AI email personalisation: relevant emails at scale

How AI personalises email content, timing and recommendations, where it genuinely helps and how to use it without feeling intrusive or breaking privacy rules.

By the UK Technology Hub teamPublished 4 minute read

The short answer

AI can personalise emails by predicting what each subscriber is interested in, recommending products, choosing send times, generating content variations and identifying people likely to buy or churn. It works best with good first-party data, clear consent, human review of generated content and limits that keep personalisation helpful rather than intrusive.

Personalisation used to mean inserting a first name. AI makes far more possible: emails tailored to each subscriber's interests, behaviour and timing. Done well, it feels helpful. Done badly, it feels creepy or simply wrong.

Where AI personalisation helps

Product and content recommendations

AI predicts which products or articles each subscriber is likely to want, based on browsing, purchases and similar customers.

Predictive segments

AI identifies groups such as likely buyers, likely churners or high-value prospects, so you can target messages appropriately. Our email segmentation guide explains segmentation foundations.

Send-time optimisation

Emails are delivered when each subscriber is most likely to engage.

Dynamic content

Blocks within an email change based on subscriber data: different hero images, offers or articles.

Copy variations

AI generates variations of subject lines and content for testing or for different segments.

Data foundations

AI needs reliable data:

  • Purchase and browsing history.
  • Email engagement.
  • Preferences shared by subscribers.
  • CRM data for B2B.

Clean, connected first-party data makes personalisation accurate.

  • Make sure your privacy notice explains how data is used for personalisation.
  • Obtain consent where required for tracking and profiling.
  • Respect preferences and opt-outs.
  • Avoid using sensitive data, such as health information, for personalisation without strong justification and consent.

Laws vary by country; some restrict profiling or require specific transparency.

Avoid the creepy line

Personalisation feels intrusive when it reveals too much, such as referencing private browsing in detail or inferring sensitive circumstances. Keep it helpful: "New arrivals you might like" rather than "We saw you looking at this three times last night."

Human review

AI-generated content needs checking for accuracy, tone, brand consistency and compliance. Product recommendations should exclude out-of-stock or inappropriate items.

Measuring impact

Compare personalised and non-personalised versions on clicks, conversions, revenue per recipient and unsubscribes. Personalisation should improve outcomes, not just engagement.

Getting started

  1. Ensure tracking and data integrations are solid.
  2. Start with recommendations in post-purchase or browse abandonment emails.
  3. Add send-time optimisation.
  4. Test AI-generated subject lines.
  5. Expand to predictive segments.

AI is reshaping email on both sides: marketers personalise more, and inboxes summarise more. Our guide to email marketing trends for 2027 explains the wider picture.

Common mistakes

  • Personalising with inaccurate or outdated data.
  • Over-personalisation that feels intrusive.
  • No human review of AI content.
  • Ignoring consent requirements.

A worked example

A pet supplies retailer used AI to personalise its weekly email. Customers who had bought dog food received content about dog health and relevant products, while cat owners saw different stories. AI chose product recommendations based on past purchases and browsing, and suggested send times for each subscriber. A human editor approved all copy and checked recommendations for obvious mistakes, such as suggesting puppy food to owners of older dogs. Revenue per email improved, and complaints stayed low.

Types of AI personalisation

  • Content selection: choosing which articles or products to show.
  • Copy variation: adjusting subject lines or intros for segments.
  • Send time: timing emails for each subscriber.
  • Frequency: sending more or less based on engagement.
  • Predictive segments: identifying people likely to buy or leave.

Explaining personalisation

Tell subscribers why they receive certain content, such as "Because you bought walking boots." Transparency builds trust and helps people update their preferences.

Fallbacks

Always provide sensible default content for subscribers with little data. Personalisation should never leave someone with an empty or confusing email.

Testing AI against simple rules

Compare AI personalisation with simple rule-based approaches. Sometimes simple rules perform almost as well at a lower cost, and they're easier to understand.

Keeping control

Set clear limits for what AI can change, review outputs regularly and keep humans responsible for brand voice and accuracy.

Starting small

Begin with one personalised element, such as product recommendations in a single flow, and measure the result against a control group. Expand only when you've seen a clear improvement and confirmed that data quality supports it.

Reviewing results by segment

Check whether personalisation helps every group of subscribers, not just the average, and adjust where it doesn't.

How we help

Our AI solutions team designs personalisation models and workflows, our e-commerce team connects store data for recommendations and our web app developers integrate CRM and app data into email platforms.

Frequently asked questions

Does AI personalisation work for small businesses?

Yes. Many email platforms include AI recommendations and send-time optimisation suitable for small lists.

Is AI personalisation legal?

Generally, with appropriate transparency and consent. Rules vary, especially for profiling and sensitive data.

Should I let AI write my emails?

Use AI for drafts and variations, with human review for accuracy and brand voice.

This guide is general information, not legal, tax or financial advice, and details such as prices, features and rules can change. Product and company names mentioned are trademarks of their respective owners and are used only to identify those products; UK Technology Hub is not affiliated with or endorsed by them.

Related guides

All email marketing guides