91% of Marketers Use AI — Only 41% Can Show What It Earned Them

The other day I was having dinner with a friend who works in marketing, and somewhere in the conversation he let out a sigh.

"Our whole team is using AI now," he said. "Every day. Can't live without it."

So I asked him: "And how much extra money has AI actually made you?"

He paused for a few seconds. "...We're definitely producing faster. But as for how much more we're earning — I've never actually sat down and calculated it."

I figured he wasn't the only one who'd never done the math.

The entire industry has been doing the AI math in a fog.

Before you rush to argue, let me put a few numbers on the table.

1. What "Everyone Is Using It" Actually Means

Jasper's 2026 annual report surveyed 1,400 marketers.

91% are now actively using AI.

Let that sink in: in the marketing world, virtually no one can leave AI out of the picture anymore.

A year ago, that number was 63%.

From 63% to 91% in a single year. Read that growth rate again. In all the years I've pored over marketing reports, I've never seen a tool penetrate an industry this fast.

Canva goes even further: 97% of marketers use it daily.

McKinsey takes a different, bigger-picture angle: 88% of the companies they surveyed already use AI as a daily working tool in at least one function.

So here's what the first batch of numbers is telling you:

"Whether to use AI" is no longer a question. The conversation moved on long ago — to whether you use it well, and whether the money ever comes back.

2. Efficiency Is Up. The Accounting Is Not.

Good news first.

ZoomInfo's research says marketers are 44% more productive because of AI, saving an average of 11 hours a week.

Jasper says half of marketers are delivering faster. 75% report higher job satisfaction, and 45% have seen operating costs come down. On McKinsey's side, 80% of companies list "boosting efficiency" as their No. 1 goal for AI.

All good news, right?

Then you look one layer deeper and you stop cold.

Still in Jasper's same 2026 report: only 41% of marketers can prove "what return AI actually delivered for me."

Last year, it was 49%.

Sit with this pair of numbers: usage climbed from 63% to 91%, while the share of people who can do the accounting dropped from 49% to 41%.

Moving in exactly opposite directions, right?

And here's the more galling comparison: teams that have figured out the math are getting frightening returns. According to Jasper, 60% of teams that changed how they measure got two to three times the ROI.

So yes, AI can make money. Most people just haven't turned it into a ledger they can actually read.

What does "doing the math" even mean? It means every "we're faster, we produce more" you've ever said can finally be converted into one answer: "here's how much more I earned."

And that's exactly the sentence most people are stuck on.

3. What Yardstick Are People Using?

In 2025, HubSpot ran a survey of 1,500+ marketers asking how they prove AI creates value.

The answers are revealing:

Productivity is up — 65% cite that as their proof. Saving time — 58%.

Then it tapers off: better personalization, 39%; deeper insights, 38%. Only 37% land on "project-level ROI."

Notice something? People are measuring "how busy am I," not "how much am I earning."

But the numbers can be calculated. Take the ZoomInfo case: a company used AI for personalized outreach, and email open rates rose 30% while conversion rose 20%.

Then there's MajorKey: after adopting predictive analytics, ad ROI rose 40% and marketing costs fell 15%.

To be clear, these are case studies, not averages. But they prove the point: the math can be done — and the people who do it are walking away with outsize returns.

4. Where Should AI Be Doing Its Best Work?

Where do you think the highest-value use is — making posters?

No. It's creative content production.

Gartner asked 418 marketing leaders (between July and September 2024) and found that 77% of organizations using generative AI are applying it to content and creative work. Among high-performing teams, that share jumps to 84%.

You can probably guess why: copywriting and visuals are instantly producible, quick to iterate, and cheap to get wrong.

Beyond that, 48% of organizations use it for strategic analysis. 47% say AI pays off most visibly in "post-campaign evaluation and reporting."

Salesforce adds one more detail: 75% of PPC professionals have AI lend a hand writing their ads.

See the pattern?

Strong teams wire AI into a full chain, from creative all the way to post-campaign review. Average teams treat it like a tireless intern.

5. The Ranks Are Reshaping. The Gap Remains.

Use it long enough, and the org chart changes.

65% of teams have created a dedicated AI role. There's also something interesting: one out of every three marketers is quietly taking on AI strategy, policy, and governance — no title change, no new headcount, just quietly absorbing it.

But the gap on the other side is still wide.

Gartner says 27% of CMOs admit their marketing teams barely use generative AI at all. 87% of CMOs hit a performance pitfall in ad campaigns last year, and 45% abandoned their campaigns midway.

The surface penetration rate is 91%. But walk into an actual company and many departments' AI deployment is still sitting in "pilot mode."

6. Then Let's Scale. What's the Holdup?

Don't worry — the blockers are perfectly clear.

In the 2025 survey, the first obstacle is data privacy, 21%. The second is output quality, 19%.

And McKinsey's conclusion cuts even deeper: two-thirds of organizations haven't started "scaling" at all.

Even the hottest word in the industry — AI agents — tells the same story: 62% of companies are still just "trying it out." Only 23% have actually rolled agents out at enterprise scale. Push agents into a single business process and you hit a ceiling of about 10%.

In one sentence: that vision of "fully automated workflows" is still a long way from you.

7. The Most Galling Number: Trust

Everything so far has been about companies. This last set lands directly on all of us.

Salesforce's data: in 2023, 58% of consumers trusted companies that "use AI." And now?

42%.

From 58 down to 42 — a 16-point slide, three years in a row.

Don't rush to say consumers just don't get the technology. They aren't afraid of AI itself. They're afraid that behind the AI, no one is willing to take responsibility.

So governance — an old, tired word — suddenly became the fastest-rising buzzword of 2026. The wider AI is rolled out, the sturdier the shoulders that have to answer for it. The bigger the rollout, the more people have to step forward and face the questions.

8. The Most Intense Users of All: Social Media Marketers

When it comes to intensity of use, nobody outworks the social media crowd.

This past year, Sociality's survey laid out the data in fine detail: 89.7% of social media marketers use AI every day, or at least several times a week.

What are they using it for?

Analyzing data, 59.5%. Finding topics, 59.5%. Writing copy, 45.9%. Creating graphics and editing videos, 40.5%.

As for tools: nearly seven in ten use chatbots, six in ten use AI visuals, and four in ten use text generation.

Here's the key set of numbers: 28.2% say the majority of their posts are written by AI. Yet 78.4% drastically rewrite their posts before publishing.

And 44.7% believe AI content performs better.

Lay those three numbers side by side and there's no contradiction at all:

AI drafts. You hold the quality gate. AI produces a first pass; you bring the judgment. The fact that seven or eight out of ten people heavily revise before posting tells you everyone knows the truth: the version you can publish — and put your name on — is the one your own brain actually touched.

So the question has never been "should I use AI." It's which team you're on. The team that posts without revising, or the team that revises until it's something you're proud of.

Ending

Writing this, I think of that friend again.

His situation is a snapshot of the entire industry in 2026. On one side, 91% of people are sprinting with AI; on the other, 41% can't even tell you what their own gold is worth.

If you happen to be standing outside that 41%, I'd love to send him the same message — and you too:

Using AI no longer counts as a skill. Being able to show what AI earned you — that's your own skill.

In the second half of 2026, keep the math a little more precise. And tonight, may you not lose sleep over this ledger.