1 min lesson
Deliverables and the demo script
Walk through each part of "Deliverables and the demo script", then explain what each one does.
Step 1 of 2
Deliverables and the demo script
The build is half the score. The other half is the short spec/runbook and how you narrate it, because this role lives or dies on written communication and clear decisions.
- Working flow
- Runs end to end on the sample list and survives the four injected failures.
- Spec / runbook
- One page: architecture, the decisions and their rejected alternatives, how to operate and debug it.
- One metric
- Match rate + score distribution + routed-bucket counts, with the definition stated.
- Demo script
- Decisions made, options rejected, the AI guardrail and how the flow generalizes to another program.
Open your demo with the decision, not the tour: "The core decision was ordering the waterfall by cost-adjusted hit rate and capping calls per lead; here's what I rejected and why." Then show it running, including a deliberately broken row so they watch your error handling catch it live. Ending with "and here's the one config change to point this at the partnerships program" closes the loop on generalizable infrastructure.
Learn more
Optional practice
Practice: Deliverables and the demo script
QIn the practical build, you add an LLM column to classify each company's segment. The reviewer asks, "What stops a wrong label from routing a high-value lead into nurture?" What's your answer?