How to make your marketing emails sound personal using AI

Your email open rate is stuck at 18%. Click-through rates are mediocre. Unsubscribes are on the rise. It’s not that your offer is bad. It’s just that your emails sound exactly like what they are: mass-mailed messages that could have been sent to anyone.

Subscribers filter out impersonal content in milliseconds. A generic subject line, a standardised greeting, a body that shows no understanding of who they are or what they need. Delete. The paradox is that using AI to write emails faster can make this problem worse, unless you know how to strategically humanise every element of the message.

Why personalisation matters more than perfection

A grammatically perfect but impersonal email loses out to one with minor imperfections but which sounds as though it was written specifically for you. The data confirms it: personalised emails have 26% higher open rates and generate six times more conversions than generic ones.

Personalisation isn’t just about inserting [NAME] into the greeting. It’s about demonstrating an understanding of the subscriber’s context: their stage in the customer journey, their previous interactions with your brand, their specific pain points. AI can scale this understanding if you guide it correctly.

Humanise the subject line: the battle is won here

47% of subscribers decide whether to open an email based solely on the subject line. If it sounds like mass marketing, you’ve lost before you’ve even started.

Avoid: ‘Weekly Newsletter #47: Special Offers’ Use: ‘Ana, this changed the way I manage projects’

The difference lies in specificity and relevance. The second subject line suggests specific content, uses a name (but not in an obvious way), and promises concrete value. A text humaniser can help you identify subject lines that sound too corporate and suggest more conversational variations.

Try these humanised patterns:

• Direct question: “Are you still struggling with this?”
• Specific observation: “I noticed you visited our pricing page”
• Time reference: “Three months on: what we’ve learnt”

Personalise the context, not just the data

Inserting a name and company is the bare minimum. True personalisation uses behavioural data to tailor the entire message.

If the subscriber downloaded your SEO guide two weeks ago but hasn’t opened the last three emails, the right context is: “I know you were exploring SEO strategies. I want to show you the next step that 67% of those who read that guide are using now.”

That paragraph shows you remember the previous interaction, understand their potential objection (they didn’t open the emails because they didn’t seem relevant), and offer logical progression. AI can generate this kind of context if you provide it with the right segmentation data.

Use snippets of real conversation

The best emails sound as if they’re continuing a conversation that’s already started. Review the responses you’ve received from subscribers: their questions, objections, and the specific language they use.

Incorporate those phrases into your AI-generated emails. If a customer wrote “I’m overwhelmed by so many options”, that exact language should appear in your email: “Many people tell me they’re overwhelmed by so many options. That’s why I created this simple comparison...”

This approach transforms generic content into communication that feels like a direct response to the subscriber’s thoughts.

Adjust the tone according to the stage of the funnel

A cold lead who has just subscribed needs a different tone to a customer who has been with you for six months. AI can generate both, but you must specify the difference.

For new leads: More educational, less presumptuous. “You’re probably wondering how this works…”

For existing customers: More direct, references to shared experiences. ‘Remember when you implemented that strategy? Here’s the next level...’

This tonal calibration is crucial. An email that assumes too much familiarity with a stranger generates rejection. One that treats a loyal customer like a cold prospect feels impersonal. Adapting the tone to your audience is essential for maintaining relevance.

Asymmetrical structure: break the mould

AI-generated emails tend to follow a predictable structure: greeting, three balanced paragraphs, CTA, closing. This uniformity gives away the automation.

Introduce deliberate asymmetry:

• Start with a direct question, without an introduction
• Use a single-sentence paragraph for emphasis
• Place the CTA halfway through the email, not just at the end
• End with a postscript that adds extra value

This structural irregularity signals that a human made specific editorial decisions, not that a template was automatically filled in.

Inject strategic imperfection

Perfect emails sound generated. Small ‘imperfections’ humanise them:

• Use contractions: ‘I can’t’ vs ‘I couldn’t’
• Occasionally include fragments. For emphasis.
• Allow ellipses for natural pauses... like when you think out loud
• Don’t be afraid to start sentences with “And” or “But” when it flows naturally

These elements mimic natural speech. Don’t overdo it (an email full of ellipses is irritating), but used sparingly they signal authenticity. These techniques help you humanise text without losing your authentic voice in every communication.

Smart segmentation multiplies personalisation

AI shines when you generate multiple variations of the same email for different segments. Instead of a generic email for 10,000 subscribers, create 5 versions for segments of 2,000 based on:

• Subscriber’s industry
• Products they’ve viewed
• Time since last interaction
• Stage of the customer journey
• Previous engagement level

Each version retains the core message but adjusts examples, tone, references and CTAs. This level of personalised segmentation was impossible to achieve manually. With AI properly applied, it’s routine.

The CTA must feel like the natural next step

Generic calls-to-action kill conversion: “Click here”, “Find out more”, “Take advantage now”. These sound like a desperate plea, not a valuable invitation.

Humanise the CTA by connecting it to the specific context:

• ‘See the 3 strategies I mentioned’
• ‘Book your 15-minute review’
• ‘Download the template I use every Monday’

A specific CTA implies that there is concrete content waiting, not a generic sales page.

A/B testing of humanised elements

Don’t assume what works. Test systematically:

• Version with a personal anecdote vs without
• Formal tone vs conversational
• Short paragraphs vs varied length
• Direct CTAs vs contextualised ones

AI tools allow you to generate these variations quickly. Testing tells you what specifically resonates with your audience, not with hypothetical audiences.

The aim isn’t to deceive, it’s to connect

Humanising emails with AI isn’t about pretending you wrote every message to 10,000 people by hand. It’s about using technology to maintain elements of genuine connection that would be impossible on that scale.

When a subscriber reads your email and thinks “this is exactly what I needed to know today”, they don’t care if AI helped generate it. They care that you understood their situation and offered relevant value. That is the personalisation that matters.

Avoiding signs of robotic content in your emails will make the difference between being ignored or generating real engagement.