AI Content Writers, SEO Rankings, LLM Models…. What’s the Real Story?

Thinking Aboout Using AI Content Writers to Optimise Your SEO… Read This First!

In the age of AI generated everything, content marketing has found itself at a crossroads.

On one side, tools like ChatGPT, Jasper, and Claude are churning out blogs, landing pages, and product descriptions at lightning speed. On the other, SEO experts are scratching their heads trying to decode Google’s evolving stance on machine made content. Is using an AI content writer the secret weapon for scale or a fast track to search engine purgatory?

We’ve heard every argument:

• “Google hates AI content, it’s a penalty waiting to happen.”
• “AI content ranks just fine if it’s helpful and edited.”
• “You can’t build trust with robotic writing.”
• “Humans can’t compete with the volume and velocity of LLMs.”

This guide was created to cut through the noise. No hype. No hand waving. Just clear, brutally honest insights into what’s working, what’s not, and how to use LLMs in your SEO strategy without tanking your rankings or your brand credibility.

At Soto Group, we’ve experimented with both extremes: human only content and LLM assisted workflows. What we’ve learned is this: you don’t have to pick sides. There’s no need to pit “traditional SEO” against AI optimisation. When paired strategically, the two approaches can serve different user behaviours from the old school Googlers to the AI native generation trusting ChatGPT to just know.

We’ve seen the good, the bad, and the outright cringe that an AI content writer can deliver. The goal? To arm you with the knowledge to scale content responsibly, balance automation with authenticity, and prepare for what’s coming next in search.

Because this isn’t just about SEO. It’s about how we communicate, compete, and create trust in a world where everyone has the same tools, but not the same judgment.

The SEO Landscape in 2025

What’s Actually Changed?

The search landscape in 2025 might be seen as evolving, but in reality, it’s fragmenting. For years, the playbook was relatively clear: write helpful content, use the right keywords, optimise your meta data, build some backlinks, and Google might reward you. But today, marketers are navigating a much more complex search environment, shaped by two dominant shifts:

• Google’s changing algorithms and evolving stance on AI-generated content
• The rise of AI-powered search behaviours (think: ChatGPT, Perplexity, and Gemini)

The “Helpful Content” Era Is Still Here

But Harder to Game

Since 2022, Google has doubled down on its “Helpful Content” updates. Algorithms specifically designed to weed out content that is thin, unoriginal, or written primarily for ranking purposes. By 2025, these systems are now deeply embedded into Google’s core ranking infrastructure.

What does that mean for AI generated content?

• AI content isn’t banned, but it must feel indistinguishable from well written human content: informative, specific, and based on lived experience.
• Google has made it clear they don’t care who creates the content (human or machine) but they do care if it’s genuinely useful.
• That means generic, surface level LLM output is a red flag, especially if it lacks first hand expertise, unique insights, or clear signals of trust.

🤖 AI Is a New Search Layer

(and it’s a different beast)

This is where many brands get caught out. While obsessing over traditional SEO tactics like keywords, on page optimisation or backlinks they miss a crucial shift: AI platforms like ChatGPT and Gemini are now major discovery layers that prioritise different things.

These platforms are trained to surface authoritative, structured, human readable content that feels “summarisable.” It’s goes beyond ranking, it’s about being referenced by the AI itself.

Take one of our recent clients:
They were using an AI content writer and on first appearance they had done the right things for SEO like steady content cadence, basic keyword research, some effort in meta tags. But they were surprised with the result…. middling rankings. Which is expected for a competitive industry – but where they really lost visibility was outside Google no visibility in AI summaries, no featured snippet presence, and no inclusion in “Perspectives” or People Also Ask sections.

Why?
Because their AI content writer wasn’t asked to write for it’s own learning and consumption models, the content wasn’t structured clearly for either human skimmers or AI parsers. No clear headings, no authoritative tone, no internal linking and nothing that made it easy for an LLM to say, “this is the best summary to include.”

The fix wasn’t about writing more content. It was about:

• Reframing structure
• Answering questions more directly
• Linking related insights
• Including topical expertise signals the AI could pick up on

E-E-A-T Still Matters

(But It’s Harder to Fake)

Google’s emphasis on Experience, Expertise, Authoritativeness, and Trustworthiness (E-E-A-T) is more important than ever and now, LLMs are inferring E-E-A-T too.

