[TL;DR: AI makes content creation nearly free, so every business creates more of it. But Byron Sharp's research shows brands grow through distinctiveness, and Rory Sutherland shows that effortless communication signals nothing. The businesses winning right now use AI for research and structure, not voice.]

The AI Content Trap: Why Generic Marketing Gets Ignored

Most Australian SMEs have discovered that AI tools can produce a blog post in five minutes, a newsletter in ten, a LinkedIn update in thirty seconds. The friction has gone. The volume has gone up. And quietly, without anyone making a deliberate choice, the content has started to look the same.

This isn't a productivity story. It's a brand story. And the research suggests most businesses are making a mistake they don't even know they're making.

When the Cost Drops by an Order of Magnitude, You Do More of It

In February 2026, Azeem Azhar at Exponential View published a detailed account of running a personal AI agent that he estimated produced the equivalent of $20,000–$25,000 worth of work in six days. Research reports, competitive analyses, presentations, a personalised CRM, a custom writing style analyser. The agent ran overnight while he slept.

His central concept is "the boundary of tedium": the threshold where a task is too complex to easily delegate but too boring to do yourself. AI has pushed that boundary dramatically. Tasks that once sat on the someday-maybe pile are now dispatched in minutes.

For marketers, this shift is obvious. Content that used to require half a day now requires an hour. The friction cost has collapsed by a factor of five or ten. And when any cost drops that dramatically, basic economics predicts the same outcome every time: people do more of it.

More blog posts. More newsletters. More LinkedIn updates. More "thought leadership." More content designed to signal expertise, build trust, and win search rankings.

And here is the problem: when the tool is the same, the output looks the same.

Investor Rohit Krishnan, who Azhar spoke with about this, put it plainly: ask an AI to write a four-paragraph essay worth reading and you get something "distinctly mid." It "lands in the middle of the statistical distribution. It is inoffensive and unengaging and you wouldn't choose to read it."

The mechanism is structural. Azhar analysed his own writing and found he uses roughly 80% Germanic root words: short, direct, Anglo-Saxon vocabulary. "Use" not "utilise." "Show" not "demonstrate." "Start" not "commence." The average large language model defaults to roughly 60% Latinate words, the longer, more formal vocabulary that entered English after the Norman Conquest. The result is writing that sounds polished but distant, technically competent but somehow flat.

More troubling is what Azhar calls coherence degradation. AI models have improved at the sentence level. A given sentence sounds fine. But that sentence inside a paragraph, inside a section, inside an argument, the quality "degrades at every level of zoom." The structure holds; the conviction doesn't.

For code, the evaluation is deterministic: does it run, does it produce the right output? For writing, the evaluation requires taste. And taste cannot be averaged.

The Brand Problem Nobody Sees Coming

Byron Sharp's research at the Ehrenberg-Bass Institute, compiled across 40+ years of consumer panel data from 130+ brands, identifies how brands actually grow. The answer is not differentiation in the traditional strategic sense. It is not loyalty programs, emotional connections, or superior quality perceptions.

What builds that "easy to think of" quality, what Sharp calls mental availability, is consistent use of distinctive brand assets: specific visual cues, tones of voice, phrases, characters, and patterns that trigger brand recognition across different contexts and over time. These work by repeatedly refreshing and strengthening memory structures in the minds of category buyers.

The critical word is consistent. And the enemy of consistency is the generic.

Jenni Romaniuk, Sharp's colleague and author of Building Distinctive Brand Assets, makes this precise: distinctiveness is not about being different in an abstract sense. It is about being recognisably, consistently you across every touchpoint, including every piece of content you publish.

When AI generates content calibrated to the average of everything ever written, it produces something indistinguishable from the category average. A financial planner's AI-drafted article reads like every other financial planner's article. A builder's newsletter sounds like every other builder's newsletter. The distinctiveness that Sharp shows drives brand growth gets systematically diluted.

This is not a minor problem. Sharp's double jeopardy law is unforgiving: brands with smaller market share suffer twice. They have fewer buyers, and those buyers are slightly less loyal. The only escape route is penetration: reaching more people in the category who currently don't think of you when a purchase need arises. Generic content does not open that route. Only distinctive content does.

Why Effort Is Part of the Message

Rory Sutherland, Vice Chairman of Ogilvy UK, has spent three decades arguing that advertising works through mechanisms economists don't understand. One of the most provocative is costly signaling.

His argument: the perceived investment in communication is itself a signal. A television campaign that costs $3 million to produce and $30 million to run communicates something beyond its content. The sheer scale of the commitment signals that the business believes in what it is selling. If they are willing to spend that much, they must be confident in the product. The spend is evidence.

Sutherland draws on evolutionary biology here. A flower is a weed with an advertising budget. The petals and nectar are costly to produce, and that cost is precisely what persuades bees the visit is worthwhile. Reduce the cost and you reduce the signal.

Apply this to AI content and the implication is uncomfortable. A blog post that took thirty seconds to generate signals something about how much the business believes in the argument it is making. Not consciously to readers, who cannot detect AI writing reliably, but structurally, in the content itself: the absence of a genuine point of view, the hedged conclusions, the argument that could have been made by anyone.

