I asked Claude for twenty subject lines for this email.

Nineteen of them were wrong.

Here are five of them, so you can see what I mean:

Unlock the Power of AI-Driven Marketing Strategy

5 Ways to Level Up Your AI Marketing

The Missing Skill in Everyone's AI Workflow

Is Your AI Strategy Actually a Strategy?

What AI Still Cannot Do for Your Marketing

The first could be any company selling anything.

The second promises a numbered recipe I don't write.

The third has a curiosity gap with nothing behind it, which is the thing I dislike most about AI-generated headlines.

The fourth is closer, but it interrogates you instead of offering you something.

The fifth one is at the top of this email.

It didn't arrive that way. It arrived as What AI Can't Do, and I added the rest. "Still," because I think this is a current limit rather than a permanent one, and I'd rather not claim more than I know. "For your marketing," because the sentence needed to point at your work instead of at the machine.

Knowing why four of them are wrong is the essence of my job.

Hundreds of millions of people are using AI every day.
We have conversations with ChatGPT and Claude. We brainstorm, research, rewrite, generate images and ideas.

That's using AI, but it's a long way from using AI to communicate something that matters.

Generating options was never the hard part. Ask for twenty headlines, and you'll have them instantly. But which one is good? That's where human judgment comes in.

There is no "Make It Good" button.
Somebody still has to know what you're trying to say, who you're saying it to, and what sounds like you.

AI cannot tell whether or not the thing is any good.

What that looks like in practice is a human-directed process that isn't tidy:

Generate → Discern → Redirect → Refine → Implement

With a lot of "no, that's not it, let's come at this another way" in between.

Ann Handley's quote above is asking a question almost nobody in AI training is asking: Why do we teach people how to use these tools without ever teaching them when not to?

Take email newsletters.

Anyone can ask ChatGPT to write one.

That's very different from developing the editorial direction, the voice, and the design. And then actually producing and publishing something worth reading, week after week.

That's some of what I'm doing for clients right now. Developing, producing, and publishing newsletters on an ongoing basis, with AI folded intelligently into the work rather than substituting for judgment.

A similar approach is how my app, Life By Decade, came into being.
​Life By Decade is a personal history timeline app that I've been building. It helps people to place the moments that shaped their life story onto a visual timeline, a legacy document that allows them to place their own photographs and stories in a narrative arc.

The response from early users has been genuinely exciting.

What's relevant here is that I didn't suddenly become a software engineer.

AI let me combine development, design, writing, and experimentation in ways that would have taken a much larger team not long ago.

Here's the paradox.
AI accelerates the creative process. It doesn't remove it.

The more possibilities AI hands you, the more valuable your discernment becomes.

Twenty options are only useful to someone who can tell the difference between them.

The payoff comes when human judgment stays in the loop the whole way through—recognizing what's good, redirecting what isn't, refining the whole mess into something polished, clear, and effective.

If you want a newsletter like this one, that's what I'm here for. Editorial direction, writing, voice alignment, design, production, and publishing. Decades of communications craft, with AI folded in where it earns its place.

The email you're reading is a fair sample of the output.

If that sounds good to you, I'd love to talk. And if you've just got a subject line you can't decide on, send me the twenty. I'll tell you which one I'd pick and why.

Any questions?

— Jon

Use AI well. Stay human. Build what matters.

P.S. I've been putting what I learned building Life By Decade, plus what has become years of experimenting with AI images, into a new approach to website design and development. If you're thinking about a new website, or one that needs a rethink, please ask me for more details.