When AI Becomes the Baseline in Marketing, What Are You Still Competing On?

A few days ago, I grabbed dinner with a friend who works in consumer goods.

He had this worried look on his face. He told me his marketing budget got slashed by 30% this year, but his boss still wanted him to double their ad performance. I asked him — have you guys used AI?

He paused for a second and said they'd tried — they had an intern use ChatGPT to write a few Xiaohongshu (RED, China's leading social commerce platform) posts.

I said, that doesn't count as using AI. That's just checking a box.

The real story is somewhere else entirely.

Let Me Put a Number Out There First

PwC ran a pulse survey of CMOs, and 78% of chief marketing officers are already embedding generative AI into their marketing systems.

Not "piloting." Embedding.

And the market size is even more striking: the global AI marketing industry is projected to surpass $107 billion by 2028, with a compound annual growth rate of 36.6%.

What does that mean?

It means AI in marketing is no longer a "competitive advantage" — it's the baseline. If you don't adopt it and your competitors do, you get left behind.

But here's the problem: most people's understanding of "AI + marketing" is still stuck at the shallowest layer — writing copy. Today, let's peel this back layer by layer.

What Is AI Actually Doing in Marketing?

Let me spell this out.

When a lot of people hear "AI marketing," a chatbot pops into their head. That's like hearing "digital transformation" and only thinking "build a website."

What AI does in marketing touches every layer — from customer awareness, to consideration, to decision, to retention. It's reshaping every single one.

Let me walk you through it layer by layer.

Layer one: Awareness — getting more people to see you.

In the old days of content creation, pumping out one WeChat Official Account article a day was already pushing the limit. Now AI produces a draft in minutes. You feed it your brand voice and target audience, and it writes twenty versions for you. Same with advertising — AI analyzes consumer behavior in real time and automatically adjusts your ad strategy, making sure every penny goes where it should.

Layer two: Consideration — making interested people want to get closer.

AI uses predictive models to score every prospect. Who's more likely to buy, who's more likely to churn — it tells you in advance. And based on each person's behavioral trail, it dynamically pushes the right content right to their screen.

Layer three: Decision — giving someone who's ready to pay one less second of hesitation.

AI algorithms adjust pricing and recommendations in real time based on demand and user profiles. What to push, when to push it, what messaging to use — it figures all of that out for you.

Layer four: Retention — making sure people who already bought don't want to leave.

AI monitors churn risk, identifies who's starting to go cold, and automatically triggers win-back actions. Customer service, in particular, has already been completely reshaped by chatbots.

See? This is not just "writing copy."

Let Me Tell You Three Real Stories

Talking about principles only gets you so far. Let me tell you three stories I've come across recently.

Story one: Seven million jars sold out in a month

You may have heard of the Nutella case.

Ferrero used AI to design the labels on Nutella jars. Not one design — it generated a unique pattern for every single jar. Millions of jars, each one looking different.

The result?

Seven million jars, all sold out in one month.

Why was it that explosive?

Because consumers discovered "mine is the only one like it in the world" — they didn't just buy it, they photographed it and posted it on social media. Suddenly every feed was full of Nutella. An industrially produced sauce got transformed by AI into a "limited-edition collectible."

Think about that. That's the power of AI to turn "ordinary" into "unique."

Story two: Unilever completely reimagined product photography

Unilever is one of the world's largest consumer goods companies, with hundreds of brands under its umbrella. In the past, shooting a single product promotional image meant building a set, setting up lighting, retouching — a slow and expensive process end to end.

They used generative AI to create digital twins for their brands — pixel-level, one-to-one virtual replicas of products. Those images then went straight into TV commercials, e-commerce product pages, and social media, with speed and volume that jumped by orders of magnitude.

And the people? People went to do creative work, make judgment calls, and handle the things AI can't do.

Machines do what machines should do, and people do what people should do. That's how it should work.

Story three: Response time cut by 90%

This is also Unilever. They deployed a system called Alex in customer service, powered by ChatGPT. Every customer email that comes in is first triaged by the system — which ones are real issues, which ones are spam — then it automatically drafts a reply, which is reviewed or edited by a human before sending.

The result: response time was cut by 90%.

90%.

Think about it. Before, a customer email would sit for three days with no response. Now there's a reaction within minutes. How could customer satisfaction not go up?

And the people? They're no longer drowning in repetitive email piles — they can handle the complex issues that genuinely require empathy and judgment.

Someone summarized it perfectly: AI handles the grunt work, and humans handle the work that requires heart.

Five Real Trends

Now that the stories are out of the way, let me lay out the "skeleton" of this whole picture for you.

Heading into 2026, there are five truly noteworthy AI trends in marketing. Not the "the future is here" fluff — things that are already running inside major companies.

Trend one: Hyper-personalization

What is hyper-personalization?

It's not the "Dear Mr. Zhang" kind of fake personalization. It's AI stitching together all of a user's data from CRM, websites, purchase history, and social media to build a dynamic, living profile of them.

Then, it predicts what they'll do next and puts the right thing in front of them before they even know they want it.

That's exactly what Netflix does. They run relentless A/B testing — randomly grouping users to see which version of the UI or which recommendation format makes people more likely to watch another episode. Then they roll out the winning version to everyone.

