AI can help you post more, or turn your feed into the same beige soup as everyone else’s. This is the workflow we use to get the speed of AI without sacrificing your actual personality.
Every week, someone messages us something like this:
“I tried using AI for my posts. They sound…fine. Also like everyone else.”
They’re not wrong.
AI is brilliant at generating “pretty good” content.
The problem is that “pretty good” is exactly what gets ignored.
Most AI tools play it safe:
Safe colours in web design
Safe fonts
Safe copy – nothing that will upset anyone, or make anyone feel anything in particular
You see this all over LinkedIn.
Endless posts with the same rhythm, the same fake‑excited tone, the same “5 powerful tips to unlock your potential”.
So the question we had to answer for ourselves was:
“How do we use AI without letting it sand the edges off our clients’ voices?”
Here’s where we landed.
What we don ’t use AI for
Let’s start with the boundaries.
We don’t ask AI to:
“Write 10 LinkedIn posts in the style of a successful founder.”
“Rewrite this to make it more engaging.”
“Improve this copy.”
Not because the outputs are terrible.
Because they’re all weirdly similar.
They default to a bland “internet voice” – slightly over‑enthusiastic, full of neat little aphorisms.
That’s the opposite of what we want for our clients.
If your posts sound like 50 other people on the feed, nobody remembers you.
So for the parts of the work that matter most – the actual sentences that go out under your name – we still write by hand.
We listen to your voice notes, your rambles, your complaints about customers or suppliers, and we write the post from there.
Where AI genuinely helps
We do use AI.
We just treat it like a very fast, slightly overeager assistant.
It’s useful for:
- Organising messy notes after a call
- Pulling out themes from a long transcript
- Helping us see patterns we might miss when we’re too close to the work
For example, after a 45‑minute onboarding call, we might have:
- A transcript
- Handwritten notes
- Screenshots of your messages
Instead of reading everything three times, we’ll ask an AI tool to summarise what you talked about into rough buckets:
- Repeated stories
- Strong opinions
- Specific client examples
Then we go through those ourselves and decide what’s actually worth using.
AI is allowed to suggest buckets but it’s not allowed to decide what you believe.
Protecting your voice on purpose - the biggest risk with AI isn’t that it makes mistakes.
It’s that it makes you sound like everyone else.
So we build in a few safeguards.
We collect your “non‑AI” phrases first.
On our first call, we’re listening hard for the way you naturally talk. The phrases you always reach for. The way you describe your work when you’re not “on”. Those become our anchor points.
We keep your rough edges.
If you speak in half sentences, we don’t smooth everything into perfect paragraphs. If you understate instead of hype, we keep that. If you make odd little jokes about bank holidays, we let them stay.
We keep reference posts that are 100% human.
For most clients, there’s at least one post they wrote themselves that just feels right. The one that made their friends say, “This is so you.” That becomes a measuring stick. Before we send a new batch, we check: does this feel like it lives next to that post?
If AI starts drifting us away from that, we ignore it. Simple as that.
Using AI to make you more specific, not more generic
Ironically, AI is quite good at helping you be more specific, if you use it in the right direction.
Instead of asking:
“Write a post about why websites matter.”
You can ask:
“List the 20 most specific problems a small B2B founder might have with their website.”
You don’t publish what it gives you. You use the list to jog your own brain.
You remember:
- The client whose booking form sent people into a blank void
- The charity site that crashed before a fundraising campaign
- The founder whose site looked like it was built before their business grew up
Then you tell those stories, in your words.
We do the same behind the scenes. AI is allowed to help with prompts and reminders but you are still in charge of the opinions.
The human bit AI can’t fake.
At the heart of all of this is something AI can’t touch.
When you write about a real turning point in your life – a medical diagnosis, the first time you raised your prices, the weird relief of having a label for something you struggled with – there’s a weight to it.
You’re not just arranging words. You’re telling the truth about yourself. AI can remix that or suggest alternative lines, but it can’t feel embarrassed about telling the story.
It can’t decide that this is the week you’re finally ready to share it publicly. That’s why we don’t let AI near that part of the process.
We’d rather take the slower route and keep your voice intact. Because long‑term, that’s the only asset that actually compounds on LinkedIn.
Not tricks or templates, or the latest posting formula.
Your voice.
The way you see your work.
The way you talk when you’re not trying to impress anyone.
Our job is to capture that, protect it, and then scale it across months of content without it going flat.
AI can sit in the back of the room and help tidy up our notes.
But it doesn’t get a seat at the front of the stage.