(AI for copywriting) 10 things I've learned after 18 months of daily use

It won't replace you. But it probably won't be as good as the you who uses it.

Forget about choosing between writing copy yourself or letting AI do it. Use both.

Here's how I work with AI to produce better copy (based on 540 days of using it every day):

I work with AI tools open all day. Two tabs are always up: Claude and ChatGPT.

I use them for everything. Not just writing. They're like having two extra Brooks’s.

1. Crunching way more data than I ever could myself

Experienced copywriters know writing is only ~50% of the copywriting process.

The other half involves data & strategy.

I’m not gonna go buy a $20k per month enterprise sentiment analysis tool.

So I use ChatGPT to help me find patterns in vast, publicly available datasets like G2, Capterra or even Amazon book reviews.

I have another post that gets into the weeds on how to do that.

Net: Use AI to crunch more data.

AI can analyze thousands of reviews in seconds.

When you present copy to clients, back it up with data.?

Say things like:

"This phrase appears in 437 positive reviews."

Clients loooooove this.

2. Never treat AI models like glorified Google search

If you do, you'll get generic outputs.

Memorize this:

1. Unique data
2. Unique context
3. Iteration

Did you memorize it?

People still conflate "let the model do all the work for you" with "use models intelligently as a collaborator" which is leaving value on the table.

Without these three ingredients, AI can't do anything useful.

3. Generating options

I use AI to create multiple versions of copy within the strategic framework I built.?

This gives clients choices without straying from the core message.

It's easier to say to a client, "I've put together two options for a draft. Based on your customer data, which direction feels right?"?

This gives the whole process concrete deliverables at every step. Clients love it.

Here's a real example of how I frame it for clients:

The faster you can work through iterations on copy especially, the better.?

But again, none of this is a substitute for having the client expertise, data and intuition to be able to know the market context and feed all of that to AI.

Here’s a template you can build on:

—

"None of this is set in stone. This is a draft to get your feedback. The goal is to close the loop quickly so I don't go too far in one direction without checking in first. With that context, please provide some directional feedback."

—

This approach has saved me tons of time on rewrites.

4. Easing client approvals

Here's what I give clients:

• Copy only I've written
• Collaborative pieces (AI + me)

Each version gets my seal of approval.

Why does this work?

• Clients have choices to show their team
• It helps navigate office politics
• Stakeholders feel heard

Result: Quicker approvals, happier clients.

But it’s about quality options.

Every version is client-ready. AI helps, but I'm the filter.

This approach turns "We need more options" into "Let's pick one and go."

5. Expanding my toolkit

Tools are leverage. I've added AI tools to my workflow.

My current AI stack is simple: Otter, ChatGPT, and Claude.?

That's it for now, but I'm sure it'll grow over time.

Think of AI as just another set of tools in your business, like Google Workspace, Notion or Slack. They're powerful if you get good at using them.

6. I can handle 2-3x more client projects with AI (same quality level)

With AI, I can handle 2-3 times more client work. That means 2-3 times more revenue.

People often talk about productivity, but what's the second-order effect? What does it look like accumulated over a year? Five years? Ten years? All those efficiency gains add up and compound, meaning more revenue for you over time.

If you're really attached to the creative process of writing, this might not be for you.?

But for those who want to work smarter and earn more, AI is fucking amazing.

Don't treat it like a magic button that does all the work. It's not that.?

It's a collaborator that amplifies your execution.

7. No one’s really an expert at using Generative AI (yet)

No one's an expert at AI copywriting yet. It's early days. We're all learning.

And that's exciting because craft still matters. Practice still matters. AI is just a new tool.

It's like getting a power saw when you're used to a handsaw. Same skills but faster execution.

Being a pessimist is self-fulfilling.

There's no playbook yet, which means you can write your own unique one.

8. I treat AI like a mirror of myself, not an assistant

Think of your AI-enabled business like a law firm rather than a hierarchy.?

The AI models you use are only going to go as far as your skills go. It's the ultimate "you get out of it what you put into it" situation.

Feed it good information. Give it context. Iterate on its output. The results get better when you work with AI, not just let it work for you.

9. I still write and edit without AI?

I still don't use some of the outputs. It's not a zero-sum game where either AI or you do the work. It's both.?

I once took an improv class, and the very first thing they teach you is "YES-AND". Take that approach with AI.

10. Prompting the AI Communicating with the AI

Forget about "prompting" AI. That's short-sighted.

A single prompt is like asking a question and walking away. You might get an answer, but it probably won't be the best one.

Instead, think of it as a conversation. A back-and-forth. An iteration.

It goes like this:

1. You give the AI some context and a task.
2. The AI responds with an output.
3. You review the output and give feedback.
4. The AI adjusts based on your feedback.

Repeat steps 3 and 4 until you're happy with the result.

—

What's been your experience with AI and copywriting? Have you found use cases I didn't mention? I'd love to hear about it.