---
name: tai-ch108-appendix-templates-for-ai-quality-work
description: 'Apply chapter 108 of Testing AI, Appendix: Templates for AI Quality Work, as a workflow for evaluating AI and non-deterministic systems. Use for test planning, eval design, quality review, release evidence, examples, or coaching related to appendix: templates for ai quality work.'
---

# Appendix: Templates for AI Quality Work

Skill name: `tai-ch108-appendix-templates-for-ai-quality-work`

Based on **Testing AI: Engineering Confidence in Non-Deterministic Systems** by **Jason Arbon**.

## Purpose

Templates make AI quality repeatable without pretending every system has the same risks.

## Use This Workflow

- Identify the AI behavior or release decision being evaluated.
- Define realistic cases, slices, unacceptable outcomes, and evidence needed for confidence.
- Choose measurements that match the risk: rubric scores, samples, intervals, traces, human review, deterministic checks, or production monitors.
- Report uncertainty, severe failures, and decision impact instead of only a pass/fail result.

## Key Guidance

Templates help teams move faster. They also prevent common omissions. The trick is to use them
as scaffolding, not bureaucracy. The most useful templates are eval plans, rubrics, judge
prompts, failure-pattern reports, release memos, and model-comparison tables.

## Apply The Approach

Create representative cases, score them with explicit criteria, review severe failures separately, report uncertainty, and connect the evidence to a concrete decision.

## Deeper Guidance

When the system matters, templates should be machine-readable where possible. Structured release
records make it easier to audit, compare, automate, and mine past decisions.