CurioCity has two distinct value propositions — for two distinct audiences at two distinct moments in the sales motion. They are not competing. They are sequential. One closes the purchase. The other creates the advocacy.
India's home tutoring industry is valued at ₹34,000 crore and growing — not because parents want to pay for it, but because the school system leaves them no choice. 55% of Indian students attend private tuition. Urban families spend ₹9,950 per child per year on average, and by Class 7 many are running three tutors simultaneously. The tuition spiral is not a choice. It is the system's failure made visible.
The failure compounds at every level. The principal has no live intelligence on what is being learned across 800 students. The teacher is racing against a calendar she cannot win — 220 working days on paper, 145 in practice, with a textbook designed for 180. The student does homework nobody checks for understanding. The parent finds out their child is struggling when the exam result arrives.
No live dashboard of learning across classes and sections. No syllabus progress vs days remaining. Parent complaints arrive with no data to counter them. Teacher performance is invisible until exam results. Inspection readiness is compiled manually every time.
By November: 40% of the syllabus remains, 25% of the year remains. The hardest chapters get 3 periods instead of 8. She knows 30% of the class didn't understand Chapter 6 — but the syllabus won't wait. She moves on, carrying the weight of it.
A gap in tenses in Class 4 cascades into active/passive voice in Class 5, into reported speech in Class 6. Nobody identified it. The homework was copied — it looked complete. There was no safe space to say "I didn't understand."
No live window into their child's learning. The only signal is a filled notebook and "yes, I did my homework." The first real signal is the exam result. By then the gap is 3 months old. The tutor re-teaches the whole chapter — not the specific gap.
This is the ground truth that CurioCity is built against. Every feature traces back to at least one of these.
All lesson generation, concept tagging, and performance measurement draws from one of two sources — chosen once at school onboarding. Both map to the same internal concept taxonomy. Scores and metrics are always consistent across the product regardless of which mode a school uses.
Regardless of content mode, every generated concept maps to a canonical NCERT concept node internally. Confidence scores, syllabus progress, and Report Card data are always comparable — across classrooms, schools, and modes. The school teaches from their book. CurioCity measures against a shared standard underneath.
Three user types. Three different slices of the same data. Each tier sees what they need to act — and only that. Data visibility is a product decision, not a technical one.
Benchmarks are set manually by the teacher per topic, within school-wide defaults set by the principal. Parents never see the benchmark number — only whether their child is above, at, or below it, and by how much. This protects teacher autonomy, prevents parent gaming, and keeps accountability with the school where it belongs.
The teacher's primary interface. Browser-based, runs on her laptop or tablet. Designed around one principle: the teacher makes decisions, not content. Every screen offers a small, clear choice. The AI generates everything. She steers.
At setup, the system shows the honest maths: chapters remaining vs. actual teaching days available. Teacher assigns Deep Mode (8 periods, full depth) or Sprint Mode (3 periods, core concepts only) per chapter on a drag calendar. Syllabus progress recalculates live.
Solves: November death spiralEvery chapter has two AI-generated versions. Deep: full literature, grammar-in-context, writing activity, formative check. Sprint: core concept only, single embedded check, exam-critical content only. Teacher selects — AI generates the full flow.
Solves: Prep overloadOvernight quest data aggregates while the teacher sleeps. Before class she sees exactly: which concepts students struggled with, completion rate, and whether today's planned lesson needs a pivot. One screen. One decision.
Solves: Flying blind into classIf Pre-Flight shows >40% of students below benchmark on a concept, the system auto-suggests a 10-minute Catch-up Sprint. One tap generates the sprint content. Teacher approves. Class plan updates automatically.
Solves: Unprepared class fallbackA structured 3–4 phase lesson, auto-generated and timed. Each phase is manually selectable — teacher can swap any AI-generated phase with her own textbook exercise, personal notes, or a library activity. The old fused with the new. Detailed workflow design pending Burlington screen recording session.
In Design — Burlington Session PendingFor every concept taught, three pre-written re-explanation variants are ready: visual, story-based, and logical. When the class doesn't understand, teacher taps the variant — no scrambling, no improvising under pressure.
Solves: Repetitive explanation panicA single class-wide score per topic, benchmarked against the teacher-set threshold. Displays a bimodal variance warning when the class is split — not just an average. A "⚠ High variance" flag appears when distribution is dangerously wide.
Solves: Data overload + hidden splitsSystem flags concepts due for revision based on student gap data. Teacher retains full manual control — she triggers and schedules spaced repetition sessions herself. The AI surfaces the signal. She makes the call.
Solves: Revision = re-readingAfter micro-writing tasks, teacher sees a class-wide writing pattern summary — not 40 individual compositions. Common errors, strongest responses, suggested focus for next session. 3 hours of grading collapsed to 3 minutes of reading.
