of quick questions so I point you to the right one. Have you worked with Excel, SQL or Python before?
Excel only. I build reports at work, no coding.
That is a good base. Data Analytics Foundations starts from Excel and goes to SQL and dashboards, with no coding needed. Are you looking to switch jobs or grow in your current role?
Switch, ideally within 6 months.
Then the 16-week weekend batch fits: it ends before your six months and includes a portfolio project for interviews. The fee is ₹24,000 and the fee sheet lists what is included. Would you like to try a free session first?
Yes. And is a job guaranteed after the course?
I cannot promise a job. A counsellor can show you what past learners did with the project. Trial slots: Saturday 10 AM or Sunday 6 PM?
Saturday 10 AM.
Noted: Saturday 10 AM. A counsellor will confirm it here in this chat.
- Asked
- Background, goal and timeline — one question at a time.
- Remembered
- Excel only, switching careers, six months, weekend batch.
The result. The counsellor speaks to someone who already knows which programme they want, what it costs and when it starts.
Exam batch and demo class
Exam preparation: IELTS, entrance tests, certifications
The problem. Enquiries come late at night with a test date approaching. Demo slots get filled through a back-and-forth of calls and messages.
Admissions assistant
AI Agent · WhatsApp
need 7 bands in ielts, exam is in march, any batch?
Hi! Yes, we run IELTS batches. First, what was your latest mock band and which exam date are you aiming for?
last mock 6, exam 14 march
Going from 6 to 7 depends mostly on your weakest section. Which one is it: Listening, Reading, Writing or Speaking?
writing for sure
Our Writing Intensive batch meets three evenings a week with weekly essay feedback, and it finishes a week before 14 March. Want a free demo class? Tuesday 7 PM or Thursday 7 PM.
thursday
Thursday 7 PM it is. What name should I put it under?
Daniel Joseph
Thanks, Daniel. I have passed your request to the team with your target, your mock score and your weak section, so you will not have to repeat them.
- Asked
- Latest score, exam date, weakest section, name.
- Remembered
- Band 6 now, band 7 target, 14 March, writing.
The result. A demo request waits for the team in the morning with the target, the score and the weak section already written down.
A parent, a correction and a scholarship question
Tuition centres and school coaching
The problem. Parents ask about fees, boards, timings and location, often for more than one child, and want reassurance before they commit.
Admissions assistant
AI Agent · WhatsApp
Do you take class 10 maths? what is the fee
Hello! Yes, we have Class 10 maths. Is it for CBSE, ICSE or a state board, and would you like online classes or at our centre?
Sorry, it is for my nephew, he is in class 9. CBSE, at the centre.
No problem — Class 9 CBSE maths at the centre. The batch meets Monday, Wednesday and Friday at 5 PM and the term fee is ₹9,000. What is he finding hardest right now?
Algebra. Is there a scholarship? His marks are good.
Scholarships are decided by our centre head, not by me, so I will pass the request on with his class and the algebra note. When is a good time to call you?
After 6 pm
Done. The centre head will call after 6 PM about the Class 9 CBSE batch and the scholarship.
- Asked
- Class, board, mode of study, weakest topic, time to call.
- Remembered
- The correction: nephew, Class 9, not Class 10. Fee and timings came from the Class 9 batch.
The result. Fewer repeated calls, and the centre head only sees the requests that genuinely need a decision.
Career coaching: no promises, a discovery call
Career, life, fitness and language coaching
The problem. A coach’s time is the product. Unqualified calls waste it, and many people ask whether the programme will guarantee a result.
Admissions assistant
AI Agent · WhatsApp
How does your 1:1 career coaching work? Will it get me a promotion?
It is eight weekly 1:1 sessions of 45 minutes with a coach, plus short tasks between sessions. I cannot promise a promotion, because results depend on your goal and effort, but we begin by turning the goal into a concrete target. What is your role now, and what are you aiming for?
Team lead in IT, want to become a manager.
Is there a timeline, for example your next review?
Appraisal is in 5 months.
That helps. The best next step is a free 20-minute discovery call, where a coach checks the fit and outlines a plan. Shall I request one? Mornings or evenings?
Evenings
Requested for evenings. The coach will confirm a time here and will see your role, goal and timeline in this chat.
- Asked
- Current role, target role, timeline, preferred time of day.
- Remembered
- Team lead in IT, manager goal, appraisal in five months.
The result. The coach starts the call knowing the person’s role, goal and timeline, and spends the time on fit instead of basics.
Many of these chats begin with an ad. If yours do, the Click-to-WhatsApp ad examples include a free course preview and a demo-class ad with the first reply for each.
