From Zero to Cited, Part 6: Writing Articles With AI Without Sounding Like AI

Here's the thing about AI-drafted articles: I've read plenty that never should have gone live. Flat voice, no opinion in sight, the same five transition words doing all the heavy lifting like tired pack mules. You've probably read a few too, even if you couldn't say exactly why they felt off. Something in you just knew.

This is Part 6 of 10 in the From Zero to Cited series, and it's the one people ask about most: how to write SEO articles with AI without the result reading like it rolled off an assembly line that's never had a bad Monday. Short answer: you can. Long answer takes the rest of this post.

A Quick Note on Landing Pages Before Articles

Before we get into articles, one thing worth settling first: not every keyword deserves a blog post. Trust me on this one.

If someone's searching "best invoicing software for freelancers" or "buy project management tool," they're already halfway to the checkout page in their head. Hand them an article and you've made them work for something a landing page could just hand them.

Transactional keywords want a page built to convert: a comparison table, a signup button, pricing laid out plain. Informational keywords, the "how does this work" and "why does this happen" kind, are where articles earn their keep. Mix the two up and you quietly waste good keyword research. Google's own guidance on search intent (through Search Central) backs this up: pages that don't match what the searcher came to do rarely hold their rankings for long.

The Honest Case Against Raw AI Content

If you've seen sites lose rankings after publishing AI content, you weren't imagining it. It happened, and it's still happening, mostly to sites that skip editing entirely.

Shipping AI drafts untouched is the real failure here. AI-assisted writing is a different animal, and it needs a different fix. Keep the two separate in your head.

Raw AI output hedges everything, repeats its own structure like a metronome, and avoids saying anything a competitor couldn't also say. Search engines have gotten better at spotting that pattern. Google's helpful content guidance says plainly that content made mainly to game rankings, rather than help one specific reader, is exactly what they're tuned to push down. That's a quality bar, not a witch hunt, and a lot of unedited drafts don't clear it.

What matters is whether what published got rewritten into an actual brand voice, checked for AI content detection risk, and read by a human before it went live. Skip those steps and you're gambling with a domain you've spent months quietly building trust on.

Where AI Content Detection Actually Matters

Detection tools are unreliable at catching AI writing, and Google itself has said its ranking algorithm isn't running detector scripts against your pages. So this was never really about search engines catching you red-handed. It matters because readers notice generic writing, and generic writing doesn't earn links, shares, or a second visit. Treat a detection check as a proxy for "does this sound like a person wrote it," not a pass/fail exam for a penalty that isn't being handed out.

Hands-On: Draft, Voice, Humanize, Check, Publish in PostKing

Here's the workflow I'd actually run, and it's the one PostKing is built around. Five steps, in order. Skip one and the piece won't hold up once it's live.

1. Start from an approved brief. A brief spells out the angle, the intent, which internal links it needs to hit, and what "done" looks like. Most "sounds like AI" trouble starts here, before any model touches the page.

2. Draft against that brief. Let the model do the heavy lifting on structure and coverage. This part's fast, and it's supposed to come out a little rough.

3. Rewrite in your brand voice profile. Run the draft through your saved voice profile in PostKing, your phrasing, your tone rules, your no-go words, so it sounds like your team wrote it. This ensures the post is unique to your brand.

4. Run the humanize and de-slop pass. This strips the tells: the metronome sentence rhythm, the over-hedged claims, the polished-nothing energy. It's a separate pass from voice matching on purpose, voice and naturalness are different problems.

5. Run the AI-content check, then read it yourself. The automated check flags anything that still reads flat or synthetic. Then a human reads it start to finish. Read it thoroughly. If it doesn't sound like something you'd hand to a client you respect, it doesn't publish yet. No exceptions.

Only after all five steps does anything go live. The order matters more than people expect. Flip it, publish first and edit later, and you'll learn the case against raw AI content the hard way.

Do's and Don'ts of Content Creation

A few rules worth following no matter which tools are drafting for you:

Don't flood pages at once. Forty articles in a week looks suspicious to search engines, and it outpaces your own ability to check the work. Steady output beats a burst you can't quality-control.

Don't keyword-stuff. If your primary term shows up every other sentence, you've stopped optimizing and started annoying your reader, probably while tripping a spam filter at the same time.

Don't duplicate boilerplate. Ten articles with the same intro paragraph and a swapped headline is thin content in a costume. Nobody's fooled, least of all the algorithm.

Don't ship thin content. A 400-word post that restates the headline three times doesn't help anyone, whether it's a yoga studio blog or a cybersecurity shop explaining a tricky concept. It won't rank either way.

Don't let two pages compete for one keyword. That's keyword cannibalization, and it's sneakier than the rest because it disguises itself as productivity. You wrote two good articles, nice work. Now they're fighting each other for the same spot on page one, and neither wins.

Hands-On: Catch Cannibalization Before It Costs You

Run a cannibalization check before you publish, not after rankings start sliding. PostKing's cannibalization report compares your new draft's target keyword and intent against everything already live, and flags overlap before it turns into a real problem. If two pages are both quietly trying to own "employee onboarding checklist," you want to know on day one, not three months later while squinting at analytics wondering why neither ranks.

Fixing it after the fact usually means merging pages or re-targeting one of them, which eats far more time than the five-minute check up front would have.

Wondering if you actually need to reformat your content for AI crawlers? Here's what Google's own team says.

On-Page Basics: What's Automatic and What's Still Yours

Some of on-page SEO is mechanical work, and it should've been automated years ago. PostKing handles meta titles, meta descriptions, canonicals, sitemap updates, and JSON-LD structured data on every article without you opening a settings panel. That used to eat an entire afternoon per post. Now it's gone.

Judgment is still yours. Title choice needs a human eye for what'll actually get clicked, not just what fits inside a character count. Intent match, whether the page answers what the searcher actually came to do, is a call only someone who knows the business can make well. A family law office and a bike repair shop need different reads entirely. And internal linking still needs a person deciding which pages deserve to be connected, and why.

Hands-On: Internal Linking via CTAs and In-Article Links

Two places to link, two different jobs. In-article links, dropped wherever a term or concept naturally connects to something you've already published, help readers and search engines see how your content fits together. CTAs at the end of a post point toward the next action: another guide, a tool, a signup page. Build both into the draft itself, don't bolt them on after the humanize pass runs. A solid internal linking habit is one of the more overlooked pieces of learning how to write SEO articles with AI that actually hold their rankings past the first month.

Where This Leaves You

Writing with AI was never the risk. Publishing the first draft is. Once you've got a real brief, a brand voice profile, a humanize pass, an AI content detection check, and a human reading it start to finish before anything goes live, the whole thing stops feeling like a shortcut and starts looking like your actual content, just made faster.

That's the whole point of learning how to write SEO articles with AI the right way: speed without the tell. Settle the workflow once, and every article after that follows the same five steps, cannibalization check included, without you reinventing anything from scratch.

Next in the series: what happens after you hit publish, and how citation tracking tells you whether any of this actually worked.

From Zero to Cited · Part 6 of 10 · Course overview

• ← Previous: Part 5: Free Tools and Content Pages
• Next: Part 7: Getting Cited by AI Search →