Case studies
Every case on this page is a fictional training scenario written for practice and assessment. Names, numbers, and outcomes are invented; none describe a real customer, company, or engagement.
Case Study: When the Rules Engine Beat the AI
A fictional insurer's claims triage 'AI initiative' meets a deterministic classification problem - the FDE's best move is proving a 400-line rules service wins on accuracy, cost, latency and auditability, and shipping that instead.
Case Study: Rescuing an AI Service Drowning in Latency and Cost
A fictional document-summarization feature reaches production and meets its enemies: p95 latency of 14 seconds and a monthly invoice nobody predicted - measurement, tiering, caching and an honest fallback turn it around.
Case Study: An Approval-Gated Customer Workflow
A fictional logistics firm automates claim responses - AI drafts, a human approves, the audit trail satisfies the compliance team, and the override-rate metric tells you whether the gate is safety or theater.
Case Study: Taming Mixed-Format Supplier Data
A fictional retailer receives product feeds as CSV, JSON and one XML 'because the supplier is special' - the project is a validation pipeline with a data-loss report a buyer can actually read.
Case Study: A Permission-Aware Support Assistant
A fictional B2B software firm wants a support assistant over internal docs - the interesting work is the authorization boundary: agents must only ever retrieve what their role allows, and the test suite tries to break that.
Case Study: Cutover from a Legacy ERP to a New Order API
A fictional wholesale distributor replaces nightly CSV imports with a new order API while keeping the legacy ERP running - duplicate events and retries make the migration an idempotency exercise.