Skip to content

Get one real thing done with AI.

Independent Practice · First Win

One real thing off your own plate, in the next fifteen minutes, with the AI you already have. A reply, a long email, a list. You keep the result.

Pick something small that has been bugging you. A reply you keep putting off, a long email you have not read, a list that needs sorting. We do that one, you watch it work, and you keep the result.

Boden FullerI'm Boden. I spend my days on systems that have to be checked before you trust them. This page is one safe first step you can do on your own, in plain language.Or start with a free conversation →

// the one rule that keeps you safe

Before anything else, know this. It is true no matter what you ask it to do.

  • It reads and it drafts. You are the one who sends, spends, and decides. Every time.
  • Never type a password, a code, or an account number into it. It does not need them.
  • Anything about your money, your health, or the law goes to a real person, not the AI alone.
  • If a message rushes you about money, stop. Ask the AI if it looks like a scam, then call your real person or your bank before you act.

// do it now

Fifteen minutes. The AI you already have. One task you need done.

1Pick one small thing that annoys you. A reply you owe, a wall of text to get through, a messy list. The first one that came to mind is the right one.
2Open your AI. The chat box on Claude, ChatGPT, or Gemini. Any of them is fine.
3Say what you want in plain words, then paste the thing under it. “Write a short, friendly reply saying I can meet next week, not this one.” Then paste the email below it.
4Read what comes back. Not right? Say so plainly. “Shorter.” “Warmer.” “Drop the part about the budget.” Go back and forth until it fits.
5When you like it, you copy it in and you hit send. The AI drafts. You decide.

That is the whole rhythm. You ask, it drafts, you check, you keep what is good. You just cleared the task that was sitting on your plate. Do it once more on a different task, and the rhythm is yours.

// when it's lying

It can be confident and wrong at the same time. The confidence tells you nothing.

Your AI can make up a fact, a date, a name, or a number, and say it as smoothly as the true parts. So before you trust it or send it, you look with your own eyes. You do not ask the AI whether it got it right. It will say yes in the same voice it says everything. You check the real thing yourself: the actual email, the actual number, the actual date it came from.

It states a fact with nothing behind it. Clean, specific claim. Vague when you ask where it came from.
It did something you did not ask for. The extra it added is the part you did not check.
It is sure, with no detail underneath. Certain tone, thin substance, a little too eager to wrap up.

The longer version: how to tell when your AI is lying.

// make it stick

Two small habits turn a one-time win into a partner that gets better the longer you use it.

  • Tell it who you are, once. Your name, your people, the words you use. It stops sounding like a stranger, and you stop repeating yourself.
  • Keep what works. When you fix something, save the fix. Next time it starts from everything you already taught it, not from scratch.
Want the whole thing, set up around your life?

That first win is enough to feel the rhythm. The full partner handles email, calendar, and the projects you keep putting off, and I set it up with you and teach you to run it yourself. The first conversation is free, for you or someone you'd like to help. Start one here.

See what a partner looks like →