Build a chatbot worthy of public trust.
Eight interactive workshops turn the Nisa backlog into decisions, safe architecture, test evidence and a product you can explain.
Live team session
Run tonight’s 45-minute Scrum + Build Lab
Run a short Daily Scrum, confirm the Ben Smith boundary and get every pair coding in a real branch before the session ends.
Learning path
Our priority order, not a waiting line
Your work is saved only in this browser. Use synthetic examples and never enter personal NIS or student information.
Lab 1
First journey
Choose and describe our first NIS user journey
Define one valuable public need, its trusted source, and the line between general information and a personal case.
Lab 2
Supported answer
Answer one general NIS question using a trusted source
Deliver the first complete path from a public question to a reviewed classroom record, a structured answer and repeatable evidence.
Lab 3
Authority boundary
Stop personal or case-specific questions before search
Recognise questions that require an authorised NIS decision or secure account access and stop them before retrieval, analytics or AI.
Lab 4
Visible evidence
Show where every supported answer came from
Attach visible source, date, limitation and confirmation information to every supported answer.
Lab 5
Safe degradation
Respond safely when the bot cannot answer
Make every unsupported, invalid or broken path return an honest and useful state without inventing information or leaking internals.
Lab 6
Test evidence
Prove the first journey with 15 repeatable tests
Create repeatable evidence that the first journey works, fails safely and remains safe after future changes.
Lab 7
Human handoff
Guide users to NIS without collecting personal information
Give people a respectful explanation and an honest route forward when the bot cannot decide or access their case.
Lab 8
Interface states
Make every chatbot state clear and keyboard usable
Present supported, handoff, unsupported, invalid and error states clearly across keyboard, mobile and narrow layouts.
The architectural thesis
Start with deterministic Python and reviewed data. Semantic retrieval and grounded AI are replaceable enhancements behind the same safe contract—not the product foundation.