School management,
spoken into existence.
Lumi is the first Conversational SMS for SEA preschools — teachers narrate the child's day by voice, principals run the school through conversation. Forms and menus recede; AI drafts, humans approve, the system executes.
- A · Voice-to-Daily-Report3
- · Today / Sunflower class01
- · Recording — Minh02
- · Parsed report draft03
- B · Chat-Driven Operations5
- · Admin home04
- · Refund · confirm-before-execute05
- · Form chase · bulk reminders06
- · Natural-language analytics07
- · Audit log08
Teacher · narrating the child's day
She just speaks. Lumi listens, extracts, drafts the report — she reviews and sends to parents via Zalo within 90 seconds.
Principal · running ops through conversation
Refunds, chasing forms, asking for numbers — one line of chat. AI builds the action, the audit trail captures it, owner approves in one tap.
- ●4 parents haven't paid May tuition (due May 5)
- ●7 parents haven't submitted the health-check form
- ●Ms. Huong is off Friday — Ms. Lien is covering
📎 SOURCES · 4 DATA TABLES VIEW
Risky calls & what I'd iterate next
Riskiest call: pushing the entire teacher daily log into voice. A preschool classroom is an extremely noisy environment — kids crying, toys clattering, regional accents. If Vietnamese ASR drops below 88% real-world accuracy, the "AI parsed" highlights become counterproductive — teachers spend more time correcting than typing from scratch, and trust in the product collapses in the first week of rollout.
Next iteration: (1) add a resume / append mode so the teacher can speak about each child separately — not forced into a 90-second monologue; (2) try splitting the refund screen into 2 steps (preview → tap-to-arm → tap-to-confirm) for amounts > 5M, borrowing from banking flows; (3) test the "expat parents read English" hypothesis for real — they may actually prefer bilingual with Vietnamese first.