Ship a Week of Work, Not a Perfect Prompt
Your capstone is a governed decision workflow: sources in, claims checked, human authority preserved, and state handed off. Combine product selection, prompting, knowledge management, validation, governance, troubleshooting, and handoff design. Build a source-backed weekly briefing workflow with explicit checkpoints. Implement a deterministic Python validator for sources, claims, governance, and handoff readiness. Run normal and failure cases before recommending release. Produce evidence of readiness instead of claiming that a workflow is safe. You are the operations lead at Northstar Field Services, a fictional company with seven regional teams. Every Friday, the leadership group needs a brief covering service delays, customer-impacting incidents, policy exceptions, and two decisions for the next week. The current process is fragile. Regional leads send updates in different formats. An analyst copies facts into a document, reconciles conflicting dates, and asks three people for approval. The final brief sometimes misses a region or carries a corrected number from an old message. Leadership asks you to "automate the weekly report with Claude." That request is not a solution. The brief affects staffing and customer communications. Some inputs contain customer details. Policy exceptions need an authorized owner. A confident draft without evidence could accelerate the wrong decision. Your job is to design a bounded Claude-assisted workflow. Claude may extract, compare, and draft. Deterministic checks verify exact properties. An authorized person…
Ship a Week of Work, Not a Perfect Prompt: Your capstone is a governed decision workflow: sources in, claims checked, human authority preserved, and state…
This free lesson is part of the AI Engineering from Scratch curriculum. Read the full explanation, run the lesson code, and verify the result in the interactive reader or from the repository source.
Browse the complete course catalog or open this lesson on GitHub.