The Old Email Marketing Playbook Is Being Quietly Killed by AI

A while back, a friend of mine who runs a cross-border e-commerce business met me for tea. He told me about something keeping him up at night: the same customer list that brought in hundreds of thousands in sales from a single blast two years ago now barely registers. Open rates have cratered. The money goes out, and you don't even see a splash.

He asked me: is email marketing dead?

No. The playbook changed.

Consider one number: an average user now receives about 121 emails a day. One hundred and twenty-one. Your email sits in that pile, wedged between a bank statement and a meeting notice. Why would anyone open it?

Traditional bulk sends get open rates of around 21%. AI-personalized emails, by contrast, achieve engagement rates 25% to 30% higher. That's the whole gap: AI can deliver the right content, to the right person, at the right time.

This isn't just my gut feeling. Looking back at 2025, plenty of case studies have already done the math. Let me tell you three stories.

Story One: Amazon, a Machine That Digests 4 Billion Interactions a Day

Amazon's email system may be the most complex on the planet.

It processes more than 4 billion customer interactions every day. Buy a laptop, and its algorithms instantly know what people with similar behavior went on to buy. A mouse, a bag, an external monitor. Then the recommendations arrive.

What about newly listed products with no purchase history? The system analyzes the product's own attributes, descriptions, and tags, finds similar items, and recommends those. Browse something without buying, and within 24 hours a follow-up email lands in your inbox — that product, plus complementary ones.

The results? Email-driven revenue up 25%, customer retention up 20%, cart abandonment down 15%. Most striking of all: AI-driven email marketing delivers a return on investment of over 300%.

And the revenue propped up by this recommendation engine accounts for 35% of Amazon's total revenue.

Incredible, isn't it? Behind each seemingly casual email, three systems work in concert: collaborative filtering (finding patterns across groups of users) reads the crowds, content-based filtering reads the products, and real-time triggers read the moment.

You think it's sending emails. It's actually calculating probabilities.

Story Two: Netflix, Stopping You Before You Walk Out the Door

Amazon obsesses over getting you to buy more. Netflix worries about something else: getting you to stay.

In a subscription business, churn is the single point that can kill you. Netflix's AI system tracks more than 1,000 behavioral signals per user. How completely you finished a series, the genres you favor, which device you use, what time you watch, what you searched, what's on your watchlist.

The moment something looks off — watch frequency dropping, indifference to new releases — the system automatically kicks off a retention sequence.

Note: this is not one of those "long time no see, we miss you" throwaway emails. It's content recommendations carefully selected based on your viewing history. Week one, a personalized list. Week two, early access to new titles in your favorite genres. Week three, social proof — what people similar to you are all watching. Only in week four do exclusives and limited-time perks come out.

Four steps, and the results: churn down 15%, user satisfaction up 25%, and 75% of email engagement coming from personalized recommendations. All told, roughly $1.2 billion in subscription revenue was pulled back from the edge of churn by this system.

The best moment to retain a customer is not after they cancel — it's before they start hesitating.

Story Three: A Small Fragrance Brand, 8 Emails, £5,549

The first two are giants. You might say: they have money, data, engineers — what does that have to do with me?

That's why my favorite is this small case.

A niche fragrance brand called Jubilee Scents partnered with Retain Marketing on an AI-driven email playbook. Just three moves.

Move one: collect zero-party data (information customers voluntarily tell you themselves). They built a fun interactive quiz that helped customers discover their scent preferences. Customers had fun, the brand got its profiles, and there was no privacy risk. Everyone opted in willingly.

Move two: AI-generated content variants. For each campaign, AI generates multiple versions of subject lines and body copy, automatically tests them across audience segments, and pushes whichever version wins.

Move three: optimize send timing. Whether to send, when to send, what to send — all tuned to each person's behavioral habits.

Six days, 8 emails, £5,549 in revenue. Open rate 34%, against an industry average of 21%. Click-through rate 8.7%, against 2.6%. Conversion rate on email traffic: 12%.

Eight emails. That number makes me stop and think.

AI marketing has never been the exclusive property of giants. Giants win on data volume; small brands can win on data quality.

So How Should an Ordinary Marketer Start?

Three stories told. But if you want to get started right now, my advice is: don't try to build an Amazon-grade system on day one. One bite at a time.

Months one and two, take inventory: how are your existing emails performing, is your data quality up to par? Set up basic data collection and management, pick a platform with AI capabilities — ActiveCampaign, HubSpot, or Mailchimp all work — and define your baseline metrics.

Months three and four, roll out the basics: behavioral triggers, send-time optimization, let AI draft subject lines for A/B testing, and use surveys and quizzes to collect zero-party data along the way.

Months five and six, go advanced: dynamic content personalization, predictive segmentation, churn prediction with retention campaigns, and CRM integration.

After six months, you can start talking about continuous optimization and scaling.

Notice something? This path and Amazon's growth path are essentially the same map — just at a different scale.

A Final Word

Some worry: with personalization this deep, what about privacy? GDPR and CCPA are sitting right there. But the case studies happen to hold the answer: leading brands have proven that hyper-personalization and compliance can coexist — provided you're transparent. Clearly state how data is used, give users an opt-out switch at any time, and spell out the exchange: "you give me data, I give you value."

The data shows 72% of consumers respond only to personalized content. Customers aren't unwilling to be understood. They're unwilling to be fooled.

The future of email marketing isn't about sending more often. It's about sending with more understanding.

Back to my friend. His problem was never that email died — it's that he's still speaking, in the style of a decade ago, to an audience that AI has already spoiled.

The users have changed. Will you?

Here's hoping your next email gets opened.