87% of Marketers Are Using AI. Fewer Than Three in Ten Truly Know How to Use It.

A few days ago, a friend of mine who works as a CMO invited me out for tea.

He told me that over the past two years, his company had rolled out seven or eight AI tools in one go, and the tooling budget had nearly tripled. I asked how it was working out.

He held his teacup, stayed silent for a long moment, and gave me three words: Hard to say.

Hard to say. I keep hearing those three words more and more lately. Nearly every marketing leader talks about AI, but the moment you ask about results, they get vague.

So what's actually going on? I went and dug through the surveys from Salesforce, HubSpot, McKinsey, and Gartner. And when I finished, I realized the answer was hiding inside one stark contrast:

The number of people using AI is absurdly high. The number using it well is pitifully small.

Two Years, Thirty-Six Percentage Points

First, let's look at how many people are using it.

Every year, Salesforce runs its State of Marketing global survey — 4,450 respondents — the most continuously tracked first-hand data available on this. In Q1 2024, 51% of marketers used generative AI in at least one fixed workflow. In Q1 2025, 76%. In Q1 2026, 87%.

Two years. Thirty-six percentage points.

What does that mean? A new technology spreading through the marketing industry used to take a decade. This time, it took 24 months. Large companies (marketing teams of 250 or more) hit 94%, and even teams of fewer than 10 cleared 73%. The wait-and-see crowd has essentially vanished.

Last year, a popular question was: will AI replace marketers? Nobody asks that anymore.

The question has become: will marketers who use AI replace the ones who don't?

Having a Gym Membership and Having Abs Are Two Different Things

Seeing that 87%, you might think this story is basically settled.

Not so fast. There's another number.

The CMO Survey (Spring 2025, published by Duke University's Fuqua School of Business together with Deloitte and other partners) measured a finer-grained question: what percentage of total marketing work does generative AI cover, on average?

15.12%.

Put those two numbers side by side and things get interesting. 87% of people use AI in at least one workflow, but on average, AI is only doing 15% of the marketing work. The remaining 85% is still done purely by hand.

Whether you've used it is breadth. How deep you go is another matter entirely. Breadth is nearly maxed out; depth has barely started.

What does that look like? It looks like gym memberships. 87% of people have signed up, but only 6% to 30% have actually built the abs.

That's not a metaphor I made up — it's the measured range from two research firms, BizIQ and Digital Applied: organizations that have genuinely integrated AI into all of their marketing processes fall somewhere in that band. BCG's survey is even more brutal: 74% of companies struggle to get AI projects to deliver scalable value.

Everyone is buying the tools. Most companies still haven't collected the value.

What's It All Being Used For?

So everyone's using it daily — using it for what, exactly?

Gartner's CMO Spend Survey gave a somewhat surprising answer: 77% of AI users apply it to creative development. That ranks first among all use cases.

What would you have guessed? Ad buying? Data analysis?

Neither. It's creative work.

HubSpot's survey is even more down to earth — the three things marketers use AI for most: brainstorming topic ideas (62%), summarizing content (53%), and writing first drafts (44%).

See the pattern? It's all wordwork. AI got its foothold in marketing by working as a copywriter and an assistant.

The tools themselves are exploding too. The number of marketing AI tools went from 1,200 in 2024 to more than 3,800 in 2026 — a 3.2x increase in two years. That figure comes from Scott Brinker, who has long tracked the marketing technology landscape. The fastest-growing category is AI video: tools like Sora, Runway, and HeyGen saw usage climb 340% between 2025 and 2026.

And there's one more frontier: AI agents — systems that can run multi-step tasks on their own without a human watching. 34% of large marketing teams already have at least one agent running in production.

Where Is the Money Flowing?

Now that we've covered usage, let's talk money.

Searchlab combined data from McKinsey, Gartner, and HubSpot to run the numbers: AI tools and infrastructure now consume 19% of marketing budgets, and that figure is growing 28% per year. 63% of CMOs plan to keep increasing spend.

At that growth rate, AI will account for a quarter of marketing budgets by 2027.

Zoom out further: the global AI marketing market was worth $6.46 billion in 2018 and $57.99 billion in 2026 — a compound annual growth rate of 37.2%. Forecasts put it at $107.5 billion by 2028.

