AI Customer Personas for Content, Built From the Catalog (Simply Explained)

Most Customer Profiles Are Made-Up Stories

Here is how most companies decide who their customer is. Someone sits in a meeting and invents a person. "Busy Brenda, 42, loves yoga, drinks oat milk, never has time for herself."

They slap a stock photo on her. They write a few paragraphs about her imaginary life. They pin it to the wall and never look at it again.

Brenda is fiction. Somebody made her up before a single real customer ever placed an order. She isn't connected to anything the company actually sells. And because she's connected to nothing, she gets used for nothing.

I think that's completely backwards.

So when I built customer profiles for a health and wellness brand (supplements, lab tests, that kind of thing), I didn't start by inventing people. I started with the only thing that was real: the products on the shelf.

Build the Customer From the Products, Not the Other Way Around

Here's the simple idea. The brand already sold real things. Each product solved a specific problem for a specific person.

A sleep product. A metabolism test. A recovery formula. I didn't have to guess who the buyer was. The buyer was obvious from what each product fixed.

So I worked backwards. Instead of dreaming up a customer and hoping the products fit, I looked at what the brand actually sold and figured out who buys each thing and why.

The products came first. The people came second.

This matters more than it sounds. When you ask AI to "write a blog post about better sleep," you get bland mush that speaks to nobody. When you tell it to write for a specific person, with a specific worry, holding a specific product, you get writing that actually connects.

Same AI. The difference is whether you gave it a real target to aim at.

I Built 40 Customer Profiles, Each Tied to Real Products

Think of it like hiring a team of specialists. I set up a small group of AI assistants, and gave each one a single job.

One only worked on sleep products. One only worked on metabolism. One only handled recovery. Each assistant could only see the real products in its category, so it couldn't wander off into fantasy. It had to build its customer profiles on top of products that actually existed and money that actually came in.

I ended up with 40 customer profiles.

But these aren't vague sketches. Each profile is a real record, like a detailed file on a customer. Here's what each one contains:

• A short life story. Who this person is and what brought them to the problem.
• What they care about. Their values, fears, how they think about their own health.
• Their real objections. "Supplements are a scam." "I don't have time for lab work." "I've been burned before."
• What makes them buy. A bad night of sleep. A scary test result. Turning 40.
• The exact products they'd buy, pulled straight from the real catalog.
• The lab tests that make sense for them.
• How to talk to them without sounding like a brochure.

Every piece earns its place. The objections become the questions we answer in an article. The buying triggers become headlines. The products give the article something concrete to recommend instead of vague advice.

A profile without all that is just decoration. A profile with it is a real instruction sheet the writing system can follow.

Pointing the Writing Machine at the Right People

Building 40 great customer profiles is useless if your content doesn't actually use them. So I wired them straight into the system that writes the articles.

Here's how it works. For every topic, the system first picks one to three of the 40 profiles to write for. Not "general audience." Specific people.

A sleep article might target the burned-out parent, the data-obsessed health nerd, and the person who has tried everything and given up.

Then it saves which profiles each article was written for. Most teams skip this. But it means I can pull up any published article and see exactly who it was meant to reach.

After that, the system writes directly for those people. The article opens with their real worry, in their words. It recommends the products that actually fit them, not a random pile. It sounds like it was written for a specific human, because it was.

And because I store which profiles each article targets, I can do something most marketing teams can't. I can see which customer types produce articles that actually sell, and which ones don't. That's a real signal I can act on, not a hunch.

Letting the Team See It (and Trust It)

Customer profiles sitting in a computer file are useless if nobody on the team ever looks at them.

So I built a simple tool where the team can browse all 40 profiles. They can read the stories, see the products, scan the objections.

That sounds small. It wasn't. Before that tool, the profiles lived hidden inside the system, and the people didn't trust what they couldn't see. After it, the marketing team started using the profiles to plan campaigns and to double-check what the system produced.

Here's the honest part. This only works because humans still check the machine. The AI does the volume. The team does the judgment. That tool is what connects the two.

One Trap I Watch Very Closely

This is the part that keeps me up at night, and it should worry you too if you're in a regulated industry.

Every profile is clearly labeled as made-up. The face in the photo is generated, not a real person. The story is a type of customer, not a real case study.

And here's the rule I never break: these profiles never show up in published content pretending to be real customers.

In health and wellness, that's not optional. If you turn a made-up customer into something that reads like a real patient saying "this product changed my life," the government's rules on fake reviews can land you a fine. They don't care that you "meant it as a profile."

So the line is clean. The profiles guide what we write. They never appear in what we publish as real people. They live behind the scenes. The reader never meets them.

That boring discipline is exactly what gets skipped when someone bolts AI onto a marketing team without thinking it through.

How to Make AI Content Speak to Real Buyers

Every business owner eventually asks me the same thing. How do I get AI content to actually speak to my buyers instead of cranking out generic filler?

The answer is targeting. And good targeting needs three things. It has to be built from your real products and real customer objections. It has to be written down as data, not stuck in someone's head. And it has to be chosen per article, so each piece speaks to specific people instead of an imaginary average.

This works for almost any business with a real product or service. A supplement brand. A financial advisory firm. A manufacturer with a product line. If you sell something specific to someone, you can build the customer profile backwards from it.

The made-up approach (invent the buyer, hope the products fit) is the only version that doesn't work, because it was never built on anything real.

If your content sounds like it could belong to any company in your industry, that's the tell. That's the thing real targeting fixes.

Want to explore what AI could do for your business?

Book a free 30-minute strategy call. No pitch deck, no sales team, just a real conversation about your operations and where AI fits.

Book a Discovery Call