The core idea isn't "I'll show you what you like" — it's "before you even realize what you want to watch, I've already figured it out for you."

Trend two: Generative AI enters large-scale production

A 2025 survey found that 58% of marketers using generative AI reported that their content performance had noticeably improved.

Notice — not "faster to produce," but "better in quality."

Why?

Because AI lets marketers go from "writing one piece" to "writing ten and picking the best one." From "one version for every platform" to "twenty versions precisely targeted at different audiences."

The creative playing field just expanded tenfold, twentyfold. Things that used to be out of reach are now within reach.

Trend three: Predictive AI shifts decisions from "after the fact" to "before the fact"

In the past, when you looked at data, you were looking at "how did last month go."

Predictive AI looks at "what's going to happen next month."

It analyzes historical data to predict market trends, churn risk, campaign performance, and individual user lifetime value. Marketers no longer have to go on gut feeling — they can make decisions before things happen.

The essence of this shift is going from "reacting" to "anticipating." If a marketing team can know two weeks ahead of time that a batch of customers is about to churn and take proactive win-back action, the results are on a completely different scale compared to trying to fix things after the fact.

Trend four: AI Agents take over entire workflows

This is the hottest direction this year.

What's an AI Agent? Think of it not as a tool, but as an assistant that makes its own decisions.

It can scan data, segment users, score leads, schedule meetings, and send automated email sequences — without you watching over it constantly.

One study found that 82% of companies plan to adopt AI Agents within the next 1–3 years. Tools like Manus AI are already being discussed extensively across the industry.

My take? This isn't slowing down. Because the logic is too clean — humans go from being operators to supervisors, and machines go from being tools to being colleagues.

Trend five: Responsible AI is the floor, not the ceiling

The first four trends all sound great.

But one thing has to come before all of them: data governance and AI ethics.

AI feeds on data. Whatever you feed it determines what it becomes. Dirty data means dirty decisions. Biased data means biased conclusions.

GDPR, CCPA — these regulations are not for show. Customers hand over their data to you based on trust. Once that trust is broken, the loss is far greater than whatever cost you saved.

So governance isn't a cost — it's insurance.

How do you actually do it?

First, establish a unified data governance framework. Who collects, who stores, who can access — spell it out in black and white.

Second, compliance first. GDPR, CCPA — not a single one can be missing. Keep proper audit trails.

Third, get marketing, legal, IT, and compliance into the same room and align on the boundaries of AI usage. Don't let each department go its own way.

Fourth, be transparent with your customers. If you're using AI for personalization, tell them openly. Hiding things is what gets you in trouble.

Fifth, when evaluating vendors, treat "security, privacy, compliance" as hard requirements. Enterprise-grade platforms like Sprinklr have built this into their core selling point — data encryption, role-based permissions, and compliance audits all baked in.

So, How Do You Actually Implement?

That's a lot of talk. When it comes down to your own company, where do you start?

Here are six steps, in order.

Step one: Get your KPIs straight first.

Don't adopt AI just for the sake of adopting AI. What are you actually trying to solve? Is content production too slow? Is ad ROI too low? Is customer churn too high? Figure that out first, then deploy AI to attack it.

Step two: Audit your data.

AI feeds on data. Is your data clean? Is it structured? Can it be accessed? If this step isn't done right, everything after is built on sand.

Step three: Choose your tools.

General-purpose tools vs. custom-built — that's a trade-off. Third-party tools get you up and running fast; custom-built gives you more control. My advice: start with third-party tools to get the process working, and once it's running smoothly, then consider building your own.

Step four: Keep humans in the loop.

This is the one I want to emphasize most.

AI is not here to replace people — it's here to amplify them. Creativity, judgment, empathy — these are things AI can't do. Let humans handle those. Let AI handle the repetitive, mechanical, high-volume-but-low-judgment work.

Unilever's case already showed us — humans and AI working together beat either one alone.

Step five: Start small and iterate fast.

Don't try to overturn your entire marketing system on day one. Find one small, high-ROI scenario, get it working, validate it, then scale up. A pilot mindset is crucial.

Step six: Embed governance into your process from the start.

Privacy, transparency, bias detection — set the rules from day one. Don't wait until something goes wrong to patch things up.

One Judgment from Me

Let me come back to the friend from the beginning of this story.

I told him: AI isn't a question of "whether to do it." It's a question of "how to do it, and how deep to go."

There's one data point from the research that really stuck with me: companies that invest in AI see an average revenue increase of 3%–15%, and sales ROI improvement of 10%–20%.

This isn't the future tense. This is happening right now.

But I also want to give you a reality check. AI is not a silver bullet. It won't automatically make your marketing better just because you bought a tool or signed up for a platform.

It amplifies what you already have. If your strategy is right, it helps you go faster. If your strategy is wrong, it helps you go wrong faster.

So, back to the question from the very beginning — when AI becomes the baseline in marketing, what are you still competing on?

It's not about who has more tools. It's about who thinks more clearly.

Tools, anyone can buy. Insight — that's the real moat.

Here's hoping that in the middle of all this change, you're the one who thinks it through.