Hi Reader,
I give AI a lot of information about me, my business, and what I'm trying to accomplish.
I've uploaded things I've written, trainings I've taught, transcripts and recordings, examples of my content, information about my business, and the goals I'm working toward. I've even created a source-of-truth document that gives AI a much deeper understanding of my business, my work, and how I think and communicate.
And after all of that, it still gets me wrong sometimes.
The interesting part is that it's not always obvious. The response might sound perfectly reasonable. It might even sound like something I could say. But every once in a while, I'll read something and think, Wait. That's not quite me.
THE OPENING SIGNAL
The more context we give AI, the better it can get at giving us responses that seem to fit.
That's part of what makes AI more useful. But there's a subtle risk that comes with it. The more familiar and convincing the response sounds, the easier it can be to assume AI understands us better than it actually does.
I think that's something worth watching as we give AI more context about our businesses, our goals, and the way we work.
WHAT ACTUALLY MATTERS
Giving AI more to work with helps. The more it understands about you, your business, your goals, and the way you work, the more useful it can become.
But I've learned that more context doesn't mean I can stop paying attention.
Sometimes AI will give me an interpretation of something I've said, a point of view it thinks I hold, or even an explanation of why I made a particular decision. And because it knows so much about me and my business, the answer can sound completely reasonable.
That's when I pay attention to that little "Wait. I don't think that's quite right" reaction instead of assuming the AI must know something I forgot.
I've actually asked it, "Where in my source material did I say that?"
Sometimes it can show me exactly where the idea came from, and I can see how it made the connection. Other times, it has taken several pieces of information and made a leap I wouldn't have made.
That distinction matters because using AI as a thinking partner isn't the same as letting it become the authority on what you think.
You can give AI a lot of information about you without giving it the final say on who you are, what you believe, or how you think.
For me, that's part of staying in the process. AI can help me organize an idea, make connections, challenge my thinking, and give me something to react to. But I still have to decide what fits, what doesn't, and what actually represents me.
ONE SMALL MOVE
Choose two or three things you've written that sound unmistakably like you and give them to the AI you use.
Then ask:
Based on these examples, what do you notice about how I communicate, what I tend to emphasize, and how I approach ideas?
But don't stop with its answer.
Look for what it got right and, more importantly, what it misunderstood. If something doesn't fit, tell it. Explain what it got wrong and why.
You're not just teaching AI more about you. You're practicing staying involved in the process.
WORTH PAYING ATTENTION TO
This isn't just something I've noticed in my own work.
Michael Hyatt recently wrote about 16 patterns that can make AI-assisted writing sound like AI. He'd been using Claude for months and had given it his books, articles, and an entire style guide to help it understand his voice.
Then his longtime editor reviewed a webinar script he'd drafted with AI and immediately started flagging things that didn't sound like him.
What I found most interesting is that Hyatt hadn't caught all of them himself.
He'd gotten so used to seeing some of those AI patterns that he'd started overlooking them himself. His takeaway wasn't that AI shouldn't be used for writing. It was that the problem starts when we use it on autopilot.
That's why I think his list is worth looking at, even if you don't agree with every "tell" on it. It's a useful reminder that keeping your voice isn't only about giving AI better instructions or more examples of your writing. It's also about keeping your own judgment switched on.
Read: 16 AI Writing Tells and How to Fix Them →
I want AI to know a lot about me, my business, and what I'm trying to accomplish. The more it knows, the better thinking partner it can be.
But being a good partner in that relationship means I have a job to do, too.
Where has AI gotten you wrong? Hit reply and tell me. I'm curious what you've noticed.
Talk soon,
Lucy
P.S. I have lost track of how many times I've told AI, "I would never say that." 😂 You'd think it would know by now.