Y​our AI content sounds fine. That's the problem.

A client forwarded me a batch of blog posts her team had been producing with AI assistance.

They were good. Genuinely.

Well-organized, clean sentences, no obvious errors. Her marketing manager had spent real time on them. You could tell.

I read through all six and couldn't find a single sentence that sounded like her company.

Not one. In six posts. About a brand I'd written for before and knew pretty well.

That's not a quality problem. That's a voice problem. And it's the AI content issue nobody is talking about — because the content looks fine on the surface, and "it doesn't quite sound like us" is a hard thing to bring to your CMO when the output is technically competent and the time savings are real.

What's actually happening

AI is extraordinary at producing coherent, readable prose. It has read essentially everything — every industry blog, every brand guide template, every thought leadership piece ever posted on LinkedIn. It knows how companies are supposed to sound.

Which is exactly the problem.

When you ask AI to write content for your brand, it defaults to how brands in your category generally sound.

Competent. Professional. Indistinguishable.

It's not being lazy. It's doing exactly what it was trained to do. You just haven't told it enough about what makes you not like everyone else.

The symptom shows up subtly at first. Your team produces content faster — great. But over a few months, something starts to feel slightly off. The blog posts sound a little flat. The social captions feel generic. A colleague reads something and says "did we write this?" and nobody's quite sure.

That's the drift. And it compounds quietly until suddenly your brand sounds like a category, not a company.

The fix most people try (and why it doesn't work)

The standard advice is to give AI your brand voice guide and tell it to follow it.

I've tried this. Most brand voice guides are written for human writers — they say things like "we're conversational but professional" and "we use an active voice" and "we don't take ourselves too seriously." That's useful guidance for a person. AI reads it and produces... slightly more personalized generic content.

The other thing people try: prompting harder. Adding more adjectives. Telling AI to "write like a real person" or "avoid corporate language."

This helps a little. But you're still working around the core issue, which is that AI needs examples, not descriptions. It needs to see the specific sentences you'd actually approve, understand the patterns in your best existing content, and have that translated into prompting language it can actually use.

That translation is the part that requires a human copywriter who knows what they're looking at.

What a voice-calibrated prompt system actually is

I want to be specific here because "custom prompts" has become one of those phrases that sounds meaningful and could mean anything.

What I build for clients is a set of prompts — usually five to eight, organized by content type — that have been constructed from the inside out. Starting with your actual content. The pieces that got shared. The copy your team reads and thinks yes, that's us. The emails that got replies.

I take those, pull out the patterns (sentence length, how you handle transitions, what you do with technical language, how you pace a paragraph, when you use humor and when you don't), and translate those patterns into prompting language Claude can work with.

Then I test them. Against real content your team actually needs. I run drafts, compare them to your existing work, adjust, run them again. The system doesn't get handed over until it's producing drafts that actually sound like you — not like a brand voice guide describing you.

The reference doc that comes with it isn't a style guide. It's a working document your team can use to brief AI on any new content type, written in the specific format that actually tells AI what it needs to know.

Where this fits in your current workflow

This isn't a replacement for your content team. It's what makes your content team faster without losing the thing that makes your brand recognizable.

Your writers stop spending 40% of their time fixing AI output that's technically fine but subtly wrong. Your social captions stop going through three rounds of revisions because the first draft didn't sound like you. Your blog posts stop reading like they were written by someone who researched your company for an afternoon.

The drafts become a real starting point instead of a rewrite.

That's what I hear most consistently from people after this process: not that the content got better, but that the friction got lower. The AI output started feeling like something they could actually use, instead of something they had to fix.

What this engagement looks like

It's a two-week project. Scoped, contained, with one clear deliverable at the end.

Week one: I do a voice intake — I go through your best existing content, ask you some specific questions about your audience and the content types you produce most often, and identify the patterns that make your writing yours.

Week two: I build and test the prompt system. You see drafts, give feedback, I refine. At the end, you have a working system and a reference document.

Investment starts at $950.

This isn't a monthly retainer. It's not a "let's have an exploratory call about your content strategy." It's a specific thing that solves a specific problem — and at the end of two weeks, you have something real to use.

Who this is (and isn't) for

This works best for brands that are already producing AI-assisted content and noticing the drift. You've got the workflow. You've got team buy-in. Something just feels slightly off and you want to fix it properly instead of prompting harder and hoping.

It's not the right fit if you haven't started using AI for content yet — that's a different conversation.

And it's not for teams that want a consultant to overhaul their entire content operation. This is a targeted fix for a specific problem. The scope is intentionally tight.

If your content has started to sound like everyone else's content, and you've been quietly wondering if AI is the culprit — it probably is.

And that's fixable.

Book a 20-minute call or comment below.