You can’t fluff your way into trust anymore, whether writing your own original content or using an AI content writer:

• Real experience (e.g. first-hand case studies, behind-the-scenes insights)
• Named authors with credible bios
• Specifics over generalities (e.g. “how we increased conversions by 34% in 60 days” beats “optimise your landing page”)

Content Overload Means Rankings Are More Competitive Than Ever

Having an AI content writer has removed the barrier to entry and anyone can generate 10,000 words a week now. But most of it is noise.

In this reality, the winners are those who:

• Write less but more strategically
• Layer human insight onto AI drafts
• Build content ecosystems, not just blog libraries

AI Content Writers: What Works, What Doesn’t

Using an AI content writer isn’t inherently bad. It’s just… incomplete.

Large Language Models (LLMs) are incredibly good at certain types of content creation especially when it comes to speed, scale, and consistency. But they also have clear blind spots, and knowing where those boundaries are is crucial to producing content that performs and actually builds trust.

This isn’t a debate about whether AI can write. It’s about whether it can communicate in a way that connects, convinces, and converts. In our experience working across industries (from care platforms to creative tech startups) the answer is: sometimes, but never without oversight.

✅ What AI Content Writers Do Well

(When Prompted Properly)

Explainers & How Tos (Low-Stakes Topics)

If a client is writing a “What is X?” piece, say, “What is trauma informed care?”, an AI content writer can get you 70% of the way with a few clear prompts. It’ll cover the textbook definition, throw in a few bullet points, and structure it neatly. Pretty cool right?

Give an AI content writer a proper prompt and you will see all the difference. We once had a client in the sustainability space who needed foundational content on carbon offsetting. The AI content writer generated a draft that was accurate, but pretty bland. However, once we added a real example of how a small Aussie brand switched from certificates to land stewardship suddenly it mattered.

Takeaway: Use an AI content writer to get the bones, but insert your client stories and quirks to make it memorable.

Great for generating introductory or foundational content. Think:

• “What is ESG reporting?”
• “How to install a rainwater tank”
• “The basics of CRM systems”

These pieces work well because:

• LLMs have been trained on similar structures a thousand times.
• The risk of hallucination is low when sticking to factual, non nuanced material.
• With some human editing, these are fast, efficient, and scalable.

Structured Content at Scale

One of our clients needed 90 category page descriptions written for a product migration. Writing each from scratch was impractical. So we templated tone, key phrases, and a logic tree and let an AI content writer do the grunt work. Editing took half the time of writing from zero.

Pro move: Don’t feed your AI content writer a brief. Feed it a repeatable pattern. That’s where it shines.

An AI content writer is fantastic at creating templated formats, which can make your content simple to consume:

• Product descriptions
• FAQs
• Category pages
• Summaries of similar service offerings

When style and structure are consistent, you can feed a base template + key inputs and get 80% of the work done in seconds.

📣 AI Content Writer Ecosystem Support

Got a long form blog ready? An AI content writer is brilliant at repurposing:

• A three sentence email teaser
• Meta descriptions for SEO
• A carousel caption for LinkedIn

But again, without a human steering tone and purpose, these assets can feel like they’ve been… assembled.

We’ve found that AI content writers can mimic brevity, but it doesn’t always understand what’s worth saying short. Humans still know what matters.

Supporting Assets for Primary Content
Once you’ve created a main blog or article, your AI content writer can help spin off:

• Social captions
• Meta descriptions
• Email teasers
• Alternate intros or CTAs

It excels here as a support tool, not the lead author.

🚩 Where AI Content Writers Start to Struggle

Actual Thought Leadership
Anything that requires genuine opinion, lived experience, or controversial takes quickly falls apart:

• Executive POV articles
• Industry predictions
• Content that leans on “we’ve actually done this” credibility

These often sound generic or hedge their bets too much, the opposite of what builds trust.

We ran a test where an LLM generated a LinkedIn post for a founder announcing a partnership. It had all the right words but none of the voice, humility, or friction that made the backstory meaningful.

It’s the difference between saying “We’re excited to scale impact” vs “This nearly didn’t happen. We almost walked away twice and I want to tell you why”

AI content writers avoid discomfort. Humans lead with it.

Strategic Content or Complex Advice

An AI content writer can simulate depth, but often:

• It misses context
• Overuses buzzwords
• Gets basic facts right but nuances wrong

For example, it might say “optimise your sales funnel” but won’t account for your ICP’s buying behaviour, internal politics, or channel mix without major hand holding.

