What Is AI Marketing? Let Me Tell You 7 Stories

A while back, a friend of mine who works in marketing invited me to dinner. Halfway through the meal, he put down his chopsticks and sighed: everyone talks about AI these days. If your proposal doesn't sprinkle in some "AI-powered" this or that, you're almost embarrassed to send it out. But what does AI marketing actually look like? He couldn't come up with a single live example.

I said, stop looking up definitions. Definitions are useless.

The truth is, AI marketing reaches you every single day. Nobody ever points it out.

So here's what I'll do: I'll tell you 7 stories. When I'm done, you'll get it.

Story One: Every Netflix Homepage Is Different

My Netflix homepage looks completely different from my colleague's.

You might say, recommendations — who can't do recommendations?

But what is a recommendation, really? Most people picture a "you might like" shelf sitting there, take it or leave it. Netflix's recommendation doesn't work that way. It watches what you view, how long you watch, where you press pause, what you rate — then re-ranks your entire homepage. Every show it pushes to you lands square on your taste.

Netflix has published its own numbers: roughly 80% of what users watch comes from its recommendation system.

What does that mean? Of every 5 shows that get watched, about 4 are handed to you by a machine. As for how many users it has, I don't remember the exact figure, so I won't quote one. Just that 80% is scary enough on its own.

Think about it: before you've decided what to watch tonight, it has already decided for you. It turned "what you might want to watch" into a business.

It doesn't push one ad to 100 million people. It builds 100 million homepages — one by one, each pushed to a single person.

Story Two: That Pair of Socks on Amazon

The second story you know even better.

You searched for a pair of running shoes on Amazon. Over the next few days, "customers who bought this also bought" starts appearing under product pages, a matching pair of socks sits in your cart, and a "picks for you" email lands in your inbox.

All of that is the algorithm's doing. It combs through your browsing records and purchase history and guesses what you'll want next. McKinsey studied recommendation systems like this one years ago and reached a blunt conclusion: it is a major driver of Amazon's sales.

You only meant to buy running shoes, and ended up walking away with socks too. That's the algorithm giving you a gentle nudge right at the finish line.

Nudge accurately, and average order value goes up. It's that plain.

Story Three: Every Monday, a Surprise Right on Time

Story three: Spotify.

It has a playlist called Discover Weekly, refreshed every Monday. Almost none of the songs in it are ones you've heard — but odds are you'll like them. It takes your listening habits apart, piece by piece, matches them against tens of millions of users, and fishes out the songs "that people with taste like yours are listening to, and you haven't heard yet."

My God, the first time I used it, I was genuinely floored.

Stop and think: what kind of move is this? It makes no noise and no fuss — it just gives you a reason to open it yourself, every Monday.

The product itself became the marketing.

Notice something? Netflix, Amazon, Spotify — across these three stories, not one company is pushing generic ads. They're all doing the same thing: making the core experience feel as if it were built for you alone.

What about the second layer? Keep reading.

Story Four: When It Comes to Buying Lipstick, AI Tries It On for You

There's a headache every makeup lover knows: lipstick, foundation — through a screen, there's simply no telling how they'll look on your face.

Beauty retailer Sephora did two things.

One was virtual try-on. Open your phone camera, and the AI "paints" the lipstick onto your face; you watch the effect in real time. The other was a chatbot. What to buy, how to choose, booking an in-store beauty advisor — you chat, and it all gets done.

Both solve the same problem: hesitation.

The biggest hurdle to buying makeup is "I can't try it." AI moved the act of "trying" online. Once you dare to commit, placing the order follows naturally.

Story Five: It Even Factored in Today's Rain

Story five: Starbucks.

The offers its app pushes to you are calculated by AI. Four raw materials go into the math: your ordering history, what time it is, whether it's raining today, and what's been trending lately in your neighborhood.

On a rainy afternoon, it might push you a hot latte. On a weekend morning, it knows you want an iced Americano.

To put it bluntly: what separates a personalized offer from "20% off everything"?

20% off everything is shouting at a crowd. A personalized recommendation walks up to you and says: I know what you feel like drinking.

Feeling seen is more addictive than any discount. That's where repeat purchases come from.

Story Six: Coca-Cola Handed the Paintbrush to Its Fans

In the first five stories, AI was "guessing you." The sixth plays the game differently.

In 2023, Coca-Cola ran a campaign called Create Real Magic. It turned its classic visual assets into an AI creation tool and opened it up to fans worldwide: come, use our bottles, our imagery, and create your own work.

The campaign became a creative playground. Fans played themselves silly, the creations flooded the internet, and everyone was doing the brand's marketing for it.

This case is different from the previous five; it took another road: using AI to unlock participation. Coca-Cola's team kept a firm hold on the creative direction and the brand's boundaries throughout.

AI is the paintbrush. The hand holding the brush is still human.

Story Seven: The Machine Sorts, Then a Person Picks Again

The last story is the one I find most worth pondering.

Stitch Fix, an online clothing styling service. It works in two stages: first, AI sifts through mountains of clothes based on your style profile, sizes, and past feedback; then the human stylist steps in, and within that shortlist makes the final picks and pairings for you.

Why not simply hand it all to AI?

Because a machine can calculate "you might like it," but cannot calculate "whether this dress really suits that wedding you're attending next week." Taste and judgment are, for now, still human territory.

Machines do the heavy lifting; people make the call.

7 Stories Told — Now a Few Hard Truths

OK, that's all seven. Look back, and you'll spot a few patterns.

First, what AI is best at is where the data is more than people can process. Tens of millions of users, hundreds of millions of combinations — people can't do that work. AI can.

Second, AI handles scale, not direction. In every case, direction, strategy, creativity, judgment — all of it stayed in human hands. Netflix first decided "we must keep users," and only then did the machine calculate how.

Third, don't let the words "big company" scare you. Personalized emails, chatbots — small companies can afford them just the same.

Most people have AI marketing exactly backwards. They think it means using AI to blast ads at scale. Look at these 7 stories — which one is blasting ads? Every single one made the experience fit each person more closely.

The endgame of AI marketing is making every customer feel this business was opened just for them.

Finally, Back to My Friend

At the dinner table that day, I told him all 7 stories. When I finished, he asked: so where do I start?

I said, go back and look at your own business, and ask yourself one deeply unglamorous question: at which moment does my customer most need to feel "someone gets me"?

Find that moment, and that's your starting point. No need to buy big tools on day one. Get one email right, one recommendation slot right, and that's enough.

Here's to you becoming, soon enough, that seller who makes customers feel "they really get me."