Kelas Sekejap — Research synthesis (first 3 SPM interviews)

Student	Date	App tested?	Persona
Adif SMK Putrajaya P14 · Form 5	1 Jun	Full · 45 min	High-engagement
Dania SMK Putrajaya 5 Perdana · Form 5	2 Jun	Mic broken	Lower-fluency
Afira SMK Putrajaya 5 Perdana · Form 5	3 Jun	Full · 30 min	High-engagement
Dhia SMK Putrajaya 5 Perdana · Form 5	5 Jun	Full · 30 min	High-engagement

Findings

What shows up across the interviews

The AI tutor solves a real classroom problem, not a convenience one.

The strongest product-market-fit signal across the four sessions — and it has two distinct angles. Adif and Afira valued the app as a speaking partner; Dhia valued it as a self-assessment mirror that gives her clarity feedback her teachers don't.

My speaking partner — she doesn't like to talk to me about English stuff because she's also struggling with it. So Roly — it's like speaking to me. I like that.Afira, Form 5

I feel satisfied because I actually don't know how clear my voice is when I speak with my teacher. So this app helped me to tell me about how bad or how good my thing is.Dhia, Form 5 (scored 15/30 on her real speaking paper)

The contrast with Dania makes it sharper still: she doesn't even have a regular practice partner — she "sometimes" does speaking with her teacher in class. Same shape of need, more acute.

Implication · The value prop is plural. "Practise with an AI partner" (Adif/Afira) and "Hear yourself the way an examiner does" (Dhia) are different pitches for different students. The app already does both — worth making both legible in onboarding and marketing.

Other apps exist in this space — but none doing what we're doing.

Two of the three students named an educational app they actively use. Neither overlaps with what Kelas Sekejap is doing — one skips English entirely, the other handles only grammar quizzes in an Instagram feed.

App	Mentioned by	Covers	Format
Siap Study	Adif	Bahasa Melayu & History (Form 4/5 only)	Full study app — no English module
My English Grammar	Dania	English grammar only	Instagram account, MCQ posts in feed

Adif's unprompted comparison of the prototype to Siap Study — the only direct head-to-head in the three sessions:

One of the best design on our educational app. The layout is simple, clean and easy to use. And this app does not contain any pop-up advertisement.Adif

Dania didn't get to compare (mic failed before she could test the prototype), but the app she does use tells us something important on its own: she found My English Grammar via her Instagram feed, not the Play Store. The bar for casual daily English engagement is being set by free, scrollable, grammar-MCQ content — not by anything that looks like an exam-prep app.

Positioning · The competitive moat isn't a feature — it's a category. No Malaysian app currently does speaking practice with AI tutors for SPM. The named alternatives either skip English (Siap Study) or stop at grammar (My English Grammar). The pitch isn't "a better study app" — it's "the only place to actually practise speaking." Worth landing in marketing copy and in onboarding.

Students want more scaffolding, not less.

A surprising and consistent ask across both engaged sessions.

On part three, you guys can give a template — a template answer to student to use so that the student doesn't clueless what to answer it.Adif

Roly chose two and I only had to choose one. So it's very easy for me to elaborate. It lessens my decision making.Afira, on Part 3

Adif also described how Cikgu Natasha actually preps her class: "We just remember the template, not the topic, because the topic is actually different." Students don't memorise day-specific answers — they memorise structural skeletons they apply on the day.

Implication for the 30-day backlog · Every Part 3 day should include an explicit template — a 4–5-line skeleton (open → reason → invite partner → respond → decide). The model arguments fill the skeleton; the skeleton is what transfers.

Part 1 is more flexible than our spec assumes.

Adif described Part 1 as "Q3–5 are random" — openers fixed (name, where from) plus a few rotating fillers. Afira said the real exam is just name + 1 question. Both said the longer app version (currently 6 questions) is helpful for revision but doesn't mirror the real test.

