I don’t particularly like talking about AI/LLMs. I’m not an expert on them and I don’t feel right about seeming to just glom onto a trend. But I’m a knowledge worker in 2025 so I obviously have a point of view.
And the truth is, if you aren’t incorporating ChatGPT, etc into your various day-to-day tasks, you’re seriously missing out. I lean on Perplexity when I want to do a well-defined search (e.g. a list of Pilates studios within a 10-mile radius of my home, ordered by cost per class). I use Claude to draft my outlines and visuals for my presentations. I use ChatGPT for pretty much everything else. It’s basically my responsive journal.

That being said, if you are using these tools — and it shows — you’re doing it wrong. Some tough love:
Have Claude write all your posts for you? Er… if you’re that easily replaceable then there’s a good chance your voice isn’t all that unique.
Found out you were wrong about information you shared? Well… don’t ask ChatGPT for facts and then blindly repeat them.
You used ChatGPT for math and then the math didn’t math? Most LLMs are trained on language, not math.
I draft many of my LinkedIn posts in Claude now, but I do it very strategically. Here’s a recent example of a post that performed quite well:
I prompted Claude which wrote me several drafts. The orange highlight represents my edit derived from a Claude suggestion. Purple highlights indicate the stats/numbers that Claude pulled out and I verified. Green highlights indicate 100% me, not what Claude wrote. The non-highlighted portion is what Claude wrote and I left untouched.
I believe this post performed well because it still came from me. I “Amanda-ified” it, and most importantly, I verified the accuracy. One of the stats that Claude originally pulled from my report was worded incorrectly, so I rephrased it. A less careful person who doesn’t know the source content well wouldn’t have caught that.
It might sound like this took me longer than if I didn’t use Claude at all, but on the contrary. It got me past the blank page. It got me to the shitty first draft. And then I edited, edited, edited, and made it mine. As an opposite approach, I’ll also write something from scratch then feed it into Claude to identify grammar errors or lack of clarity. Again, I edit and make judgments before I publish.
This doesn’t just apply to creators. It applies to all of us. I really believe we can use these tools to do better output — but it’s on us to keep the bar high. Use LLMs for the work you’d outsource to a less-skilled-you. Take advantage of that time/energy regained by focusing on what only you can do.
This Thursday, I’m going to share some of my AI workflow ideas in light of modern content marketing advice. I’ll be co-hosting a 30-minute session with my friend Matt Lerner, Co-Founder of SYSTM and Paypal OG. Even if you can’t join us live, register to get the recording. We’ll have a couple of slides, but overall you can expect a value-packed, casual discussion.
» Register for our free webinar! See ya Thursday.
Now tell me… what do you use ChatGPT and other AI tools for? I’d love to hear your specific tips and tricks in the comments!
So far, my only consistent use case for AI is as a line editor. Particularly if I get stuck in the grammar of a sentence, or find myself running on too long. I'll literally just say "can you make this sentence less bad?" and it usually can - although nothing I could copy/paste, just the format/structure.
I also like using it to trim down LinkedIn posts to fit the character limits.
I think there are risks to using it to write the shitty first draft for you - or even write an outline. That process of going from zero to something is part of what gives a piece of content its originality. I enjoy content that brings together diverse frameworks, studies or lines of thought - which AI will never do. I've found that reaching for AI too quickly can start to become a crutch - and cause my creative muscles to atrophy.
I just worry about the ecological costs of of all this. AI is harming the environment, and it's wild no one talks about that in articles like this. Is it ignorance? Laziness? It seems AI is so popular with Gen-Z and this fact should not be ignored.