# LingoTask — Deep Research Brief **Date:** 20 June 2026 **Source:** LingoTask website, EdCity eLAFP, Learning & Teaching Expo 2025, HKET Education (信報教育), saasbrowser.com, CUHK faculty pages --- ## 1. Overview **LingoTask** is an AI-powered **全科 (all-subject) smart learning and teaching platform** purpose-built for Hong Kong primary and secondary schools. It covers **Chinese, English, Maths, and History**, with a flagship focus on automated essay marking and speaking assessment aligned to DSE/TSA/HKAT/IELTS standards. | Key Data | | |---|---| | **Serving** | 190+ Hong Kong schools | | **Founded** | 2021 (pre-ChatGPT era) | | **Backed by** | CUHK (Chinese University of HK) + SpeechX Limited | | **Pricing** | HK$29,500–49,500/year per school | | **Grant alignment** | EDB「智啟學教」HK$500K per-school AI funding | --- ## 2. The Team — Academic Heavyweights This isn't some fly-by-night startup. The founding team comes straight from CUHK's engineering faculty with serious credentials: | Name | Role | Background | |---|---|---| | **Prof. Helen MENG** | Chief Scientist, SpeechX | MIT (B.S., M.S., Ph.D.), Chair Professor @ CUHK | | **Dr. Bruce LI (李坤)** | CEO & Co-founder, SpeechX | Zhejiang Univ (B.S.), Tsinghua (M.S.), CUHK (Ph.D.) | | **Prof. Xixin WU** | Assistant Professor, CUHK | Tsinghua (M.S.), CUHK (Ph.D.) | Dr. Bruce Li's previous startup attempt (circa 2016) was a "full AI teacher" concept that raised millions but burned through funding without achieving product-market fit. That failure taught them the lesson that shaped LingoTask: **AI should augment teachers, not replace them**. --- ## 3. What It Actually Does ### Core Modules | Module | Capabilities | |---|---| | **Writing Marking** | 1 min per essay, 15 min for 100 essays. Handwriting OCR at 99% accuracy (English, Simplified & Traditional Chinese). DSE/TSA/HKAT/IELTS rubric-aligned. 90% agreement with human markers. | | **Speaking Assessment** | 99% speech recognition accuracy. Handles Hong Kong-accented English, Cantonese, Mandarin. Covers朗读, 看圖說話, 個人短講, one-on-one dialogue, group discussion. | | **Reading & Listening** | One-click conversion of PDFs, web pages, YouTube videos, MP3s into interactive quizzes. Auto-grading + diagnostic reports. | | **Maths** | Handwritten OCR step-by-step diagnosis | | **History** | Data-based question logic analysis | | **Ask AI** | Integrated with GPT / Gemini / DeepSeek — unlimited usage across all plans | ### Key Innovation: Focused Marking Instead of hitting students with all errors at once, LingoTask does **targeted correction by ability level**: - Beginner students get simplified feedback (e.g. "the -ing form") - Advanced students get technical labels (e.g. "gerund") This came directly from teacher feedback — overwhelming weak students with every mistake kills motivation. ### Personalised Model Essays The system preserves the student's original context and generates a custom "high-score version" of their own essay. Students see *how their own writing* could improve, not a generic example. --- ## 4. Pricing Breakdown | Plan | Annual Fee (HKD) | Subjects | Teacher Accts | Student Accts | |---|---|---|---|---| | **A** | 29,500 | English only | 12 | 3 grades or 12 groups | | **B** | 49,500 | English + Chinese | 25 | Whole school | | **C** | 49,500 | English (speaking focus) | 12 | Whole school | | **D** | 49,500 | English + Chinese (speaking focus) | 12 | Whole school | - Semi-annual: ~HK$19K–32K - Free trial: 2 weeks, up to 2 teachers + 50 students - **All plans include unlimited** writing marking, reading exercises, listening exercises, and Ask AI (GPT/Gemini/DeepSeek) - Speaking assessment is only in Plans B/C/D ### EDB「智啟學教」HK$500K Grant Alignment This is the killer go-to-market move. The EDB (Education Bureau) offers a **HK$500,000 per-school grant** for AI learning/teaching under the「智啟學教」scheme. LingoTask positions itself as the ready-made compliant solution: - Provides full compliant quotations - Drafts proposal documents - Handles procurement procedures - Other applicable grants also listed on their site For a school, the effective cost under this grant is **zero** — the government foots the HK$29K–49.5K bill. --- ## 5. Origin Story & Key Lessons ### The Spark (2021) At a dinner party, an English teacher friend complained about marking essays — spending nights and weekends on repetitive corrections. He transferred to a primary school because "primary essays are shorter, I can mark faster." His exact words: *"If only AI could help mark essays!"* ### The First Failure (2016) Before LingoTask, the team raised millions to build a "full AI teacher" — voice recognition, dialogue systems, virtual digital human teachers. They ran out of money. The lesson: **Product-Market Fit matters more than technology**. ### The Hard Lessons in Product Development 1. **99% OCR or bust** — They thought 90% handwriting accuracy was fine. Then they watched students in class: a 300-word essay with 10% error rate = 30 words to manually fix. Students gave up. Teachers saw it as "creating more work." 90% = failure in education. **99% was the minimum viable**. 2. **Speed is non-negotiable** — If feedback takes a week, it's useless. Students in 2026 expect instant responses. The system is optimised for sub-minute marking even at scale. 