Free AI Implementation Templates
The four working artifacts from the book's FDE Toolkit: copy them, adapt the vocabulary, and run your first disciplined AI pilot this quarter.
- A one-page pilot charter for AI projects: workflow, baseline, success metric, eval set, permission line, and kill criteria — written before any building.
- How to build a minimal evaluation set for an AI system: collect real cases, define pass/fail dimensions, grade with an LLM judge plus human audit, and use it on every change.
- An honest ROI template for AI pilots: benefits, inference costs, build, integration, change management, and run costs — with the lessons that generalize.
- A ten-field register for every AI deployment: autonomy class, data touched, EU AI Act tier, injection surface, evals, oversight, monitoring, labeling, and review.
The method that ties them together is Chapter 14: Working Like a Forward Deployed Engineer.