Sutherland is too pragmatic to argue against efficiency. His point is subtler: understand where the signal lives, and do not accidentally outsource it. The research synthesis, the factual accuracy checking, the structural outline, these can all be AI-accelerated without cost to the signal. But the voice, the specific word choices, the deliberate position-taking that reflects someone who has a genuine stake in being right: these are where the signal lives.

The Right Division of Labour

The mistake is treating AI as a writing tool. It is not, or not primarily. It is an extraordinary research, analysis, and structure tool that happens to produce text as a byproduct.

Azhar ran his agent to produce a 10,000-word analysis he described as "the best analysis I have read on" its subject. But he also notes that first-draft emails from the agent "sounded like they were fresh from an LLM," and he rewrites around 40% of them. The research and structure: excellent. The voice: requires human intervention.

The right workflow for content that actually builds a brand:

• Brief it as research first. Ask AI to synthesise everything known about the topic, surface the strongest counter-arguments, and produce a structured outline. This is where AI is genuinely extraordinary.
• Build the argument yourself. Decide what position you actually hold. What do you believe that your competitors would not say? What have you seen in your own client work that the research confirms or contradicts?
• Write the introduction and conclusion by hand. These are the anchors for brand voice. Get them right and the middle becomes easier.
• Use AI to strengthen, not generate. Once there is a genuine draft, AI can improve transitions, check consistency, and flag gaps. This is a very different task from generating from scratch.
• Never let AI set the angle. The positioning, the specific insight, the provocative take that makes someone forward an article to a colleague: these require the kind of taste that cannot be averaged.

What This Means for Your Business

Run this test on your last five pieces of content. Remove your logo and business name. Could any of them have been published by your closest competitor without anyone noticing?

If the answer is yes, those pieces are not building your brand. They may be actively eroding it, because every generic article you publish teaches the market that your voice is indistinguishable from the category average. Sharp's research suggests this is the opposite of how brand growth works.

The businesses building real brand equity right now are not the ones generating the most content. They are the ones generating content that could only have come from them. Content that takes a genuine position. Uses the specific language of their industry without drowning in jargon. References real client outcomes with real numbers. Sounds like a person who has strong opinions and is willing to be wrong.

AI can help you get there faster than ever before. The research synthesis, the competitive analysis, the draft framework: all of this is now trivially fast. But the moment the output could have come from any business in your category, you have stopped building brand equity and started creating noise.

The boundary of tedium has moved. The threshold for distinctiveness has not.

FAQ

Does AI-generated content hurt SEO?

Not directly, not yet. Google's guidance focuses on quality and helpfulness rather than how content was produced, and AI-generated content that is accurate, well-structured, and genuinely useful will still rank. The SEO problem with generic AI content is indirect. If your content provides no unique perspective, no original data, and no distinct voice, it competes only on topical coverage, an increasingly crowded space. Lily Ray and other E-E-A-T researchers are consistent on this point: search engines are rewarding demonstrated expertise and genuine authority. Content that could have been generated by anyone is, by definition, not demonstrating either. The smart approach is to use AI for research and structure, then layer genuine expertise on top. That combination is both more distinctive for brand building and more defensible for search.

How do I develop a distinctive brand voice if I'm not a natural writer?

Voice comes from opinion more than from craft. The most distinctive business content is not the most beautifully written. It is the most specifically argued. Start by listing five things you believe about your industry that most of your competitors would be reluctant to say out loud. These positions, backed by your real experience and client outcomes, are the raw material of a distinctive voice. AI can then help you articulate those positions clearly and consistently. The mistake is asking AI to generate the positions themselves. What you believe and why you believe it: this must come from you. Once those positions are clear, even plain writing becomes distinctive, because the argument itself is impossible to replicate.

What content tasks should AI handle versus what should stay human?

Use AI for anything where the measure of quality is accuracy or completeness: research synthesis, factual accuracy checks, competitive landscape summaries, keyword research, structural outlines, headline testing, and editing for consistency. These tasks benefit from AI's breadth and speed without sacrificing brand distinctiveness. Keep human judgment in anything where the measure of quality is resonance: the central argument, the specific position, the examples drawn from real client experience, the introductory hook, and the final call to action. The practical test is whether the task requires you to have an opinion. If it does, do it yourself and use AI to support. If it doesn't, accelerate it with AI and reinvest the time saved in the things that require your judgment.

Further Reading

• How Brands Grow, Byron Sharp - The empirical case for mental availability, distinctive assets, and penetration as the drivers of brand growth
• Building Distinctive Brand Assets, Jenni Romaniuk - How to measure and build brand distinctiveness systematically
• You Already Have an AI Agent, Azeem Azhar - A detailed account of AI agents in practice, including the honest assessment of where they fall short on writing
• Alchemy, Rory Sutherland - The full argument for costly signaling, psycho-logic, and why perception beats reality in marketing
• Ehrenberg-Bass Institute for Marketing Science - The research foundation behind Byron Sharp's work on brand growth