

Tue 08 / 09 / 26
How to make AI sound more like you (without arguing with it for an hour)
Richard Daniel, Director, Marketing and Digital Strategy Lead at Emerging Digital Partners, shares advice on getting generative AI to sound more like you and your business. Richard explores how a little more guidance can lead to better first drafts, while keeping human at the heart of your content.
By Richard Daniel of Emerging Digital Partners
This one’s for business owners and teams who use generative AI tools in their day-to-day work and are still figuring out how to get the most out of it.
Getting it to write something is easy. Getting it to sound like you is where things become interesting.
Most people don’t approach that training deliberately. They give it a task, dislike what comes back and begin a negotiation.
“Make it less corporate.”
“More concise.”
“No, you’ve removed the useful bit.”
“I would never say ‘delighted’.”
Each version changes, but the AI is still working from a description of your style that amounts to “not that”.
AI can produce perfectly readable writing from a basic instruction. The problem is that “readable” and “recognisably yours” are two different things.
“Sound human” means almost nothing
Ask an AI tool to make something “human, professional and engaging” and it has to decide what those words mean. The result will often be polished, enthusiastic and interchangeable with thousands of other AI-generated posts and pages. Your new service will “mark the beginning of an exciting journey”, whether you like it or not.
The AI hasn’t ignored your writing style. It simply doesn’t know what it is yet.
For most everyday uses, you’re not training a new AI chat model in the technical sense. You’re teaching the tool how to approach your writing through examples, preferences and corrections.
At Emerging Digital Partners, we see the tool as most useful when it supports human expertise rather than attempting to replace it. It can also help identify what may be missing when someone knows their subject but not necessarily what a particular piece of content needs.
See what changes when you give AI more to work with
Here’s a quick experiment you can try now using a real writing task from your business.
Choose one real writing task from your business and approach it from two separate AI chats (you can use whichever AI tool you prefer).
In the first attempt, you’ll use a fairly standard prompt. For the second, you’ll add an example of how you write and ask what information is missing.
You could choose a case study, website page, customer update, LinkedIn post or something else entirely. To make it your own, pull together any notes and facts you already have regarding your chosen content piece.
Use the prompt below in the first AI chat, replacing the words in square brackets with your own information.
“I need to write a [case study, website page, customer update or post] using the notes below.
Make it professional and engaging, explain the important points clearly and finish with an appropriate next step if needed.
Notes
[Paste your notes/facts here]”
That may produce a usable result, but “professional and engaging” still leaves the AI to decide how your business should sound. Keep the answer for comparison.
For the second attempt, find one writing example that sounds like you. This could be an email, blog, website page, social post or customer message. A closer match to the content piece you are creating can help, but what matters is that you wrote or approved it and still think it sounds right.
Open a separate AI chat and try this prompt:
“I need to write a [case study, website page, customer update or post] using my notes below.
I’m not exactly sure what information it needs or how it should be laid out. Read my notes and ask me any important questions before you start writing.
I’ve also included something we wrote before that sounds like us. Use it to get the tone right, but don’t copy it.
Keep to the facts and don’t make anything up.
Notes
[Paste the same notes you used in the first AI chat here]
Writing example
[Paste your approved writing example here]”
The instruction to ask questions before writing is important. You don’t have to predict every question or know every structure yourself. Answer what it asks, let it write and compare the results.
The second prompt isn’t better simply because it’s longer. It gives the AI an example of your voice and asks it to find missing information instead of filling the gaps with things it predicts you’d like to hear. AI can help with the structure, while you provide the facts and final judgement.
Explain what actually sounds wrong
Even a stronger prompt won’t always work first time. When something feels wrong, explain why.
“This sounds more alarming than it needs to.”
“This makes claims about quality without explaining what proves them.”
That gives the AI something useful. “This doesn’t sound like me” leaves it guessing again.
You don't have to know everything before you start
Giving AI more to work with doesn’t mean you need to write a perfect brief or know exactly what every type of content requires. Give it the facts you have, explain what you’re trying to create and ask it to help identify what’s missing.
What it needs will depend on the job. A client-facing document may need more background, evidence or a clearer next step. A social post may need a stronger point, a more natural opening or a different tone. AI can help you work out what “more” looks like before it starts writing.
You still answer the questions, correct what feels wrong and decide whether the finished piece is accurate and sounds like you.
The goal isn’t to create a digital clone and hand it your LinkedIn or email password. It’s to get a first draft that starts closer to your voice, takes less time to correct and still leaves the final judgement with you.

If you want to contribute to the Chamber blog, contact us on hannah@brightonchamber.co.uk