An AI content writer might suggest “build a brand community” as part of a growth strategy, which is fine, except the client has three FTEs and no customer list. It doesn’t know context unless you train it heavily.

In one case, we watched AI confidently suggest “run webinars” to a palliative care SaaS startup whose audience was 85% aged care providers… on dial-up connections. We had a very human laugh at that.

Localised or Culturally Aware Writing

We’ve caught an AI content writer throwing “Mom” in Australian case studies, referencing Thanksgiving in October, and using “Z” spellings in federal government comms.

If your brand sounds like it’s outsourced its voice to another continent, it loses credibility…. instantly.

• LLMs often default to American spelling, idioms, or case law.
• They struggle with tone matching across diverse cultural or regional audiences.
• Subtle signals like slang, humour, or implied context can fall flat or even backfire.

Reality Check: How to Tell Good AI Content from Bad

Even if content sounds polished, here’s how to spot whether it’s doing its job:

The Sweet Spot: AI as Draft Partner, Human as Editor in Chief

The strongest content teams we’ve worked with treat an AI content writer like a junior employee: great for getting something on the page, but nowhere near ready to publish solo. They are new after all, they need training and usually that training includes all the cultural and experience based nuances that only come with practical action. Our recommendation? Use an AI content writer to start. Use humans to shape. Use experience to sharpen.

The most effective content workflows in 2025 are hybrid:

• Use your AI content writer to accelerate first drafts, formatting, or ideation.
• Then let humans refine, challenge, and contextualise the result.
• The end product? Content that’s efficient to produce and meaningful to read.

Warts and All (The Real Risks of AI Content)

Let’s be honest: most conversations about using an AI content writer are either breathless or blindly optimistic. What’s often missing is a clear look at what happens when the output of an AI content writer goes unchecked. The reputational risks, brand consequences, and SEO sinkholes that businesses only discover once it’s too late.

We’ve seen the following mistakes firsthand, either through rushed internal experiments, panicked audits, or calls that start with, “Can you take a quick look at why our traffic just halved?”

🚫 Risk #1: Being Demoted by Search Engines

Google has been abundantly clear: it doesn’t care if content is written by humans or AI it cares whether it’s helpful, original, and aligned with user intent.

Yet in practice, many AI content writer generated articles:

• Overuse keywords unnaturally
• Repeat ideas from top ranking content without adding anything new
• Fail to match the intent of the search query

We’ve reviewed AI generated content that ranked for the wrong terms entirely, not because it was badly written, but because it never understood what the user was actually looking for.

In competitive industries, that can mean:

• Lower rankings
• Ineligibility for featured snippets
• And being completely ignored by AI overviews in the SERP

😬 Risk #2: Hallucinations, Outdated Info & Legal Liability

Your AI content writer is only as good as the data in its language model. LLMs are trained on information up to a certain date (or, in some cases, with limited real time context). That means:

• Stats might be outdated.
• Laws or guidelines might be misrepresented.
• And worst of all your AI content writer will confidently make things up when it doesn’t know the answer.
  • Sometimes you’ll even believe something that is outright fiction

We’ve seen an LLM invent entire case studies, fabricate citations, and attribute quotes to authors who’ve never said them. If a client hits “publish” without catching that? It’s bad form and a legal risk.

🧱 Risk #3: Erosion of Brand Voice and Trust

This one’s subtler, but just as serious.

When an AI content writer is used to create mass content, and the style and tone of voice aren’t checked or provided – this dominates your output, and your brand starts to sound like everyone else. Even if readers can’t articulate why, they’ll feel the blandness.

One client’s blog metrics looked fine on paper (bounce rate, time-on-page), but interviews revealed that customers had stopped reading altogether. “It just doesn’t sound like you anymore.”

Without a strong editorial voice, your content becomes wallpaper (visible, maybe, but forgettable).

💸 Risk #4: False Sense of Efficiency = Wasted Budget

Here’s the paradox: using an AI content writer may feel like you’re producing more, faster but if the content isn’t performing, converting, or ranking, you’re just publishing faster into the void.

Worse, you can end up:

• Burning hours editing unusable drafts
• Paying for volume when what you needed was precision
• Or investing in content that never reflects your strategic POV

We had one client who created over 50 AI assisted articles in a quarter and not one ranked in the top 20. After auditing, we found most lacked structure, original data, or semantic relevance. They could’ve written five good ones and achieved better results.

