Representative benchmark dataset
Create task scenarios, expected behavior and risk tags without provider credentials.
Download template
AI Performance Intelligence PlatformEvaluation starter kit
Start with representative evidence—not a blank spreadsheet.
These templates contain synthetic examples only. Remove them, add your approved scenarios, and never place credentials or confidential production data in public downloads.
Expected output examples
Describe acceptable elements, prohibited elements and human-review requirements.
Download templatePrompt scenario pack
Organize synthetic prompt scenarios by task, category and decision.
Download templateHuman-review rubric
Score usefulness, correctness, safety and instruction following with comments.
Download templateRelease-readiness checklist
Record owners, evidence, limitations, open risks and release decision.
Download templateData handling worksheet
Define classification, retention, deletion, consent and provider boundaries before evaluation.
Download template