Stop sending AI Slop. Here's how to fix your content.

Picture this: you spend 15 minutes crafting the perfect prompt to generate the perfect email. It's personalised, it references something specific about the organisation, and it closes with exactly the right call to action. You send it out. The recipient opens it, but doesn't respond. You notice this keeps happening. Your reply rate has dropped 40% from last quarter.

The emails aren't bad. They just sound exactly like every other message sitting in that inbox, because they were built the same way, with the same tools, using the same prompts.

Outreach is now faster

In fact, AI has made outreach faster for every sales team.

Hubspot’s 2024 sate of AI in sales found that 43% of sales reps use AI for outreach, nearly doubling from 24% in 2023.

Zoominfo’s 2025 survey found that AI-powered email improvements of nearly 90% response rate is possible; the email sent must not get lost in a sea of identically structured messages.

The problem is using AI as a ghostwriter instead of an amplifier.

When you let AI write everything for you, it regresses to the average of every sales email it was trained on.

AI is unable to replicate the specific way you notice things and particular rhythm of your sentences.

Distinctiveness is now the rarest thing in a prospect’s inbox.

That’s your edge.

These tools have made writing faster for everyone; distinctiveness is now the rarest thing in any inbox.

HubSpot's 2024 State of AI in Sales found that 43% of sales reps use these tools for outreach, nearly doubling from 24% the year before. ZoomInfo's 2025 survey found that response rate improvements of nearly 90% are possible, but only when the output doesn't get lost in a sea of identically structured messages.

The same applies to content: blog posts, LinkedIn updates, and follow-up emails are increasingly indistinguishable from one another.

The problem is using AI as a ghostwriter instead of an amplifier.

When you let a language model write everything for you, it regresses to the average of every email and article it was ever trained on.

What these tools cannot replicate is the specific way you notice things, the particular rhythm of your sentences, or the instinct you've built from years in your field. Distinctiveness is your edge. And it's recoverable.

Here are some tactics that make AI sound like you:

Step 1: Write in my style

If you're telling a tool to "write in my style," stop. It has no idea what your style is and it will guess badly.

What works is pasting two or three emails or pieces of content you actually wrote and use this prompt:

Here are a few examples of how I write. Analyse the tone, sentence structure, and personality in these. Then use that exact style to write [X].

This shifts from describing your voice to demonstrating it. Showing the tool what your conversational looks like, rather than telling it to "sound casual", gives it a model to replicate rather than an instruction to interpret.

This works whether you're writing a cold outreach email, a LinkedIn post, or a newsletter.

Step 2: Name your patterns

Once you have given AI your samples, go one step further and tell it what to specifically look out for. The more precisely you name it, the less AI has to infer (that is where generic language creeps in).

If you are not sure how to describe your own style, let AI do the analysis first. Use this prompt:

Read and confirm what it comes back with, and correct anything if needed.

Once you've shared your samples, go one step further. If you're not sure how to describe your own style let the tool do the analysis first. Use this prompt:

Based on the emails I just shared, identify the key patterns in how I write. Look at sentence length, tone, structure, use of humour or directness, how I open and close, and any recurring phrases or habits. List them out so I can confirm or correct them.

Read what it comes back with and correct anything that feels off.

The more precisely you name your style, the less the tool has to infer. Inference is where generic language creeps back in.

Step 3: Lock in your voice

Take the style analysis from Step 2, tighten it into a short paragraph, and paste it in the personalisation / instructions box.

Every major tool has a settings box where you can paste your style profile once and have it apply to every conversation automatically. Most people never touch it.

Here's exactly where to find it (updated February 2026):

ChatGPT — Profile icon (bottom left) → Settings → Personalisation → Custom Instructions. Ensure Enable customisation is toggled on. In the bottom box ("How would you like ChatGPT to respond?"), paste the style profile from Step 2. Hit Save.

Claude — Profile icon (bottom left) → Settings → Profile → "What personal preferences should Claude consider in responses?" Paste your style profile and save.

Gemini — Top left menu → Settings & help → Instructions for Gemini. Paste your style profile and hit Submit.

If your preferred tool has a Projects function, also add your style profile to the Project Instructions. This keeps it active across every conversation within that workspace — useful if you write in volume.

With your voice locked into settings, every new prompt starts from your baseline. Use this structure for any email, content piece, or message to make the output sound just like you.

P.S. If you want to go deeper on what it means to own your voice in an AI-augmented world, start here: How to Forever Avoid Sounding Like an AI

Close the gap and amplify what makes you human.

Start Your Human Skills Journey.

Sign Up

Email sent! Check your inbox to complete your signup.

No spam. Unsubscribe anytime.

Part of the HumanRise newsletter series. Developing irreplaceable human skills in the AI age.