Dominate the BOT
10 Ways to Get Control Over Your AI and Technology.
I wrote about freedom & bondage this week…power, control, who’s really in charge.
The same rules apply to your AI tools and your use of technology.
Many people open AI and just... talk. No structure or rules. Then they wonder why the output feels generic or wrong.
You’re the top here…act like it.
Here are 10 ways to stay in control of your bots. Copy these prompts and use them.
1. Give it your safe word
Most major AI tools let you save information about yourself so you don't have to repeat it every chat.
Fill it out once, so you can stop repeating yourself in every chat.
Try this: “I’m a [your role]. I write for [your audience]. I hate corporate jargon and fluff. Keep responses direct.”
Most people never touch this setting. That’s like never learning your partner’s boundaries and wondering why things feel off.
2. Set the scene before you start
Don’t just ask a question. Give the bot a role first. It changes everything about the output.
Try this: “You are a blunt editor who cuts anything that sounds like AI wrote it. Review this paragraph.”
Vague requests get vague answers. Specific roles get specific results.
3. Use a leash, not a whip
Being polite doesn’t help. Being vague doesn’t help either. Precision is what gets you what you want.
Skip “Can you help me with this email?” Try telling it exactly what you need, in what tone, for what reader.
Most people ask AI the way they use Google...a few words and a lot of hope.
4. Make it ask before it acts
Good AI tools will ask clarifying questions if you let them. Most people skip this step entirely.
Try this: “Before you answer, ask me anything you need to know to get this right.”
You’ll be surprised by what it flags that you forgot to mention.
5. Don’t let it freelance on facts
AI will hand you a confident, well-written, completely made-up statistic if you let it. This happens more than people think.
Try this: “Flag anything in this response you’re not fully certain is accurate.”
Then verify anyway. Always.
6. Correct it out loud, every time
If it gets your tone wrong or misses your point, don’t just quietly fix it yourself and move on. Tell it what went wrong, specifically.
Try this: “That was too formal. I write like I talk. Try again, looser.”
Bots don’t learn from silence. They learn from correction.
7. Keep your receipts
When a prompt works, save it. Build yourself a running list. This may be the biggest AI time-saver nobody talks about.
A folder of five prompts that consistently work for you is worth more than a hundred random ones you’ll never find again.
8. Never let it leave the room unsupervised
If you’re using AI for research, client work, or anything with your name on it, do not let it wander off and cite sources you haven’t checked.
Try this: “List every claim in this response that needs a source, separately, so I can verify.”
It’s your name on the byline. Not the bot’s.
9. Set a curfew
Know when to stop. AI is fast at drafts and slow at judgment. The longer you let it run unsupervised, the more generic everything gets.
Use it for the first 70 percent. Bring your own voice for the last 30. That’s not a rule I made up for content. That's how I stay in control.
10. Remember who’s really on top
At the end of every session, you make the call. You edit. You cut what’s wrong. You decide what goes out under your name.
I also make the BOTS fight… not literally, though that would be entertaining.
I routinely ask ChatGPT, Claude, Perplexity, Copilot, and Gemini the same question. Sometimes they agree. Sometimes one completely blows it. When they disagree, I dig deeper. AI shouldn’t replace your thinking. It should sharpen it.
The bot drafted this list with me. Then I made a few competing bots critique it. I cut half of it, rewrote the rest, and I’m the one hitting publish.
That’s the whole point.
Researched with Beulah, my AI collaborator. Direction, edits, and final judgment: Nancy AF, the robot overlady.
Stay spicy, Nancy AF