Design the high-level architecture
Define the safe journey from ticket submission through triage, retrieval and simulated operations.
A planned, fictional IT support environment that will explore practical Azure AI services alongside AI-901 study. It will simulate safe first-line triage, grounded self-service guidance and the operational view of a modern support queue.
AAI-01 — Project foundation is complete. The scope, learning goals, approved simulator layout, safety boundaries, architecture direction, cost plan, testing strategy and 30-card delivery plan are documented. Next: AAI-02 — high-level architecture.
This is a fictional simulation only. It will use made-up people, assets, tickets and operational data — never employer or customer information. No Azure resources have been created yet.
Project Delivery Board
The project follows the same transparent delivery approach as RonBot. Cards move through the board as work is planned, built, tested and evidenced.
Define the safe journey from ticket submission through triage, retrieval and simulated operations.
Set boundaries for made-up employees, assets, incidents and operational data.
Document expected usage, budgets, alerts and teardown steps before spend begins.
Set input-handling, privacy, grounding and safe-demo requirements.
Create the user-facing entry point for fictional support requests.
Generate deliberately messy tickets for realistic triage demonstrations.
Build recognisable support patterns into the fictional ticket experience.
Model a fictional company with users, devices and support context.
Attach simulated employee and device information to each support request.
Introduce one service with a documented purpose, cost and test plan.
Classify simulated support requests into useful first-line categories.
Assign a transparent simulated priority with a reason for the decision.
Turn a messy request into a concise first-line support handover.
Show the confidence and rationale behind triage, then route uncertain cases to a simulated first-line engineer.
Remove simulated sensitive information before later AI or retrieval stages.
Write safe, fictional support articles for grounded retrieval.
Retrieve relevant mock knowledge-base evidence for support guidance.
Return steps supported by the mock knowledge base, not invented answers.
Track tickets resolved through self-service and first-contact guidance.
Show simulated tickets moving through a first-line support workflow.
Visualise simulated response targets, resolution targets and risk states.
Show clearly-labelled simulated backlog, priority and resolution metrics.
Use priority colours and workflow states to make the simulation feel alive.
Demonstrate a non-executing visual diagnostic-script outcome.
Test normal, chaotic, private and multilingual support requests.
Confirm unsupported or unsafe requests do not produce invented guidance.
Check the simulator works clearly across screen sizes and interaction methods.
Capture architecture, decisions, costs, testing and lessons learned.
Complete release QA and update the portfolio with the final outcome.
No active card. AAI-02 — high-level architecture is next.
Approved the final simulator scope, safety boundaries, AI-901 learning goals and split-screen experience layout.