How It Works in WABB: Knowledge, Agent, Handoff
The four use cases above run on the same six pieces. In WABB they are set up once in AI Lab, and the step-by-step guide to building the agent shows every screen. Two things come first: the WhatsApp number has to be connected to the WhatsApp Business Platform — the API requirements checklist lists what to have ready, including the business website Meta checks when you verify the business — and AI Starter has to be active under Settings → Subscriptions, with credits available in Billing → Usage.
| Piece | What it does for an institute | Where in WABB |
|---|---|---|
| Knowledge | Course list, fee sheet, batch timetable, FAQs, refund and demo policies. This is where the agent finds the fee and the batch dates. | AI Lab → Knowledge databases |
| Role and instructions | What the agent is responsible for and how it behaves in repeatable situations, such as asking the goal before recommending a course. | Create AI Agent → Basic setup and Logic |
| Boundaries | What it must not do: promise results or jobs, decide discounts, handle refunds. | Logic |
| Fallback and handoff | A short message, a retry limit and an action that moves the chat to a counsellor when the agent cannot help. | Fallback & Launch |
| Entry point | How a conversation reaches the agent: a keyword from an ad, or an existing chat you assign. | Automation → Keywords or Chats |
| Review | Questions the agent could not answer, collected so you can add the missing answer. | AI Lab → Unanswered |
For an institute, a first set for the knowledge database looks like this:
Courses: each programme with who it is for, the prerequisites, the duration and the outcome you can truthfully claim.
Fees: the fee, what it includes, instalments if you offer them and the offers you have approved.
Timetable and demo slots: batch dates, timings, online or at the centre, and the slots you give for a trial.
Policies and FAQs: refund and transfer rules, documents needed and how admission works.
Instructions and boundaries are where an admissions agent gets its judgement. The screenshot shows both for an institute: ask the goal and level before recommending, never promise results, and hand discounts and scholarships to the centre head. An early-bird price printed on your fee sheet can sit in the knowledge database and be quoted as written; anything beyond it goes to a person.
Once the agent is live, AI Lab → Unanswered shows the questions it could not answer. Add the answer to the knowledge database, and the next parent gets it.
A fixed menu in front of the agent
Not every enquiry needs reasoning. Start with a short flow in the chatbot builder: a greeting and up to three reply buttons, such as Course info, Fees and batches and Free demo class. A tap gets an instant, fixed answer. Open questions, such as “which course is right for me?”, go to the AI Agent, which a keyword or a chat assignment starts.
What needs an integration
In the conversations above, every chat ends with a request to a counsellor, who confirms in the same chat. That works out of the box. Writing a confirmed demo into a calendar, or a lead into a CRM or a Google Sheet, goes through webhooks, Zapier or the API — see the plans for what each includes, and ask the WABB team what fits your tools. The agent can answer freely while the student’s last message is less than 24 hours old; after that, a follow-up — or a reminder the day before a demo class — has to be an approved template message. The student enrollment use case shows how the pieces fit for admissions.
Source: Meta for Developers — Pricing on the WhatsApp Business Platform, page updated 30 September 2026, checked 14 October 2026.
Business Impact: What You Are Aiming For
Results depend on your courses, your traffic and how well the agent is set up, so there are no percentages here. These are the changes an institute should look for:
Fewer lost enquiries
Every message gets an answer at once, including nights, weekends and the hour after an ad goes live.
Faster first reply
The student hears back while the interest is fresh, not after the morning team meeting.
Less retyping
Counsellors stop answering the same fee and timetable questions and spend their time on real conversations.
More qualified conversations
A counsellor opens a chat that already has the goal, level and timing, so the first call starts at the useful part.
A clearer view of demand
The list of unanswered questions shows what people ask that your materials do not yet cover.
How to tell it is working
Write down your own numbers for a normal month, then track four things every week:
Enquiries answered within a minute
Enquiries that reached a demo request or a counsellor
Unanswered questions per week, and how many you closed
Handovers where the counsellor did not need to ask the basics again
One agent answers first. Your team closes the admission.
FAQ
Common Questions
Can an AI Agent replace admissions counsellors?
No. It takes the first, repetitive part: the quick replies, the questions and the qualification. A person still confirms the demo, handles scholarships and exceptions, and closes the admission. See Four Use Cases: From the First Message to the Handover.
Can it recommend the right course?
Yes, when your course information is in its knowledge database. It asks about the goal and level, then matches them to the courses you have described. Without that information it should say so and hand over. See What an AI Agent Can Actually Do.
Can it book a demo class directly?
It can offer the slots you allow, collect the name and pass the request to your team in the same chat. Writing the booking into a calendar or a CRM needs an integration, so check what your setup supports. See How It Works in WABB.
What if a parent asks something that is not in the materials?
The fallback sends the chat to a person, and the question lands in the Unanswered list. Add the answer once and the agent has it next time. See How It Works in WABB.
Does this work for coaching and not only for schools?
Yes. The pattern is the same for tuition, exam preparation, online courses and personal coaching: answer fast, qualify, then move to a demo or a discovery call. The materials and the boundaries change, not the workflow. See Four Use Cases: From the First Message to the Handover.
Does it understand messages written in mixed languages?
Test it with the way your students actually write, including short, lower-case and mixed-language messages, before you launch. The WhatsApp AI Agent overview explains how to test.