A training center goes digital, part 2: why build the info and enrollment entry point first, not a learner system or AI

This series follows one training center through several phases of going digital. The journey is built from real implementation experience in this industry — the stages and challenges should feel familiar if you run this kind of school. Each part covers what was built, what was deliberately not built, and what evidence justified moving to the next phase.

Let’s say one honest thing up front

The owner of this center knew exactly what the long term held: a portal students could log into with their schedule, homework and grades — and, further down the road, AI that helps write test questions and answers students’ questions online, freeing admin and teachers from repetitive work.

Those needs existed from day one, they did not appear later. So why did phase one not build the portal, and not plug in AI?

Because “the need exists” and “it can be built correctly right now” are two different things. Installing a learner portal and AI into an unprepared school produces:

• Enrollment data still scattered. However polished the portal, it would be fed with enrollment details scraped from chats in fragments. Garbage in, garbage out.
• Nobody knows what the system should manage. Schedule fields, who updates grades, how transfers and refunds work — the school had not even written these rules down, so any system design would miss the real exceptions.
• Staff get pushed too far at once. Adapting to a new portal while facing AI-generated content and deciding whether to trust it — most people would bypass the system and keep working in LINE.
• AI has nothing to lean on. AI test generation should match “this school’s style,” and Q&A should come back to “this school’s materials.” Without a unified learner entry point and structured data, AI is just a text generator with no context.

So why build the entry point and enrollment convergence first

Building the entry point first is not a smaller goal — it is the correct order:

1. The entry point lets strangers understand first. Enquiries drop, and admin and teachers are freed to do what actually needs a human — instead of answering the same question twenty times.
2. Convergence cleans the data first. Enrollment moves from scattered chat descriptions to structured enrollment records. When the data is clean, any later system has usable input.
3. Data shows you the truth. Where enquiries come from, how many enroll, which course actually fills. These numbers can finally answer the owner: “which system should I build next?”
4. People start adapting from this step. Admin gets used to unified replies and following up enrollments; teachers see the pages deflect repeat questions for them. Readiness grows in small steps — it is not installed along with a system.

What belongs to the entry point, and what belongs to a system

Be clear, to avoid promising everything at once:

• Entry point and data phase: course info pages, an enrollment and enquiry form, and enrollment records landing in one searchable place.
• Independent systems: a learner login portal, online testing and grading, a grades system, a full student management system (CRM), online payment.
• Later still, plugged in over an API: AI test generation and online Q&A. They only come after the portal is stable, the data is reliable, and teachers are willing to review the output.

Over an API means one program calls another program’s capability through a defined interface. In this story the AI is not desktop software installed on a teacher’s computer — it is a capability the center’s portal plugs in over an API, and students use it inside the portal.

What to watch before moving to phase three

Before agreeing to build the next thing, the center promised only to watch these signals:

• Whether hotline-style LINE enquiries actually dropped after the unified course pages went live;
• Whether enrollment records reliably and completely land in one place;
• Whether admin is willing to use this converged flow instead of going back to handling each chat individually.

If none of these hold, entry point and convergence are not done — and building further on top is just building on sand.

Part 3 looks at how this center converged the enquiries scattered across LINE and Facebook into a genuinely trackable enrollment data trail.

---

The boundary matters: the school website, course info pages and an enrollment/enquiry form are a normal website project with a configured data-delivery channel. The learner login portal, online testing, CRM, online payment, and AI test generation and online Q&A plugged in over an API are separate system projects with their own scope — they are not part of a standard website package.