Volunteer Guide to Data Collection & Knowledge Entry
How to feed accurate, well-structured school knowledge into Coronation AI's brain.
Why This Guide Exists
Coronation AI only knows what we teach it. It cannot guess RCHS's routine, rules, or syllabus on its own — every correct answer it gives a student later comes from data that a volunteer collected and entered correctly today.
This guide explains, in plain language, how the AI's "brain" is organised, what kind of data goes where, how to format it so the AI can actually use it, and how to check your work before submitting it. Follow it closely — a well-structured entry makes the AI accurate; a messy one makes it wrong or confusing.
- Never dump a raw PDF, photo, or paragraph into the system and call it done.
- Always break information down into small, labelled, structured pieces before entering it.
- Structured data = the AI can find and use it. Unstructured data = the AI has to guess.
How Coronation AI's Brain Is Organised
Think of Coronation AI's knowledge as four separate shelves. Each shelf holds a different kind of information, and each has its own rules for how often it changes and who is allowed to update it:
| Layer | What it holds | Who updates it | How often it changes |
|---|---|---|---|
| 1. Official External | WBBSE / WBCHSE syllabus, board notifications, government circulars | Academic team lead | Once a year, or when the Board publishes updates |
| 2. Official School 🌟 | School profile, rules, calendar, notices, facilities | Admin office + volunteers | A few times a term |
| 3. Academic Content 🌟 | Textbook content, chapter breakdowns, learning outcomes, teacher-approved explanations | Subject volunteer teams | Ongoing, chapter by chapter |
| 4. Live Data | Class routine, homework, attendance, events — pulled from school ERP | Automatic (via ERP), not manual entry | Daily / real-time |
💡 Note for Volunteers: As a student volunteer, you will almost always work on Layers 2 and 3 (school information and academic content). Layer 4 updates automatically from the ERP, and Layer 1 is handled by the academic team lead.
The Golden Rule: Structure Before You Submit
Coronation AI reads data as small labelled records — not as free-flowing paragraphs. Every fact should have a clear label (a "field") and a clear value. This is the single biggest difference between data that works and data that doesn't.
What Data To Collect (Phase 1 — Priority)
These categories come first because Coronation AI cannot answer basic questions without them:
| # | Category | Examples of Questions it Answers | Where to get the source data |
|---|---|---|---|
| 1 | School Profile | "When was RCHS established?" "Where is the computer lab?" | Admin office, school prospectus, headmaster's office |
| 2 | Classes & Sections | "Which sections does Class IX have?" | Admin office academic structure records |
| 3 | Subjects & Teacher Mapping | "Who teaches Physical Science in IX-B?" | Staff room, subject allocation list |
| 4 | Class Routine | "What does IX-A have on Tuesday period 3?" | Current routine chart — entered into ERP, not typed directly |
| 5 | Academic Calendar | "When do Second Summative exams start?" | Office calendar / notice board |
| 6 | WBBSE / WBCHSE Syllabus | "What is in the Class IX Physical Science syllabus?" | Official WBBSE / WBCHSE website and publications |
| 7 | Textbooks | "Explain latent heat the way our textbook does." | Approved textbooks for each class and subject |
| 8 | Examination Info | "What is the Madhyamik exam pattern?" | WBBSE notifications + school exam office |
| 9 | School Rules | "What are the library rules?" | Student handbook, admin office |
| 10 | Contact Info | "How do I contact the school office?" | Official school records only |
The Volunteer Workflow — Step by Step
Every piece of data goes through the same five steps before it becomes part of Coronation AI's live knowledge:
Collect
Get the information from an approved, official source. Never rely on memory, rumour, or an outdated printout.
Structure
Break the information into labelled fields using the templates in Section 6. Use the exact field names given.
Bilingual Check
If the entry is student-facing academic content, add both an English and a Bengali version wherever possible.
Submit for Review
Enter your structured data into the Coronation AI volunteer module. It will be saved as "Pending Review".
Reviewer Approval
A subject lead or teacher-in-charge checks the entry for accuracy and publishes it live.
Volunteers collect and structure data. Only an authorized reviewer (teacher / subject lead) publishes it live to ensure 100% safety and accuracy for our students.
Entry Templates by Category (Ready to Copy)
Use these exact field formats when entering facts. Click "📋 Copy Template" to copy the template directly to your clipboard:
Bilingual Entries (English + Bengali)
Coronation AI serves both English- and Bengali-medium students, so student-facing academic content should be entered in both languages wherever possible:
- Always keep English and Bengali versions linked: Keep both versions under one entry rather than two separate disconnected records.
- Use simple, textbook-level Bengali: Avoid literal automated machine translations. Use approved terms (e.g. লীন তাপ for Latent Heat, সালোকসংশ্লেষ for Photosynthesis).
- Mark pending if not ready: If a Bengali translation isn't available yet, submit the English version and note
Bengali pending.
Quality Checklist — Self-Audit Before Submit
Run through this checklist for every single entry before clicking submit:
Do's and Don'ts
- Break every fact into its own labelled field.
- Cite where the information came from (handbook, textbook, notice).
- Ask your team lead if a source looks outdated.
- Use the school's exact official spellings.
- Submit small, complete entries often.
- Paste a whole PDF or textbook page as one giant entry.
- Guess at a fact you're not sure about — flag it instead.
- Publish an entry without teacher review.
- Mix two topics or classes into one record.
- Add personal phone numbers or private data.
Roles in the Data Pipeline & Project Lead Contact
Gathers source material and structures it using the templates in this guide.
Checks submitted entries for accuracy and approves them for live AI ingestion.
Oversees the AI pipeline and ERP integration.
For questions regarding system access, data guidelines, or category assignments, contact Debasis Mandal (Assistant Teacher, Computer Science / Coronation AI Project Lead).
🚀 Submit Structured Knowledge Entry
Directly feed structured records into the Coronation AI review queue.