One Person Is an Entire Market

A few days ago, I decided to book a getaway.

I typed "seaside B&B" into a search box. Three times. Just three.

Then something strange happened. My news feed: B&Bs. My video feed: B&Bs. A friend's photos — and there, in the corner, another B&B. For days on end, the whole internet seemed to have held a meeting behind my back, and kept leaning in to ask: this ocean-view room — care to take a look?

If this has happened to you too, hold off on feeling offended.

Think about it: behind all of this sits an enormous business. As someone once put it, marketing in 2026 is in the business of the "market of one."

One person is an entire market.

And behind that one-liner sits a stack of hard-number estimates from consulting firms — the kind they charge real money for.

First, Let's Run the Numbers

The AI marketing market was valued at $47.32 billion in 2025 — that's the Statista and Grand View Research reckoning. Compound it forward at 36.6% a year, and by 2028 you land at $107.5 billion.

Three years. More than double.

The bigger number comes from McKinsey. Their estimate: generative AI can create roughly $463 billion a year in additional productivity for the global marketing industry — equivalent to 5% to 15% of total digital marketing spend.

More than $460 billion. That's the annual GDP of quite a few countries.

Where the money flows, the people follow. Marketing, as an industry, is being rebuilt by AI.

But the label "AI marketing" is too vague. Let's take it apart first.

What Does "Generative" Actually Mean?

You've surely received one of those "smart" marketing emails. They open with "Dear {姓名}", swap in your name, and blast ten thousand people at once.

That's traditional automation. In plain terms, it's a photocopier: the template is dead, the variables are alive, and it's the same medicine in a different bottle.

Generative AI is another matter entirely.

It's a copywriter and an illustrator in one. Tell it who you're writing to and what it's about, and it starts from a blank page, producing something different for every single person. Text, images, video, audio — all made to order.

Plenty of people are already using it. eMarketer's research shows 58% of marketers already use generative AI for content. A good share of the articles and posts you scroll past right now likely had their skeletons built by AI first.

And there's one more difference — the killer one: to change traditional automation's behavior, you wait for a programmer to rewrite the rules; generative AI learns from new data on its own, and the more you use it, the better it knows you.

A photocopier never gets better. A writer does.

So What Can It Actually Do?

Talking in concepts gets you nowhere. Let me tell you three stories first.

Story one: saving money.

Xerox needed sales-training videos, in multiple languages at that. Under the old playbook, the whole translation-to-filming chain would hemorrhage budget at every step. They used Synthesia, an AI video tool, to get digital avatars speaking French, German, and Spanish. The result: video-training costs cut in half, and delivery 30% faster.

Impressive.

Story two: having fun.

Coca-Cola once ran a campaign called "Create Real Magic": it opened up the brand's assets and let users generate their own Coca-Cola imagery with AI. Users couldn't stop playing — and spreading the word? The users took care of that themselves.

Story three: showing care.

Carvana is an e-commerce platform that sells cars. Once you take delivery, it generates a video made just for you, telling the story of you and this car. A used car, somehow delivered with a real sense of ceremony.

Three stories, three uses. But set them side by side and you'll see that what AI does in marketing comes down to four buckets, over and over.

Bucket one: it writes and draws for you. Copy, blog posts, product descriptions, images, video — tools like Midjourney, DALL·E, and Adobe Firefly turn out visuals in seconds, and you can retire the revision ping-pong with your design agency. Writers, meanwhile, are done "sitting and staring at an empty document": AI hands you ten angles, you pick one and go deep.

Bucket two: it finds the right person and says the right thing. What you've viewed, what you've bought — AI remembers it all, then computes what you're most likely to buy next. Infosys surveyed consumers: 86% admit that personalized content influences what they buy, and one in four of those call the influence "significant." You're eyeing a lipstick in a particular shade — it recommends the same color family at the same price point, rotating through brands. Customer service works the same way: AI chatbots don't sleep, replying in a second at 3 a.m. and slipping in suggestions for matching items and add-ons while they're at it. Salesforce's data is sitting right there: 82% of companies using AI say their customers' experience of browsing and choosing products has genuinely improved.

