Take a Look Inside
Open a focused preview of the supplied pages, sheets or working views.
112-page paperback, 3 inherited working practices, KDP cover, fillable and printable digital editions, evidence and exercise-lineage registers, and marketplace population package.
What It Is For
A reproducible record of the task, baseline, system version, test cases, failures, review burden and deployment decision.
Intended for: Founders, operators, product teams and consultants running bounded tests of prompts, workflows or agents.
Three Decisions It Helps You Make
- 01
What real situation or decision will you use this for?
- 02
What evidence, assumptions and limits must be visible before you act?
- 03
What result or review point will tell you to continue, adapt, pause or stop?
Price and Package
What this edition does not promise
- A working aid, not a guarantee of a commercial, career or personal outcome.
- It does not replace legal, regulatory, clinical, financial or employer-specific advice.
- Human release review and fulfilment testing remain required.
Source Corpus Acknowledged
This journal contains 3 named working practices adapted from The Evidence-Based Life contract v1.2.0. Each records its retained mechanism and domain boundary. Named research, official guidance, standards and relevant primary provider material informed selected prompts and boundaries. Provider material is operational guidance, not independent validation; no named source endorses this Stratogenic product.
6 named sources and transfer limits
- OpenAI. Working with evals. — Informs criteria, representative test data and repeatable evaluation runs. Limit: Provider guidance is product-specific and is not independent evidence of safety or fitness. Review: Release review required.
- Anthropic. Demystifying evals for AI agents. — Supports evaluating multi-step behaviour, tool use and outcomes. Limit: Provider engineering experience does not remove the need for independent, use-specific review. Review: Release review required.
- NIST AI Risk Management Framework 1.0. — Structures govern, map, measure and manage prompts. Limit: A voluntary framework is contextual; using it is not certification. Review: Release review required.
- NIST AI 600-1. Generative Artificial Intelligence Profile. — Adds generative-AI risk and measurement questions. Limit: A risk profile does not make a system safe. Review: Release review required.
- OWASP. Top 10 for LLM Applications 2025. — Informs adversarial cases, excessive-agency and sensitive-data boundaries. Limit: A checklist cannot replace threat modelling or a system-specific security review. Review: Release review required.
- Regulation (EU) 2024/1689, Artificial Intelligence Act. — Keeps role, risk, transparency and human oversight visible. Limit: Obligations and dates are context-dependent; this is not legal advice. Review: Release review required.
Corpus review: 2026-07-19. Provider documentation is operational guidance, not independent validation or endorsement.
Frequently Asked Questions
What is The AI Experiment Logbook?
The AI Experiment Logbook is field journal designed to help founders, operators, product teams and consultants running bounded tests of prompts, workflows or agents. a reproducible record of the task, baseline, system version, test cases, failures, review burden and deployment decision.
What do I receive with The AI Experiment Logbook?
112-page paperback, 3 inherited working practices, KDP cover, fillable and printable digital editions, evidence and exercise-lineage registers, and marketplace population package.
Can I buy or download The AI Experiment Logbook now?
No. This is a forthcoming preview. Checkout and delivery stay disabled until named human approval and fulfilment checks are recorded.
