Skip to content
← All articles AI & MARKETING

How to Use AI for Marketing Without Sounding Like Everyone Else

Take your latest marketing email. Remove your company’s name. Could one of your competitors send it without changing anything else?

The grammar might be perfect. The benefits might sound impressive. But if the message could come from almost anyone, it gives the reader very little reason to remember you.

AI makes it easier to produce polished marketing. Making that marketing relevant, distinctive and worth responding to takes more thought.

The opportunity is to use these tools to express something your business has a credible reason to say.

The same paintbrush. A different picture.

A paintbrush in my hand is not the same as a paintbrush in Michelangelo’s.

We could have the same materials in front of us. What he brings is an understanding of composition, proportion, light and where to place the next stroke. He can see what belongs in the picture and what needs to be left out.

That is how I think about AI in marketing.

ChatGPT or Claude can help you explore ideas, question an argument and improve a sentence. They can make difficult work more accessible. But there are still decisions to make: who you want to reach, what matters to them, which claims you can support and why they should choose you.

The tool can help you work through those decisions. You still need to judge the answers.

There is good evidence that AI can help people improve. In an experiment involving 453 professionals, access to ChatGPT reduced the average time spent completing writing tasks by 40%, while the assessed quality of their work increased by 18%. Weaker performers benefited particularly strongly. These were controlled professional-writing tasks, rather than a test of campaign sales, but the productivity gains were substantial. Noy and Zhang, Science, 2023 ↗

The question is what you do with that extra capability.

Better writing can still become more alike

In another experiment, researchers asked participants to write short stories, with some given access to AI-generated ideas.

The AI-assisted stories received better evaluations for creativity and writing quality, particularly among less creative writers. Yet they also became more similar to one another. The experiment involved short fiction and constrained prompts, so it does not establish that every AI marketing campaign will become generic. It does illustrate how individual improvements and greater similarity can happen together. Doshi and Hauser, Science Advances, 2024 ↗

For a business, that creates a useful challenge.

Your copy may become smoother while losing the observations, language and point of view that make it recognisably yours.

Simply asking for something “more persuasive” gives the model little direction. Persuasive to whom? About which problem? With what evidence?

Those questions are where the useful work begins.

Give AI something worth saying

A brief such as “write a LinkedIn post about our excellent customer service” leaves a great deal unexplained.

What does excellent service mean in your business? What do customers actually value? What happens when something goes wrong? Is there an example you can substantiate?

Specific answers give you stronger material to work with.

Consider these two illustrative descriptions of an imaginary reporting service:

“We help businesses unlock efficiency through innovative reporting solutions.”

And:

“We bring sales figures from your existing systems into one weekly report, so your team can review performance without rebuilding a spreadsheet each time.”

The second explains what the service does and why someone might care. It also gives the buyer something concrete to assess.

That clarity comes from understanding the offer.

Before asking AI to draft, give it:

  • The audience: their role, situation and relevant needs.
  • The offer: what you provide and how it works.
  • The difference: why someone might choose it over the alternatives.
  • The evidence: facts, examples and results you can substantiate.
  • The objections: concerns buyers have actually raised, separated from your assumptions.
  • The next step: what you want the reader to do.

Then ask it to identify what is missing before it writes anything.

A useful instruction is:

“Read this brief as a sceptical prospective customer. Identify vague claims, missing evidence and questions the message needs to answer before drafting.”

You may discover that the biggest improvement needs to happen in the argument before it happens in the wording.

Keep the details that make it yours

Your own explanations are valuable source material.

How do you describe the service on a sales call? Which question do customers repeatedly ask? What do you find yourself correcting when people misunderstand your offer?

A few paragraphs in your own words can give AI a clearer sense of your voice than a string of adjectives such as “professional, confident and approachable”.

Ask it to preserve the substance of what you mean. Then read the result aloud.

Would you say that sentence to a customer? Does it explain anything? Has an important qualification disappeared because the revised version sounds more impressive?

A distinctive voice can be straightforward. It comes through in what you notice, what you believe matters and how clearly you explain it.

You do not need to make every sentence unusual. You need to give the reader something specific enough to recognise and remember.

Make personalisation earn its place

Mentioning a prospect’s latest post or congratulating them on an announcement can demonstrate that you looked them up. It does not automatically establish why they should be interested.

The detail needs to connect to the offer.

A company opening another location might be relevant to a service that supports multiple sites. It is a starting point for investigation, rather than proof that the company has a problem or intends to buy.

AI can help examine that connection: what the development establishes, which needs it might create and what remains unknown.

That matters before a message is written. Account selection, research and the reason for approaching someone all shape whether the eventual copy has anything worthwhile to say.

Producing hundreds of variations becomes more useful when the underlying reasons are sound.

Test what happens after someone reads it

A message can sound convincing in a meeting and still fail to produce a useful response.

Research published in 2026 tested AI-personalised ads on Meta platforms, using news articles as source material. Personalisation did not significantly improve engagement overall, and some approaches reduced it for particular groups. Audience ratings of appeal also differed from observed behaviour. El Fraihi and colleagues, ICWSM, 2026 ↗

That study does not settle whether personalisation will work for your business. It reinforces the importance of testing it in context.

Review drafts for accuracy, relevance and voice. Then follow what happens in the market: useful replies, qualified enquiries, opportunities and customers.

Look at objections as well as successes. They can reveal what the message has failed to explain—or where the offer and audience do not fit.

At So Sure Media, we bring market research, competitor analysis, key decision-maker research and persuasive writing together. AI supports that work, helping us investigate and develop relevant approaches at scale. We qualify interest against agreed criteria and give your sales team the context to continue the conversation.

The paintbrush gives you possibilities. The decisions you make with it shape the picture.

If your marketing sounds polished but gives people little reason to respond, let’s look at the thinking behind it.

Let’s talk about your marketing
← Back to the blog