Something happened to the internet around October 2024. You could feel it if you were paying attention. Blog posts started sounding the same. LinkedIn feeds filled up with articles that hit every keyword, ticked every SEO box, and said absolutely nothing. The AI content floodgates opened, and the water is rising fast.
ChatGPT crossed 100 million weekly users last year. Claude, Gemini, Copilot, and a dozen others followed. Suddenly every business with a WordPress login and a monthly subscription could publish five blog posts a week. So they did. And the result is a content market where volume has exploded but quality has collapsed off a cliff.
Here's the problem nobody wants to acknowledge: if everyone's using the same tools with the same prompts, everyone produces the same output. I've seen it across industries. A managed print provider's blog reads identically to a recruitment firm's blog reads identically to an IT consultancy's blog. Same structure. Same tone. Same non-committal advice wrapped in the same polished sentences. Readers aren't stupid. They can smell it.
The maths seems obvious. If one blog post a week generates X traffic, then five blog posts a week should generate 5X. That's how most SMEs are thinking about AI content right now, and it's wrong.
Google's helpful content updates throughout 2024 made this explicit. Pages that exist purely to rank, that add nothing a human expert wouldn't already know, get pushed down. Not penalised exactly, just quietly deprioritised. The sites publishing three mediocre AI articles a day are discovering that their traffic actually declined compared to when they published one decent human article a week.
I ran the numbers on a client's competitor earlier this year. They'd gone from publishing twice a month to publishing daily in November 2024. Their domain authority hadn't changed. Their backlink profile hadn't grown. But their indexed page count had tripled. Six months later, their organic traffic was down 12%. More content, worse results. That's the volume trap.
None of this means AI is useless for content. Far from it. The businesses getting this right are using AI as a research assistant and structural tool, not as a writer.
Here's what works. Using AI to analyse a topic and identify angles you hadn't considered. Feeding it three competitor articles and asking what they all missed. Getting it to outline a piece based on your rough notes, then rewriting every section in your own voice. Using it to check facts, find statistics, suggest sources. Running your draft through it for structural feedback: is this paragraph in the wrong place, does this argument hold together, have I repeated myself.
What doesn't work is hitting "generate" and publishing whatever comes out. Even with detailed prompts, AI writing has tells. It reaches for the safe middle ground. It hedges where a human would commit. It uses transition phrases that no actual person would say out loud. "Furthermore", "it is important to note that", "in conclusion". Real people don't talk like that.
I tested this with a client's audience last quarter. We sent two versions of an email campaign to matched segments. One drafted by AI with light editing. One written from scratch by a human who knows the industry. The human-written version had a 34% higher reply rate. Not open rate, reply rate. People could feel the difference even if they couldn't articulate it.
The real competitive advantage in 2025 isn't publishing speed. It's voice. Your voice, specifically.
Every AI tool is trained on the same internet. It produces content that sounds like the average of everything ever written online. That's useful as a starting point, terrible as a finished product. When your blog sounds like every other blog, you've become invisible despite being technically present.
The businesses winning the content war right now are the ones with a recognisable voice. An opinion. A perspective that comes from actually doing the work, not summarising what others have written about doing the work. A seal manufacturer who explains the specific failure modes of nitrile vs EPDM in chemical environments is more valuable than an AI article about "choosing the right seals for your application." A software company that describes the exact moment a warehouse manager realises their picking system is costing them 30 minutes per order beats any generic "warehouse efficiency tips" article.
That specificity, that lived experience, is the one thing AI can't generate. It can only remix what already exists.
So here's what I actually recommend to clients. It's boring, but it works.
Start with your own knowledge. What questions do your customers ask you repeatedly? What mistakes do you see prospects making before they find you? What's changed in your industry this year that most people haven't caught up with? Write those answers down in your own words. Don't worry about polish. Record a voice note if writing feels unnatural.
Then use AI to structure it. Take your rough thoughts and get AI to organise them into a logical flow. Let it suggest where you need supporting data. Let it identify gaps in your argument. Let it check your spelling and grammar.
Then rewrite the AI's structure in your voice. This is the step most people skip, and it's the only one that matters. Read each section and ask: would I actually say this to a client over coffee? If the answer is no, rewrite it until the answer is yes.
The whole process takes longer than pure AI generation. A good article might take two hours instead of twenty minutes. But one article that sounds like you, that contains genuine insight from your experience, that readers remember and share, is worth more than fifty articles that sound like everyone else's.
The AI content arms race rewards the wrong behaviour. It rewards volume over value, speed over substance, presence over memorability. If you're competing on volume, you're competing against every business with an AI subscription, which is all of them.
The alternative is competing on something AI can't replicate: your actual expertise, delivered in your actual voice, backed by your actual experience. That's what a proper content strategy looks like. Not a content calendar filled with AI drafts, but a publishing rhythm built around what you genuinely know that your competitors don't.
We build content strategies for clients that start with their expertise and work outward. What do you know? What have you learned? What would your best customer wish they'd known sooner? Those questions produce better content than any prompt.
The businesses that figure this out in 2025 will own their space online. The ones that keep feeding the AI content machine will keep wondering why nobody's reading.
Martin Dugan, AA2
