“Hi [First Name]” is not personalisation. Neither is sending your entire database the same email, just split into three segments based on purchase history. These are table stakes. They were impressive in 2015. In 2026, they’re the bare minimum.

Real personalisation means each subscriber receives content that’s relevant to them specifically — not just their demographic group or purchase segment, but their individual preferences, behaviours, and stage in the customer journey. AI is finally making this feasible for businesses that aren’t Amazon.

What AI Personalisation Actually Means

Traditional email segmentation groups subscribers into buckets. Past purchasers. High spenders. Inactive subscribers. Geographic regions. Each segment gets a different version of the email. Maybe three versions. Maybe five. But everyone in the segment gets the same thing.

AI personalisation operates at the individual level. Machine learning models analyse each subscriber’s behaviour — open patterns, click-through history, purchase timing, browsing data, engagement frequency — and make predictions about what content, timing, and offers will drive the best response for that specific person.

The difference is substantial. A segment-based approach might send all “high-value customers” a 15% discount on Tuesdays. An AI-personalised approach sends subscriber A a product recommendation email at 7:14 AM on Wednesday (because that’s when she consistently opens emails), while subscriber B gets a different product recommendation at 9:30 PM on Thursday (because he’s a late-night browser who engages most with evening emails).

Send Time Optimisation

This is the most immediately impactful AI feature in email marketing, and it’s available through most major platforms now.

Mailchimp, Klaviyo, and Braze all offer AI-driven send time optimisation. Instead of scheduling a campaign for 10 AM and hoping it lands at a good time for most subscribers, the platform analyses each subscriber’s historical engagement patterns and delivers the email at the time they’re most likely to open it.

The results are meaningful. Most brands see 10-25% improvements in open rates from send time optimisation alone. That’s not a rounding error — it directly translates to more eyeballs on your content and more clicks to your site.

The catch is that you need enough historical data for the model to learn from. New subscribers with limited engagement history default to general best-practice send times until the model accumulates enough data to personalise.

Dynamic Content Blocks

Beyond timing, AI enables content personalisation within the email itself. Dynamic content blocks display different products, images, or text to different subscribers within the same campaign.

A clothing retailer’s weekly email might show winter coats to subscribers in Melbourne, summer dresses to subscribers in Cairns, and new arrival sneakers to a subscriber who’s been browsing the footwear category. Same email structure, completely different content.

This used to require manual setup of multiple content variants and complex conditional logic. AI systems now automate the content selection based on predicted relevance, pulling from your product catalogue and matching items to individual subscriber interests.

Working with an AI consultancy can help businesses build the data pipelines that connect their product catalogue, customer behaviour data, and email platform into a system capable of genuine dynamic personalisation. It’s more involved than flipping a switch, but the infrastructure, once built, drives results across every campaign.

Predictive Product Recommendations

Amazon’s “customers who bought X also bought Y” model has been around for decades. What’s new is that similar recommendation engines are now accessible through standard email marketing platforms.

These models analyse purchase history, browsing behaviour, and similar customer profiles to predict which products a specific subscriber is most likely to buy next. The recommendations are generated automatically and inserted into emails without manual curation.

For e-commerce businesses, product recommendation emails consistently outperform generic promotional emails. Barilliance research shows that personalised product recommendation emails generate 5-6x higher transaction rates than broadcast promotional emails.

The quality of recommendations depends entirely on the quality of your data. If your product catalogue is well-structured with accurate categories and attributes, the models produce relevant suggestions. If your data is messy, the recommendations will be too.

Churn Prediction and Win-Back

AI doesn’t just help you sell more to engaged subscribers. It identifies subscribers who are about to disengage, often before traditional metrics would flag them.

Churn prediction models analyse patterns in engagement — declining open rates, fewer clicks, longer gaps between interactions — and identify subscribers at risk of becoming inactive. This triggers automated win-back sequences tailored to the individual: a special offer for a price-sensitive subscriber, a content-focused email for an information seeker, or a simple “we miss you” message for a brand loyalist.

The alternative is waiting until someone hasn’t opened an email in six months and then sending them a generic re-engagement campaign. By that point, they’ve mentally unsubscribed. Early intervention, driven by predictive models, catches them while they’re still reachable.

Subject Line Optimisation

AI tools can now predict subject line performance before you hit send. Platforms like Phrasee and built-in features in major ESPs analyse your historical data to predict which subject line variants will perform best.

Some systems go further, generating subject line variations automatically based on patterns in your highest-performing historical campaigns. The AI learns your brand voice, your audience’s preferences, and the language patterns that drive opens.

This doesn’t replace human creativity. It augments it. A marketer generates ideas; the AI evaluates and refines them based on data rather than gut feeling.

The Privacy Balance

Personalisation runs on data, and data collection faces increasing scrutiny. Australia’s Privacy Act reforms are moving toward stronger consumer protections, and the global trend is clearly toward more transparency and consent requirements.

Smart personalisation respects these constraints. It uses first-party data — information that subscribers have directly provided or that you’ve collected through your own platforms with proper consent. It doesn’t rely on third-party tracking data that’s being phased out across the ecosystem.

The businesses that build personalisation on first-party data foundations are the ones that will thrive as privacy regulations tighten. The ones relying on third-party cookies and purchased data lists are building on sand.

Getting Started

You don’t need to implement everything at once. Start with send time optimisation — it’s the lowest-effort, highest-impact AI personalisation feature. Then add dynamic content blocks for product recommendations. Then explore predictive churn modelling.

Each layer builds on the same foundation: clean subscriber data, proper tracking of engagement and purchase behaviour, and a platform capable of acting on AI-generated insights.

The goal isn’t to create emails that feel creepily personalised. It’s to send emails that feel relevant — that arrive at the right time, with the right content, for the right person. That’s not surveillance. That’s respect for your subscriber’s inbox.