The AdPipe Pulse Report 2026: Six Findings on AI in Marketing

We asked marketing, creative, and agency leaders how AI is really going. Then we read 21,038 ads to check.

Estimated Read Time: 4 minutes

Key Takeaways

• AI adoption isn't the question anymore. Proving it works is.
• The teams that can prove it do a few specific things differently, and none of them come down to having better tools.
• We also ran a creative diagnostic on 21,038 ads. What it found about how spend gets allocated is worth a second look at your own ad sets.
• The full findings, plus a playbook for marketing and creative leaders, are in the report.

Why We Ran the Pulse

Every marketing team has an AI story right now. Most of them start with a tool and end with "it's helping."

We wanted to know what sits underneath that. So this summer, we surveyed marketing, creative, and agency leaders about how they use AI, what it has delivered, and where they're spending next.

Then we did something most surveys don't. We read the creative itself: 21,038 ads and $81 million in Meta spend, each one compared head to head against the others in its own ad set.

Here are the six findings at a glance. The details, and the data behind them, live in the full report.

1. The Experiment Phase Is Over

Not one team we surveyed is sitting AI out. Most have moved past testing and into regular use.

That changes the conversation. "Are you using AI?" has a settled answer. The question that's left is harder, and it's the one the rest of the report is about.

2. Only 22% Can Prove It's Working

This is the headline. Most teams can feel AI helping. Very few can put a number next to it.

We call it the proof gap, and it's the defining AI story for marketing teams this year. The report breaks down what's holding teams back, and why the obstacle changes as they get further along.

3. Depth Is the Whole Game

The teams that can prove results aren't the ones with the best tools. They're the ones who went deeper.

There's a clear line in the data where proof starts to show up. It sits further out than most teams would guess. At the low end, no team could prove results at all.

Want to see where the line is? Get the Pulse Report.

4. Agentic and Predictive AI Are Stuck at "Heard of It"

More than half the field knows these tools by name. Far fewer use them.

The distance between buzz and use is one of the widest in the report. It says a lot about where the hype is running ahead of the work, and where the next real opening might be.

5. Budgets Are Moving Before Strategy

AI budgets are going up, and nobody we surveyed is cutting. The surprise is how many teams don't know where the next dollar is headed.

The report shows where the money is going, and the one place almost nobody is putting it.

6. Almost Nobody Is Doing Diagnosis

Only 12% of teams use AI to understand why their creative performs. It's the least-adopted use case in the survey, and about half the wish list for what teams want from AI asks for exactly that.

Marketers know what won. They rarely know why.

What the Creative Told Us

So we ran the diagnosis most teams aren't running. Computer vision read every ad frame by frame, and we set each one against the others in its own ad set.

One finding stopped us: how often the ad Meta funds most turns out to be one of the weakest in the set. The answer is in the report, along with what wins head to head, which styles fatigue first, and how $20 million ended up on the weaker ad.

Who Should Read It

Marketing leaders get a playbook for auditing ad sets and reporting creative in the metric their CFO already trusts.

Creative leaders get a playbook for briefing by funnel stage and matching style to the moment.

Anyone who gets asked "what's AI doing for us?" gets numbers to answer with.

Read the full findings. Download the Pulse Report 2026

The Pulse surveyed marketing, creative, and agency leaders between July 7 and August 26, 2026. The creative diagnostic covers $81.0 million in Meta spend over 90 days.