Resolution · Default mode asks 3–5 questions with rotation (matches Adif's reality). Add an exam-mode toggle that strips down to 1 for the days right before the trial / real SPM.

The persona split is about anxiety profile, not fluency level.

After three sessions I'd drawn a clean line: Adif and Afira are one persona, Dania is another. Dhia's session breaks that read. Same school, same class as Dania. Self-described her English as "broken". Scored 15/30 on her real speaking paper. And still finished the prototype — Part 1, Part 2, Part 3, 5/5 on the linking-words exercise. The fluency floor for engagement is lower than the Dania session suggested.

The right way to split these students isn't how good their English is — it's what they're afraid of on the exam. Three out of four name the same fear; one names different ones.

Persona A · Speaking-anxious (3 of 4)

Adif · Afira · Dhia

Anxious about Part 3 speaking (and for Dhia, Part 4 listening too)
Self-described fluency spans B2 (Adif) down to "broken English" (Dhia)
All three engaged with the prototype and gave actionable feedback
Strong in writing and reading by their own account
Each one explicitly said they'd use the speaking module daily

Persona B · Writing-anxious (1 of 4)

Dania

Most nervous about essay writing (Paper 2) — not speaking
Struggles to generate ideas on a topic; common topic: food
Vocabulary comprehension the other key barrier — "I don't understand the words"
Couldn't access the speaking prototype — but the cause was mic failure + no session prep, not fluency alone
Already uses an Instagram English account daily; discovers tools via her feed

Strategic implication · The speaking prototype probably serves a wider slice of the SPM cohort than I'd argued three sessions ago. Dhia is the proof that lower-fluency students engage with it when device + onboarding aren't blockers. The Word-of-the-Day skill is still relevant for Persona B (and useful as supplemental daily practice for Persona A), but it isn't a second product line — the speaking module is the front door for more of the audience than originally scoped.

The character-led UI is doing real emotional work.

All four students made some version of this comment, unprompted:

It's so cute … reminds me of Shopee.Afira

One of the best design on our educational app. The layout is simple, clean and easy to use. And this app does not contain any pop-up advertisement.Adif

The photo is cute, maybe.Dania, in a session where she barely commented on anything

They are so cute and adorable. So I more semangat [when I see them].Dhia (semangat = enthusiastic, motivated)

Four unprompted positive responses across four different self-described fluency levels. Dhia's semangat framing is the most useful — character design isn't just aesthetic, it's measurably moving motivation.

Implication · Keep the warm-orange palette and character-led UI. They're paying off. And keep the no-ads stance Adif flagged — for him it was a trust signal, not just a hygiene one.

Dhia's account is the clearest read we have on the Listening paper.

Across the four sessions, only Dhia walked through the structure of Paper 4 (Listening) in detail. She's anxious about it (alongside speaking — her two weakest papers) and described all four parts from memory:

Part	Format	Notes
Part 1	Objective, ~8 questions	Answered one at a time. Short audio per question.
Part 2	Objective, ~8 questions	One longer audio covering all 8 questions.
Part 3	Matching	5 speakers · 7 answer options. Match each speaker to an answer.
Part 4	Fill-in-blanks	Paragraph with blanks; audio plays through once; single-word answers in blanks.

[Part four] is pretty hard, actually.Dhia

Worth noting Afira also flagged the Listening paper as painful, but for a different reason — "the audio in our hall echoes through the room." The hall acoustics in some Putrajaya schools genuinely make Part 4 harder than it has to be. The prototype's "Hear what good sounds like" model-answer audio is clean by comparison, which both Afira and Dhia noticed.

Implication · Don't scope a Listening module from this alone — but it's the closest thing to a verified Paper 4 spec we have from the field. Confirm with Cikgu Natasha before building. The clean-audio side-benefit of the existing Part 2 model answers is a feature worth marketing in its own right: "Practise listening to clear English before exam day."

Seven things we learned from four SPM students.

Who we spoke to