The Free Library
Your practical AI starting point.
Every guide begins with a job, a decision, or a problem. You get enough detail to try the method, inspect the assumptions, and decide whether it belongs in your work.
Three guides. More in progress.
I am building this library one finished piece at a time. I would rather publish three guides that explain the work than fill a page with thin summaries written to look busy.
The first set covers the decisions that come before a useful AI result: defining the job, judging whether the idea is worth pursuing, and designing the process around the tool.
3 guides · Free to read
Before You Prompt, Define the Job
Write down the result, reader, source material, limits, and review before the model starts guessing. Includes a practical five-part worksheet you can reuse.
Read the guide →
A Five-Question Check for Any AI Idea
Test the complete workflow against time, cost, complexity, employee load, and quality. The fast step alone does not tell you whether the idea works.
Read the guide →
Build the System Around the AI Tool
Map the input, AI task, review, decision, and next action. Then give errors a route back before people start repairing the system in secret.
Read the guide →What Belongs in a Guide
Enough context to make your own decision.
A prompt without its source material can be impossible to judge. A result without the repair time can make a weak method look efficient. A recommendation without its limits pushes the decision onto the reader while pretending it has already been made.
My guides include the conditions that shaped the answer. For tool tests, that means the product, plan, date, input, steps, checking, and outcome. For methods, it means the question the method answers, the situations where I would use it, and the places where it needs a stronger review.
The aim is practical independence. You should leave with a method you understand, not a phrase you have to trust because I wrote it.
Help Shape the Library
Bring me a task.
Tell me what you are trying to do, where AI is getting in the way, or which claim you want tested. Specific questions make better guides.