Solves: Weekend correction loadEvery April, the incoming teacher receives a class gap profile — which concepts which students are weak on, based on the previous year's quest data. She walks into her first class already knowing the room. The 6-week figuring-out period is eliminated.
Solves: New year blindnessA single interface. Two tabs. The primary job is the quest — the child learns here, on the parent's phone. The Report Card tab is always accessible to the parent. No separate parent app. No child device required. Simple by design.
The student's primary experience. Gamified, milestone-framed, impossible to copy.
Each chapter is delivered as a fixed 7-stage quest. Stages progress from comprehension → application → creation. Student cannot skip ahead. Each stage unlocks the next. The homework loop is gamified by structure, not gimmick. Detailed stage design pending worksheet format mapping session.
In Design — Worksheet Session PendingSeparate optimised quest flows per subject. English: passage + grammar-in-context + micro-writing. Maths: scaffolded problems with hint progression. Science: concept application + diagram interpretation. Not one generic quiz skin across all subjects.
Solves: Grammar-literature disconnectNo scores visible to the student. No percentages. Progress is milestones: "4 of 7 stages complete." Deficits appear as "2 milestones remaining" — never as failure. Low confidence is always a direction, never a verdict.
Solves: Weak student anxietyEach practice problem has three embedded layers: hint → partial solve → full solve. Student attempts first. Copying is eliminated — each layer requires an active response before unlocking. The student cannot fake their way through.
Solves: Homework copying ecosystem3–5 sentence writing tasks embedded within the quest. AI provides instant, age-calibrated feedback — not marks, but specific guidance. Writing skills are no longer the first casualty of time pressure. The teacher sees a class-wide pattern summary, not 40 compositions.
Solves: Writing skills droppedWhen a student falls below benchmark on a specific concept, the system generates a targeted micro-quest on that concept alone — not the whole chapter again. This is the tutor replacement. Surgical, not repetitive. The gap identified tonight, closed tonight.
Solves: Tuition duplicationThe parent's view. Always accessible. Progress-framed, never deficit-framed. Benchmark numbers are confidential to the school — parents see relative distance only.
Per-topic performance shown as relative distance from benchmark: ↑ 12% above, → at benchmark, ↓ 8% below. No raw scores. No benchmark numbers. Just direction and distance. Every parent, regardless of education level, understands this instantly.
Solves: Report card shock + no live windowA single subject-level score aggregating all topic confidence scores. Trend direction shown week-on-week with a simple arrow. Parent sees whether English is improving, stagnant, or declining — without needing to interpret raw numbers.
Solves: No early warning systemWhen any topic shows below-benchmark, one action appears immediately: "A quest is ready for [child's name] on [topic]. Tap to start." No advice. No alarm. No explanation required. Parent hands phone to child. Learning begins.
Solves: Tutor trust problem + tuition spiralAfter the final quest stage, a glanceable summary appears for the parent: today's milestone progress, subject direction this week, one next step. No raw scores. Three elements. Designed to be read in 10 seconds before the parent takes the phone back.
Solves: Homework supervision anxietyThe Live Report Card is the primary USP of CurioCity. Every student. Every topic. Every subject. Visible to teachers, principal, and parents — each seeing the right slice at the right resolution. Real-time, not post-exam.
Class vs class confidence by subject. Weakest topics across the school. Syllabus progress per class vs days remaining. All in real time.
Per-student milestone progress. Topic confidence with bimodal variance flag. Pre-Flight readiness before each class. Benchmark set by teacher.
Relative distance from benchmark per topic. Subject aggregate trend. One-tap quest when below benchmark. No raw scores. No benchmark numbers.
Report Card data draws from student quest completion — the digitised equivalent of today's physical worksheets and homework. Exact data sources to be finalised after mapping current CBSE homework and worksheet formats in a dedicated session. Exit ticket data from the Co-Pilot flow will also feed the Report Card once the 60-min workflow is defined.
The two interfaces connect through a continuous overnight data loop. The student's quest at night becomes the teacher's intelligence in the morning. The teacher's lesson becomes the student's quest that evening. The loop closes every 24 hours.
Teacher closes today's class on the Co-Pilot dashboard. System auto-generates tonight's student quest based on what was taught and what the exit ticket revealed. Teacher approves with one tap. Class join code makes quest accessible to all linked parent accounts immediately.
Student works through the 7-stage quest on parent's phone. Responses, milestone completions, micro-writing, and stuck points are captured per student. End-of-quest summary appears for the parent. Data submitted to server. Report Card updates in real time.
All student responses aggregated. Topic Confidence Scores calculated. Bimodal variance detected. Pivot Protocol trigger evaluated against teacher-set benchmark. Pre-Flight dashboard built and cached. Tomorrow's lesson plan updated if pivot is needed. Lesson package downloaded to teacher device via any available WiFi.