Bucket three: it handles search for you. Keyword research, on-page optimization, competitor watching, outreach emails — tools like SurferSEO and Frase.io wrap it all up. How should your homepage title be written to please the search engine and the human heart at once? It hands you ten options in one go. What is the competition doing with its SEO? It watches for you, every single day. And one more shift worth noting: eMarketer estimates that in 2025, 121.1 million Americans used generative AI tools to search. Which means whatever you write now has to do more than court the rankings — it needs clear structure and citable sources, or the AI won't pick it to quote.

Bucket four: it does the books and it tells fortunes. The fortune-telling is a joke — the professional term is predictive analytics. Throw an industry's historical data into one pot, and AI can smell trends that haven't taken shape yet. For example: it notices "plant-based" and "superfood" conversations climbing on social media, and may tip off a beverage company that it's time to stock up on kale-flavored kombucha. When the wind actually arrives, the goods are already on the shelf. It can also keep an eye on the stray online remarks about your brand; the moment something looks off, it adjusts your ads while they're still running — no waiting for the next campaign cycle.

But this is where I have to hit the brakes.

But Everything Has Its Flip Side

Flip side one: shoddy output gets punished. Google's March 2024 core update (the periodic revision of its search-ranking rules) took direct aim at low-quality, unoriginal, assembly-line content — a large chunk of it bulk-generated by AI. AI-written material survives only when it has a point of view, verified facts, real detail, and genuine usefulness to readers; thin, templated content will only slide down the rankings.

Flip side two: your brand's voice drifts. In 2025, more than 40% of practitioners said "inconsistent brand voice" was their number-one headache with AI. The more AI writes, the more your brand sounds like no one in particular.

Flip side three, the one that stings most: buying the tools doesn't buy growth. One gap says it all: 88% of marketing teams are using AI, but only 26% say they've actually captured value from it. Why? Because only 49% of teams have seriously run the numbers on AI's ROI. If you never run the numbers, how do you know whether you made money?

IBM has a figure: for every $1 enterprises invest in AI, they get $3.50 back. Sounds lovely. But that 3.5 is the average earned by the people who know how to run the numbers.

The tools are neutral. What opens the gap is the person using them.

So — What Does It Cost?

Three tiers.

Lightweight: off-the-shelf SaaS tools — Jasper, SurferSEO, Chatfuel and the like — at $50 to $500 per seat per month, with almost zero implementation cost. If your team used to write everything by hand, the switch can pay for itself within weeks.

Middleweight: wire it up yourself via API, plugging the models into your CRM and your content workflow. A one-time investment of $15,000 to $80,000, plus usage fees. Set your goals clearly, and the payback arrives in three to six months.

Heavyweight: enterprise-grade customization — fine-tuned models, a knowledge base, your own data connected. Starting at $100,000 with no upper limit, and a payback period of 12 to 18 months.

Worth it? In 2025, Gartner surveyed 822 business leaders: companies that adopted generative AI early saved an average of 15.2% in costs and lifted productivity by 22.6%.

And there's a smarter play — I call it "both hands": generative AI supplies the ideas, churning out dozens of ad variants a day, while deployment AI like Google's Performance Max or Meta's Advantage+ does the picking, spotting which variant can fight. There are also platforms like Albert.ai that squeeze both jobs into one. Vanguard does exactly this — a 2025 Harvard Business Review analysis noted that its LinkedIn ad conversion rate rose by 15%. Generation is the left hand, deployment is the right, and one fast hand alone is useless.

Back to That Ad

Remember the B&B ad that chased you around at the beginning?

The next time you see it, try a different lens: behind that one ad sits an entire business built on treating one person as a market of one.

Netflix uses it to decide which show to push you next — and off you go, one series after another. Nike uses it to guess which shoes you'll want next season. Starbucks uses it to plan the day's store schedules and recipes. The business has already passed $47 billion, and it's still growing at more than 30% a year.

Whether to get on board is your call.

But if you do, I have exactly one reminder: learn to run the numbers first, then learn to use the tools.

And may you be the one who boards early — and still keeps the books straight.