This document is a companion to the main Guruthuma stakeholder brief. The pricing tiers and headline Year 1 numbers there stay unchanged — what's new here is a from-the-ground-up cost verification for each use case, with a deliberate 2× buffer applied throughout, so leadership can see both how conservative our original budget was and how much margin headroom that leaves.
01
Use Case 1
After-class doubt solving
Students message their question on WhatsApp or Telegram at any hour. Without Guruthuma, a teacher either answers every message personally — eating into evenings and weekends — or simply doesn't, and the student is left stuck until the next class.
How Missaka Isuru solves it
Missaka Isuru runs in a fast, lightweight configuration tuned for quick, simple question-and-answer exchanges. It's given the teacher's own subject area and syllabus up front, so every answer stays grounded in what that teacher actually teaches, in natural, exam-appropriate Sinhala rather than a stiff translation.
What the teacher actually gets
24/7 WhatsApp or Telegram doubt-solving for their own students
Answers grounded in the teacher's own subject, syllabus, and way of explaining things
Natural, exam-register Sinhala — not machine-translated English
Automatic handling of repeat questions, so the teacher isn't pinged for the same thing twice
Capacity for 10,000+ questions per teacher per month
Covered under a flat monthly subscription — no per-question billing surprises
Our cost to deliver
LKR 0.60–1.20
Original brief baseline / question
≈ LKR 0.04
Verified engine cost, pre-buffer
≈ LKR 0.08
Cost we plan around (2× buffer)
Even at this doubled figure, a Starter teacher using their full 10,000-question monthly capacity would cost us about LKR 800 to serve — under a quarter of the LKR 3,500 Starter subscription price, with room to spare.
02
Use Case 2
Practice paper & worksheet generation
Writing a fresh, syllabus-aligned model paper every week, by hand, in Sinhala, is one of the most time-consuming parts of a tutor's job — and it's the first thing that gets skipped when a teacher is short on time.
How Missaka Isuru solves it
For paper and worksheet generation, Missaka Isuru switches to a more careful, higher-capability configuration that prioritises structure and syllabus accuracy over raw speed — appropriate for content a student will study from directly, not just a quick chat reply.
What the teacher actually gets
On-demand model papers and worksheets from a topic description, aligned to O/L or A/L syllabi
Sinhala-medium formatting matching real exam paper conventions
Editable output the teacher can adjust before sharing with students
Unlimited generation volume that scales with demand
Included in the subscription, or available pay-per-use depending on tier
Our cost to deliver
LKR 8–15
Original brief baseline / paper
≈ LKR 0.53
Verified engine cost, pre-buffer
≈ LKR 1.06
Cost we plan around (2× buffer)
Even a teacher generating several papers a day stays well under a single rupee of cost per paper after the buffer — a rounding error against any tier's subscription price.
03
Use Case 3
Hall-class & institute assistant
A large institute or hall class with thousands of students can't run on one teacher's phone. Routine questions — schedules, payments, "what did I miss" — pile up fast, and answering them all by hand means hiring admin staff just to keep up.
How Missaka Isuru solves it
The institute assistant runs the same fast, lightweight configuration as after-class doubt solving, just deployed at much higher concurrency, so it can hold thousands of simultaneous student conversations under one branded interface without slowing down.
What the institute actually gets
One branded chatbot covering the entire institute or hall class, under the institute's own name
Handles class schedules, payment reminders, and model-answer distribution automatically
Answers routine student queries without adding admin headcount
Usage analytics dashboard for the institute owner
Capacity for tens of thousands of interactions per month
One flat monthly fee, regardless of exact interaction volume within plan limits
Our cost to deliver
LKR 3,000–5,000
Original brief baseline / institute / mo
≈ LKR 209
Verified engine cost, pre-buffer
≈ LKR 417
Cost we plan around (2× buffer)
Against the LKR 35,000 Institute tier price, a buffered cost of LKR 417 a month implies a margin on this use case alone of roughly 99% — well above the figure used in the main brief, which was set conservatively before this verification.