Teacher opens dashboard. Sees class readiness at a glance: Confidence Score per concept, completion rate, Pivot Protocol suggestion if triggered. Makes one decision — run today's planned lesson or trigger the catch-up sprint. Lesson runs offline from cached local copy.
Phased lesson runs from cached local copy. No internet required. Teacher selects and customises phases manually as needed. Exit ticket in final minutes. Responses captured. Report Card updates. Syllabus progress advances. Loop restarts from Evening step.
The student interface is the input layer. The teacher interface is the intelligence layer. They never interact directly — only through the overnight aggregation layer. This separation keeps the classroom experience clean and the data loop reliable.
Every major edtech product in India has solved either the student's in-the-moment doubt — or the teacher's content creation problem. Nobody has solved the teacher's pacing problem, the principal's visibility problem, or the parent's early warning problem simultaneously.
A comprehensive competitive analysis — covering Classklap, Embibe, Burlington, BYJU'S, Khan Academy, Doubtnut, and the LMS landscape — requires dedicated research and is deferred to a separate session. This section to be completed before any external pitch or investor conversation.
SaaS subscription per school, priced per student enrolled annually. The school is both the buyer and the distribution channel — the principal purchases CurioCity and presents it to parents as part of the school's value proposition. The school sells the tutor-elimination promise. CurioCity delivers it.
CurioCity collects learning data on minors. India's Digital Personal Data Protection Act (DPDP) 2023 is in force. CurioCity is designed with student data privacy compliance built in from architecture upward.
The full vision above is the destination. What ships first is chosen to validate the core loop — that a teacher can deliver a better class because of CurioCity, and that a principal can see the result. Everything else is Phase 2.
This is the teacher's daily touchpoint. If the Co-Pilot is useful, she uses CurioCity every single day. If it isn't, nothing else matters. Shipping this first validates the core teacher behaviour change — from improvised delivery to structured, AI-assisted class flow. Design pending Burlington screen recording session.
The principal is the buyer. If she can see live class performance and syllabus progress on day one, the sales conversation closes. The Principal Dashboard is built on top of the Co-Pilot's output data — it cannot ship before Feature 01 generates meaningful data. Exact metrics TBD.
The tutor-elimination promise and the parent-facing Report Card are the long-term moat. They require quest completion data flowing from MVP Feature 01 before they are meaningful. Ship once the teacher loop is validated and the data layer is live.
Feature 01 changes teacher behaviour. Feature 02 closes the sales conversation with the principal. The Parent Interface turns school customers into parent advocates. Each phase creates the conditions for the next. Shipping in any other order breaks the dependency chain.
| Decision | Resolution | Status |
|---|---|---|
| Product Name | CurioCity | LOCKED |
| Hero Statement | Eliminate Home Tuitions with the Help of AI | LOCKED |
| Product Tagline | Your child's AI companion for learning, development and live tracking | LOCKED |
| Institutional Tagline | Stay Informed, Stay Intelligent | LOCKED |
| Primary USP | Live Personalised Report Card + Tutor Elimination | LOCKED |
| Subject Scope | All subjects — English, Maths, Science, Social Studies, Hindi, others | LOCKED |
| Target Segment | CBSE Class 4–8, all Tiers | LOCKED |
| Interfaces | Two only — Teacher Dashboard + Parent/Student Interface | LOCKED |
| Student Device | None — student uses parent's phone for quests at home | LOCKED |
| In-Class Hardware | Teacher device + projector only. Zero student devices in class. | LOCKED |
| Content Modes | NCERT Mode (LLM/cached) + Local RAG Mode (SLM). Both map to same concept node layer. | LOCKED |
| Benchmark Setting | Teacher sets per topic, within school defaults. Principal sets school-wide starting defaults. | LOCKED |
| Benchmark Visibility | Relative distance only to parents (↑↓%). Benchmark number confidential to school. | LOCKED |
| Below-Benchmark Action | One-tap targeted quest. No advice, no alarm. Immediate action only. | LOCKED |
| Student Onboarding | Parent self-onboarding via class join code. Parent owns child's profile. | LOCKED |
| School Admin Layer | Principal — account management + school-wide defaults + performance view. TBD: teacher-level metrics visibility. | LOCKED |
| Parent Interface | Two tabs — Learn (quest) + Report Card. No separate parent app. | LOCKED |
| Business Model | SaaS per school, priced per student enrolled annually | LOCKED |
| Data Privacy | DPDP 2023 compliant. Data within school account. No third-party sharing. | LOCKED |
| 60-Min Co-Pilot Workflow | 3–4 manually selectable phases. Full design pending Burlington screen recording. | IN DESIGN |
| Report Card Data Source | Quest completion + exit tickets. Exact format pending CBSE worksheet mapping session. | IN DESIGN |
| Principal Dashboard Metrics | Class and school-wide performance. Teacher-level visibility TBD — policy decision pending. | TBD |
| Competitive Landscape | Dedicated research session required before any external pitch. | TBD |