The mid-sized company view is even more tangible: the median monthly spend on AI tools by marketing teams rose from $1,200 in Q1 2025 to $3,400 in Q1 2026. A 183% jump in one year.

183%. That single-year growth outpaces what cloud services and SaaS saw when they first took off.

The money is real, and there's more of it every year. The problem has never been willingness to spend — it's whether the spending pays off.

The Time Savings Are Real

So is it worth it or not? Start with time.

HubSpot's AI Trends report found that the average marketer saves 6.1 hours per week thanks to AI.

Let me do the math for you. A 20-person marketing team, each person saving 6.1 hours a week, comes to 122 hours a week. That's three full-time hires appearing out of thin air. What could those three people do? Strategy, creative quality control, the work that genuinely requires judgment.

And the more senior you are, the more you save. Senior practitioners save 8 to 10 hours a week; heavy users reach 11 to 13 hours.

Output is climbing too. Companies using AI produce 42% more content per month. 84% of marketers say AI speeds up content delivery. The most striking case is blog writing: in 2024, 65% of marketers didn't use AI to write blog posts; by 2026, only 5% still don't.

Two years, sixty percentage points. The complete replacement of one specific job happened quietly, almost without anyone noticing.

The Returns Are Real, but There's a Strange Pattern

McKinsey's global AI survey broke returns down to the application level.

AI copywriting delivers an average ROI multiple of 3.2x — the highest of all. Personalization engines: 2.7x. Audience research: 2.4x. Ad copy: 2.3x. At the campaign level, AI-driven marketing campaigns deliver 22% higher ROI than traditional ones, 32% more conversions, and 29% lower customer acquisition costs.

Here's where it gets interesting.

In the world of enterprise software, the usual pattern is: the higher an application's return, the harder it is to implement, so the lower its adoption rate. AI in marketing runs completely backwards: copywriting, the highest-return use case, happens to be the easiest to pick up — which is why everyone is already doing it.

What does that mean?

The easiest pickings have already been shared out to everyone. The next wave of returns is hiding in the tough builds: personalization infrastructure, predictive audience modeling, agentic campaign management.

This work is hard and few are doing it, which is exactly why the value is still sitting there. BCG has estimated that AI can unlock between $0.8 and $1.2 trillion in value annually across sales and marketing. What the entire industry has captured so far is only a tiny bite.

The rest of that prize is stuck behind the execution barrier.

They Complain It's Slow — and Keep Paying

Now let's return to that "hard to say" from the beginning.

Why, with so many users and so much money spent, is the answer still "hard to say"? Two numbers from The CMO Survey, read together, are particularly painful.

91% of marketing leaders say generative AI adoption is moving "too slowly."

93% of marketing teams have still budgeted for continued AI investment.

Complain it's slow, keep paying; complain again, pay again. That's not a contradiction — that's the most honest picture of marketing in 2026. The cost of not using AI is now rising faster than the cost of using it badly. Nobody implements AI because it's easy. They implement it because without it, you lose faster.

Where's the bottleneck? Look at the numbers. Only 17% of marketers have received systematic AI training, while 58% name the skills gap as their number-one challenge. Only 13% dare to trust AI conclusions without human review. Gartner also found that 45% of top marketing executives admit AI has created more chaos than order within their teams.

The organization can't keep up with the tools. People are holding the machines back.

And there's a personnel shift underway: in 2025, 23% of agencies cut junior copywriter roles, and 31% plan to keep cutting in 2026. The hand doing the cutting turns right around and hires senior strategists who can keep AI under control.

Junior seats are shrinking; senior seats for people who can command AI are getting more expensive. Which side of that divide you plan to stand on is a choice you need to start making now.

Back to That Cup of Tea

Back to my CMO friend.

Toward the end of our tea, I told him: saying "hard to say" already puts you ahead of most people in self-awareness. The industry average is only 15% of the work handed to AI, and 74% of companies are still circling in front of the value barrier.

But the direction is clear: the easiest pickings — copywriting — have already been shared out to everyone. The next returns belong to the companies that fix the training gap, wire AI into their workflows, and dare to let AI touch the hard stuff like personalization and agents.

They also belong to the people who neither deify AI nor pretend it isn't there.

May every dollar of your AI budget come with a return you can clearly account for.