How to create AI-generated content (that doesn't suck 🥴)

‍Writing is thinking. To write well is to think clearly. That's why it's so hard. — David McCullough

Please stop using AI to write your first drafts.

One of the first (and most obvious) uses of generative AI is writing. We see this reflected in research; one McKinsey study found the most commonly reported use case was “content support.”

And anecdotally we know that professionals are using ChatGPT for everything from blog posts to emails to LinkedIn posts (and LinkedIn comments 🙃).

But it's also one of the worst ways to use AI. (See: LinkedIn comments.)

Yes, AI can generate content instantly. But should it?

At its core, generative AI finds patterns in data and produces outputs that represent the statistical average of what it's seen. When writing, this means it's likely to produce the most common — and therefore least original and interesting — text.

Here's why this matters: Outside of producing mediocre content that no one wants to read, the very act of creation, of putting thoughts down on paper or creating a first draft, is often the exercise that clarifies and sharpens our thinking.

So when you hand over that first draft to AI, you're not just outsourcing your writing. You're outsourcing your thinking. And that means potentially:

• Missing insights that come from wrestling with ideas
• Limiting your own growth and learning
• Arriving at less inspired conclusions and content

But here's the good news: AI can still be an incredibly powerful creative tool — when used at the right time and in the right way.

The right way to use AI in your creative process

AI isn't just a writing tool — it's a creative partner that can work across multiple modalities (text, images, video, audio) and languages (natural language, programming languages, design specifications). Here are some ways to leverage it effectively:

• Write reports: After you've outlined your key points and gathered your thoughts, use AI to help structure and polish your writing
• Generate code: Once you've mapped out your program's logic, use AI to help implement specific functions or optimize your code
• Design visuals: After sketching your concept, use AI to help refine layouts or generate variations
• Create presentations: When your narrative is clear, use AI to help with slide design and transitions
• Develop training materials: Once you've identified key learning objectives, use AI to help create supporting materials and exercises
• Draft communication templates: After establishing your message strategy, use AI to help create consistent, scalable templates
• Produce multimedia content: When your creative direction is set, use AI to help generate supporting assets

Notice the pattern? In each case, AI comes in after the initial thinking and planning. It's an enhancer, not a replacement, for your creative process.

✏️ Creative Process Starter Prompts

Here are some ways to integrate AI into your creative workflow:

I've written a first draft of [content type]. Can you help me identify areas where the structure could be improved?

Here's my outline for [project]. Can you suggest different ways to organize these points for maximum impact?

I've sketched out these key messages for [audience]. How could we adapt the tone and style to better resonate with them?

I have a rough draft of [content]. Can you help make it more concise while maintaining the key points?

Here's my current version of [content]. Can you suggest ways to make it more engaging while keeping the same core message?

I've written this technical explanation. Can you help adapt it for a [specific audience level]?

💡 Tips for Better Results

When using AI in your creative process:

Start with your own thinking

• Develop your core ideas and key messages first
• Create rough outlines or sketches before involving AI
• Use AI to refine and enhance, not replace your creative process

Be specific about your needs

• Clearly define your audience and objectives
• Provide context about tone, style, and constraints
• Include examples of what you're looking for (and what you're not)

Iterate and refine

• Use AI as a collaborative partner, not a one-shot solution
• Try different approaches and compare results

Consider different modalities

• Consider how different formats could enhance your message
• Think about how to combine text, visuals, code, or other elements
• Use AI to help create consistent experiences across formats

Most importantly, remember that AI is most valuable when it enhances rather than replaces your creative thinking.  

Use it to polish your ideas, not generate them. Use it to refine your voice, not substitute for it.

How I used Claude to write this article

I decided to illustrate my point by allowing Claude to write what you just read. (I made a few minor changes, but left about 90% of the writing as-is.) But it still looks and sound very much like the rest of this series — on using AI for Research, Analysis, Decision-Making, and Innovation.

Here's what I did:

1. Uploaded the previous installments.

I had four finished articles that followed a clear structure and captured my tone and writing style. Because I'd already done the work of establishing a clear tone of voice and point of view, I could upload nearly 5,000 words of text for Claude to use as inspiration.

2. Outlined my point of view.

Using a few Claude-provided prompts, I shared my thoughts on using AI to create. Now, Claude can write about a more general topic (AI-generated content) through the lens of a particular point of view.

3. Reviewed and revised.

I find that Claude is currently the best frontier model for matching tone and style, so it was no surprise that its first draft was pretty close to what I might write — Turning a couple paragraphs into nearly 800 words.  

Here you can see a Claude artifact with a summary of its work on the left and the draft it created on the right:

Here's the thing: I could save myself a lot of time by just using AI to write this newsletter. But I know that I (and you, my readers) would suffer for it. I wouldn't have to wrestle with difficult questions like When should we use AI? and How do we keep up with exponential change? And those shortcuts — that lack of struggle — would show.

We're not just aiming to create more, but to create more of what matters: what resonates, what educates, what convinces, what works.

So use AI to create. But don't lose sight of the humanity you bring to your work. The world still needs your wisdom, your experience, and your voice.