THE PLAYBOOK
AI agent guides for everyday tasks
Reusable methods for briefing, evaluating, and working with AI agents.
Give your agent a brief it can actually follow.
Turn a vague request into a clear task, with a definition of done.
Test the task. Not the demo.
A repeatable scorecard for accuracy, effort, and useful results.
Build a research shortlist with traceable sources.
A practical workflow you can adapt to your next decision.
NEW FIELD GUIDES & TOOLS
01Laya Model: Decisions for AI Workflows
Understand Laya decision models, where they fit in agent workflows, and how to evaluate routing on your own data.
Run Laya Locally: Setup and First Check
Set up an isolated Python environment for Laya and record model versions, raw outputs and first-run timing.
Laya Email Triage: A Reproducible Test Kit
Download synthetic email-routing cases and a Laya test script. Review per-case decisions before connecting a support workflow.
Sonnet 5.5 vs Opus 5.5: Cost and Test Plan
Compare published Sonnet and Opus 5.5 API rates, estimate token costs, and download a repeatable task evaluation worksheet.
Hotel Lobby AI Prompt Builder
Build and copy a Hotel Lobby video prompt with subjects, framing, motion and continuity instructions. No video generation or photo upload.