AI will make your marketing cheaper. Used badly, it will also make it invisible.

You can spot AI copy at a glance now. Competent, tidy, weightless. Nothing wrong with any individual sentence, and nothing in it you'll remember four seconds later.

Most of the noise about AI in marketing is about jobs. That's not the risk that should worry a brand owner. The risk is sameness. AI doesn't write badly. It writes averagely, at enormous scale, for everyone at once. And average, in marketing, is where brands go to be ignored.

Why everything is starting to sound the same

Language models are trained on roughly everything, so left to their own devices they produce the statistical middle of everything. Ask for a product description and you get the average of every product description ever written. It's grammatical, it's inoffensive, and it belongs to nobody.

Now consider that your competitors are using the same handful of tools, with roughly the same prompts, at the same time. The output converges. Feeds and inboxes fill with copy that's technically fine and completely interchangeable.

Distinctiveness has always been the thing customers actually buy from brands. It's now also the thing AI erodes by default. That combination should shape how you use it.

Where AI earns its place

None of this is an argument against the tools. We'd make the case that AI belongs in most of your marketing operation. Specifically, in the production line:

Research and analysis. Summarising reviews, mining support tickets for objections, pulling patterns out of campaign data. Work that used to take a day now takes an hour, and the machine doesn't get bored on row 400.

Versioning and volume. Turning one strong concept into twelve sizes, three placements and four audiences. The idea is done; this is manufacturing.

First drafts of structural work. Briefs, outlines, reporting summaries, internal documents. Anything where the shape matters more than the voice.

Operations. QA checks, tagging, naming conventions, campaign checklists. The unglamorous work that actually determines whether an account runs well.

The pattern: AI is exceptional wherever speed and volume are the point and taste is not.

Where it quietly does damage

Then there's the other category. Voice. Judgement. Taste. Strategy. The decision about what to say, rather than the labour of saying it in twelve formats.

Hand those to the machine and the damage arrives quietly. Product pages that could belong to any brand in your category. Social posts that are "on brand" for a brand that doesn't exist. Support replies that answer a question nobody asked. Each one is a small withdrawal from an account you spent years filling, and no single withdrawal is big enough to notice.

The tell is always the same: it reads fine and it sounds like nobody.

The working rule

Automate the production line. Protect the point of view.

AI generates options, volume and speed. A human owns the voice, the judgement calls, and the final look at anything a customer will read. Not as a philosophical stance, as a commercial one: marketing that costs half as much to produce and sounds like everyone else isn't cheaper. It's worthless at a discount.

In practice that means a few unglamorous things:

A real voice guide. Not "friendly but professional". Banned words, sentence rhythm, actual examples of your best writing. Detailed enough that a machine, or a new hire, could be held to it.

A review gate on customer-facing output. Everything a customer reads gets human eyes before it ships. Internal work can move faster.

The same measurement as everything else. AI-assisted content isn't special. If it converts worse than what it replaced, the saving was an illusion.

Humans on strategy. What to sell, to whom, at what price, with what message. The tools can inform those calls. They shouldn't be making them.

Where we stand

We use AI every day inside our own agency and in client work, which is exactly why we have a view on where it breaks. The brands getting this right aren't the ones using AI most or least. They're the ones who decided, deliberately, which parts of their marketing are production and which parts are the point.

One more thing

We're putting together a working session for D2C and brand-led businesses on exactly this: where AI belongs in your marketing and customer comms, where it will cost you, and how to set it up so the brand survives the efficiency. If you'd want a seat when it runs, register your interest below and you'll be first to hear.