3. **Actionable > Accurate** — Early versions copied Grammarly's approach: flag errors, give the right answer. Teachers pushed back. The key is *guiding* students to self-correct, not handing them the answer. The system now gives progressive hints and recommends remedial exercises. 4. **No user manual** — Their design principle: "No manual, no training needed." They chose native app + web (higher dev cost) specifically for smooth camera upload, voice recording, and offline cache. 5. **Robustness is invisible until it breaks** — At 130 schools, a DDoS attack took them down for 30 minutes. Teachers called within 5 minutes. They hardened defences. A second attack a week later was absorbed without impact. ### The Pivot to Chinese At LTE 2025 (Learning & Teaching Expo), teachers kept asking: "Can it do Chinese?" The Chinese module was already in development. The expo feedback gave them the classroom context to prioritise and align it properly. Result: a full Chinese learning suite for both primary and secondary within months. --- ## 6. Competitive Landscape ### Direct Competitors | Platform | Focus | Key Difference vs LingoTask | |---|---|---| | **EC Marking (改卷易)** | DSE Chinese/English writing marking | Narrower scope — writing only. No speaking, no multi-subject, no grant alignment. | | **Traditional tutoring centres** | In-person | Labour-intensive, no AI efficiency | | **Grammarly / ProWritingAid** | General English | Not HK-curriculum aligned. No DSE/TSA rubrics. No Chinese. | ### LingoTask's Moat 1. **190+ school adoption** — Network effects. Teachers share workflows. Schools benchmark each other. 2. **CUHK academic credibility** — Parents and school boards trust university-backed products. 3. **EDB grant alignment** — Zero cost to schools (effectively). Hard to compete with "free." 4. **Deep HK curriculum integration** — DSE/TSA/HKAT/IELTS rubrics baked into the AI models. A general AI writing tool can't do this without years of training data. 5. **Teacher co-creation flywheel** — 100+ workshops per year. Every session generates feedback that improves the product. Competitors without school access can't replicate this. 6. **Multi-subject expansion** — Started with English, now covers Chinese, Maths, History. Each new subject deepens the dependency. --- ## 7. Relevance to Simply English ### Could LingoTask Be Used at Simply English? **Short answer:** Not directly. LingoTask is designed for **schools** (whole-school licensing) and doesn't seem to offer individual tutoring centre plans. The smallest plan (Plan A: HK$29,500/year for 12 teachers) is priced for institutions, not small learning centres. **However — the *model* is relevant:** | LingoTask's Approach | Application to Simply English | |---|---| | AI marking reduces teacher workload by 70%+ | Your satellite model could integrate AI pre-marking → teacher final review → higher throughput per teacher | | Focused marking by level (targeted feedback) | Your differentiated class levels could use AI to give personalised writing corrections without burning teacher time | | Personalised model essays | Students at Simply English could get instant "this is your essay improved" examples — huge value add | | Instant speaking assessment | DSE speaking practice with AI feedback is a **direct sellable product** | | One-click worksheet generation from any source | Prep time drops from hours to minutes — lets your teachers focus on teaching | ### Strategic Angle — Partnership or White-Label? Given: - LingoTask has CUHK/SpeechX backing but no visible learning centre track - Simply English has 3 centres + growing satellite model - DSE prep is one of Simply English's growth areas There might be a **partnership opportunity**: Simply English becomes a pilot/promotion partner for LingoTask in the learning centre space. You get discounted access; they get a new channel. ### Competitive Watch If Simply English competes with: - **Monkey Tree / Headstart / other HK tutoring chains** — those centres could adopt LingoTask and gain a significant efficiency edge over you - **British Council** — already has resources; less likely to need it - **Independent tutors** — LingoTask makes them more competitive too **Implication:** The AI-marking wave is coming to HK education. If your competitors adopt it and you don't, you'll be at a disadvantage on both cost and quality. Worth evaluating an AI marking solution — if not LingoTask, then a lighter-weight alternative — as part of your satellite expansion. --- ## 8. Verdict | Dimension | Assessment | |---|---| | **Team credibility** | 🟢 Excellent — CUHK professors + experienced CEO | | **Product quality** | 🟢 Strong — 99% OCR, 90% marker agreement, native app | | **Market traction** | 🟢 Strong — 190+ schools, EDB grant-aligned | | **Pricing for small centres** | 🔴 Weak — designed for schools (HK$29K–49.5K/year) | | **Relevance to Simply English** | 🟡 Medium — model is instructive but direct use unlikely | | **Threat if adopted by competitors** | 🟡 Moderate — gives adopting centres significant efficiency gains | **Bottom line:** LingoTask is a well-funded, academically-backed, school-focused AI marking platform with strong traction. It's not a direct fit for Simply English at current pricing, but its *model* is highly relevant — and it signals that AI-powered marking is becoming a competitive necessity in HK education. Worth monitoring as a potential partnership target or as a benchmark for what an AI-augmented learning centre looks like.