04
Use Case 4
Auto-grading & feedback
Marking short-answer scripts one by one, subject by subject, is hours of a teacher's week that produces no teaching at all — just red ink. The backlog grows fastest right before exams, exactly when students need feedback the most.
How Missaka Isuru solves it
Auto-grading uses Missaka Isuru's document- and image-reading capability, paired with the same higher-capability configuration used for paper generation, so it can read handwritten or photographed scripts accurately and produce fair, specific feedback rather than generic comments.
What the teacher actually gets
Upload or photograph short-answer scripts for automatic review
Per-script marks plus written Sinhala feedback for the student
Handles handwritten answers and imperfect smartphone photo quality
Capacity for hundreds of scripts per teacher per day
Available as an add-on module on top of any core tier
Our cost to deliver
LKR 3–6
Original brief baseline / script
≈ LKR 0.11
Verified engine cost, pre-buffer
≈ LKR 0.22
Cost we plan around (2× buffer)
A teacher grading 300 scripts a day, every school day of the month, would still cost us under LKR 2,000 a month at the buffered rate — comfortably inside add-on pricing headroom.
05
Year 1 cost reconciliation
Rolling the four use cases above up to Year 1 volumes, here's how the verified, doubled engine cost compares to what was originally budgeted in the main stakeholder brief. Only the Missaka Isuru service-cost line changes — hosting & infrastructure cost is unaffected and carries over unchanged.
LKR 686K
Originally budgeted engine cost
LKR 329K
Verified, buffered engine cost
-52%
Under original budget, buffer included
~88.9%
Potential blended margin (was 85.6%)
| Line item | Original Year 1 plan | Verified + buffered | Difference |
|---|
| Missaka Isuru service cost | LKR 686,000 ($2,036) | LKR 329,000 ($976) | -LKR 357,000 (-52%) |
| Cloud & hosting infrastructure | LKR 857,000 ($2,542) | LKR 857,000 (unchanged) | — |
| Total cost to deliver | LKR 1,543,000 ($4,578) | LKR 1,186,000 ($3,521) | -LKR 357,000 |
| Gross profit (on LKR 10.71M revenue) | LKR 9,167,000 | LKR 9,524,000 | +LKR 357,000 |
| Blended gross margin | 85.6% | ~88.9% | +3.3 pts |
USD figures use an indicative planning rate of LKR 337 = $1. "Verified + buffered" reflects a fresh engineering cost check on each use case, with every figure deliberately doubled before being used in planning. This does not change published pricing — only our understanding of the cost floor beneath it.
What this means in practice: the 85.6% blended margin in the main brief was already conservative. This verification doesn't ask leadership to change pricing — it simply confirms there's more headroom than originally disclosed, which leadership can choose to hold as extra safety margin, reinvest into lower pricing to grow faster, or both.
06
How we verified this
In the interest of transparency, here's what stands behind the numbers above — described in plain terms, without getting into engineering specifics that sit with the founder and technical team.
Real conditions, not estimates
Each use case was run through Missaka Isuru under realistic conditions — the same syllabus context, system instructions, and typical question or answer lengths a teacher would actually see in production.
Two operating configurations
A fast, lightweight mode handles simple, high-volume tasks like doubt-solving and the institute assistant. A higher-capability mode handles tasks needing stronger structure or visual reading, like paper generation and script grading.
Doubled, on purpose
Every figure in this brief doubles the measured real cost before we plan around it — a deliberate buffer for usage variance, currency movement, and room to upgrade quality over time without surprising the budget.
Exactly which configurations, providers, or models sit behind "fast" and "higher-capability" mode is a technical and competitive decision that sits with the founder and engineering team — intentionally out of scope here, consistent with how we treat Missaka Isuru everywhere else in our stakeholder materials. What matters to everyone else is that the numbers above are real, verified, and already padded for safety.