Stop asking AI.Start working with it.

Stop Asking AI is a free field guide that teaches business owners to work with ChatGPT, Claude and Gemini by delegating real work instead of just asking questions.

Last updated October 24, 2026

Bring a half-formed idea. Talk it through until something clicks. Then hand AI the real work. Free lessons from years of working this way, mistakes included.

Find your spark, free Build your first assignment

The method

Three moves. You'll go back and forth between them, and that's fine.

Bring the mess

Start before you're ready. A ramble, a screenshot, a problem that keeps bugging you. Context beats a perfect prompt.

Find the spark

Push back. Ask what you're missing. Keep going past the first answer until something clicks.

Give it the work

Say what done looks like, check a small sample, and approve anything that goes out under your name.

Field note 001

Why didn't the AI finish it overnight?

A business owner handed AI a huge lead-generation job, walked away, and came back to nothing useful. The fix wasn't a smarter model. It was ten minutes of questions and a ten-lead sample first.

Read the field note

The field guide

Short lessons you can try on your own work today. Start anywhere.

0You don't need a perfect promptStart with an unfinished thought and shape it by talking.Start 1Stop asking. Start delegating.Ask for an answer, or hand over responsibility for a result.Start 2Give AI the whole pictureWhy the real material changes the quality of the work.Start 3Think out loud with AIBrainstorm until the real problem shows up.Start

4Argue with your AIDemand evidence, and notice when it's confidently wrong.Coming soon

5Turn conversations into thingsFrom an idea to a draft, a prototype, finished work.Coming soon

6Build an AI chief of staffHand off recurring work and keep the decisions.Coming soon

7Teach AI how you workTurn repeated corrections into rules that stick.Coming soon

8Know when to let goWhen it can keep going, when it must ask, when only you decide.Coming soon

9When AI gets it wrongRecovering from bad results and false claims of done.Coming soon

10From one assistant to a teamSeveral AIs, tools and checks, only after the basics.Coming soon

Numbers from real work

These lessons come from running a technology company with AI every day. A few measured results, not estimates:

77% fewer AI tokens

In October 2026, moving routine checks from AI turns to simple scripts cut our daily token use from 3.43 billion to 0.77 billion, while the work delivered stayed about the same: 23 merged code changes against 27 the day before.

Since 1994

Over thirty years in enterprise IT, including engineering work at Dell, Apple, AMD and the IBM WebSphere lab, and a Red Hat Certified Engineer credential. More about Jason.

4,000+ merged changes

Merged pull requests on GitHub as of October 2026, nearly 500 of them to other people's open-source projects. See them on GitHub.

What these lessons build on

The ideas here come from daily practice, and from the people building these systems. If you want the source material:

Building effective agents (Anthropic)
Effective context engineering for AI agents (Anthropic)
Prompt engineering guide (OpenAI)
Prompting strategies (Google Gemini)
ReAct: synergizing reasoning and acting in language models (Yao et al., 2022)

What's the idea you can't stop thinking about?

Start with lesson 0. It's free and takes about fifteen minutes.

Start lesson 0 free Share what clicked