Generic words are free now, sounding like yourself is not

The market price of fluent text fell to zero in about two years. Every company can now produce unlimited copy that is correct, polite, and structurally flawless, which is exactly why it has stopped saying anything. When everyone writes with the same machine, everything starts to sound the same: the same rhythm, the same word choices, the same careful friendliness. The machine is not the fault. AI writes many things better than a tired human at eleven at night, and we use it ourselves. The fault is the sameness, which nobody ordered and everybody received. Once fluency is free, differentiation moves elsewhere: to what you are willing to state precisely, and to whether the text sounds like anyone at all.

Fluency was a signal, and the signal died

Polished writing used to carry real information. It meant somebody spent time, knew their subject, and cared about the result. A reader could use the quality of the language as a proxy for the quality of the operator, because a sloppy business rarely produced careful prose.

That proxy is now broken. Polish is the default output of a button, so readers are already hunting for the next signal. The next signal is specificity: a claim that can be checked, a number someone stands behind, an opinion it is possible to disagree with. Generic text never takes a risk, and the absence of risk is precisely how you spot it.

The world’s biggest brand bought the same backlash twice

In 2024 Coca-Cola released an AI-generated remake of its classic Christmas ad and collected a wave of public irritation.¹ The following year it shipped another AI ad and collected the same reaction again, this time as a deliberate choice.²

Coca-Cola can afford this. Its brand absorbs an annual controversy, and the attention may even serve it. A small business has no such balance sheet. Your stock of trust is your entire estate, and a reader who clocks your text as machine filler quietly extends the conclusion: if the communication is filler, perhaps the service is too. The conclusion may be unfair. It gets drawn anyway.

Voice is a set of decisions, not a mood

Tone of voice gets mistaken for atmosphere, as if the question were whether to be more casual or more formal. In practice a voice is a stack of decisions: which words you refuse to use, which claims you will not make without cover, which sales line you will not write even though it would convert.

We run one test on every sentence, and it is harsher than it sounds: would you say this out loud to a client, across a table, unprompted. If not, it does not get published, however elegant the grammar. A machine cannot make these decisions for you, because they require something it does not have: something to refuse.

Let the machine write, as long as you are the one speaking

The honest framing is not human versus machine. AI is a good drafter, a merciless proofreader, and a fast summariser, and it would be theatre to pretend this very text never met one. The difference is in who owns the claims.

Where a voice is defined before the machine arrives, the machine accelerates the work. Where no voice exists, the machine fills the vacuum with an average, and the average sounds like everyone else. The order of operations decides everything: first the decisions about what you will and will not say, then the tool that helps you say it faster.

The inconvenient part

Content is now produced faster than anyone’s willingness to read it, and most of the new supply is polite fog that sticks to no one. In that environment, sounding like yourself is the cheapest competitive advantage left: it needs no media budget, only the nerve to say something you will stand behind. The test is simple. If your company’s name were removed from your website, would a single customer recognise the text as yours? If not, you do not have a communication problem. You have a voicelessness problem, and producing more text will only make it larger.

From words to done.

Sources

1. NBC News, “Coca-Cola causes controversy with AI-made ad” (2024). https://www.nbcnews.com/tech/innovation/coca-cola-causes-controversy-ai-made-ad-rcna180665
2. Forbes, “Coca-Cola Sparks Backlash With AI-Generated Christmas Ad, Again” (2025). https://www.forbes.com/sites/danidiplacido/2025/11/04/coca-cola-sparks-backlash-with-ai-generated-christmas-ad-again/
3. Marketing Dive, “Why Coca-Cola keeps pushing the limits of generative AI despite backlash”. https://www.marketingdive.com/news/why-coca-cola-keeps-pushing-limits-generative-ai-despite-backlash/804739/