🧨 Risk #5: Public Embarrassment (or Internal Tension)

In the worst cases, your AI content writer not only underperforms it creates serious damage:

• A blog post that misrepresents your stance on a sensitive topic
• A team wide email generated by AI that lands cold or tone deaf
• An investor update that sounds robotic and disconnected from reality

We’ve seen content teams quietly undo weeks of AI work after internal reviews flagged tone, inclusivity, or even regional relevance issues. It can create trust fractures inside your team especially if one side is pushing for automation, and the other is begging for human nuance.

What is worse, is that we expect an AI content writer to sound like an expert with 50 years experience in our business or company when in actual fact, they are the intern finding their feet. If you don’t keep them in check, they’ll go rouge. If you don’t nurture them, they’ll get lazy. If you don’t performance manage them, they’ll start to poison others around them.

🗣️ How You Talk to AI Matters

(Better Prompts, Better Results, Better Partnerships)

Most AI generated content flops because it was poorly briefed. That’s right, it is not AI’s fault. It’s yours…..

When teams treat LLMs like vending machines (type prompt, hit “go”), they get surface level results. But when they brief with context, constraints, and clarity just like they would a freelancer or strategist, the output becomes exponentially more valuable.

Here’s how to shift from “meh” to “magic.”

1. Describe the Purpose, Not Just the Output

Poor prompt:

“Write a blog about sustainable packaging.”

Stronger prompt:

“Write a blog post that helps FMCG brand managers understand how new packaging reforms will affect their 2025 compliance targets. Include real-world examples, risks of inaction, and a comparison table of EPR models.”

Why it works: It gives the AI content writer a goal (educate brand managers), a context (2025 reforms), and a structure (examples + comparison) which leads to content that actually helps someone, not just fills a page.

🧩 2. Give the AI Content Writer Your Lens

If you want content to sound like you, give it your positioning.

Prompt add on examples:

• “Our brand voice is confident but never arrogant, speak plainly and skip filler.”
• “We challenge common wisdom and offer first-hand insights from consulting projects.”
• “Use UK English, not American spelling, and avoid buzzwords like ‘cutting-edge’ or ‘game-changer’.”

💬 Your voice is an asset, don’t make the AI guess it.

📚 3. Use “Meta Prompts” to Guide Structure and Strategy

LLMs are fantastic at thinking about thinking. You can use that to your advantage by asking it to work like a strategist, not just a writer.

Try:

• “Before writing the article, list five different angles this topic could take and tell me which aligns best with a senior decision maker audience.”
• “Suggest a structure for this blog that would maximise its chance of being used in an AI Overview or featured snippet.”
• “What follow up content pieces could be created from this topic to build a cluster?”

This helps shift the AI’s role from “doer” to collaborator, which is where its power truly lies.

4. Ask for Its Opinion

(Yes, Really)

Most users don’t realise that an AI content writer can critically evaluate its own output. So if you are thinking about how to optimise your AI content writer:

Try:

• “What’s missing from this article if I want to appear as a subject matter expert?”
• “If you were a sceptical reader, what would make you mistrust this page?”
• “What edits would you recommend to make this align more closely with E-E-A-T principles?”

The goal is not to blindly trust AI’s answer, it’s to pressure test your work from a different angle, and perhaps learn a thing or two along the way.

5. Let the AI Content Writer Push Back

(When Appropriate)

Some of the best collaborations happen when AI is told not to be a yes-man.

Prompt examples:

• “Challenge my assumptions in this draft, what’s something my audience might disagree with?”
• “Where might I be oversimplifying this topic?”
• “Play devil’s advocate: how would a competitor position against this article?”

This technique is especially useful for positioning work, brand strategy, or rebuttal planning.

The AI Is Only as Good as the User’s Imagination

LLMs can do way more than just write they think, strategise, structure, and stress test. But only if you ask them to.

When businesses learn how to use an AI content writer as a layered thinking partner, rather than a glorified rewriter, they unlock its real potential. That’s the shift from “generate content” to create competitive advantage. The cincher is, this entire article was written by ChatGPT. But it certainly didn’t get to this stage of insight and depth without plenty of back and forth.

So while all of this context is great (and perhaps you have learned a thing or two) the real value (and what can be enthralling) is watching a layered, intellectual conversation between a human and AI content writer. What does that look like you ask? Let’s peek behind the scenes…. here is an snapshot of how this article came to life.

ChatGPT said: