The $107 Billion Bet: Where AI Marketing Dollars Are Actually Going in 2026

The numbers are staggering. Global AI marketing spend is projected to exceed $107 billion by 2028, according to Statista. But here's what most boardrooms won't tell you: the majority of that investment is being poured into the wrong places.

I've spent the last seven years watching marketing budgets shift. From my time at BBDO and Leo Burnett to building Kastling, I've observed a consistent pattern. When new technology arrives, companies do the same thing: they automate what they already do instead of reimagining what they could do.

Where the $107 Billion Is Actually Flowing

Personalization Engines take 40%+ of budgets. This is the clear leader. Netflix's personalization engine alone saves them approximately $1 billion annually in customer acquisition and retention costs. When you can predict what a customer wants before they know they want it, everything changes.

AI-Generated Creative at Scale is the next frontier. Coca-Cola used generative AI not to make one ad, but to generate thousands of variations, testing messaging across demographics, regions, and platforms simultaneously.

The Distinction That Matters

Here's the critical insight that separates winners from the rest: the real ROI isn't in content generation. It's in decision intelligence. Most marketers are using AI to make more stuff. The velocity is impressive, but it's often noise.

The winners are using AI to make better decisions about what stuff to make. Which customer segment should we prioritize? What is the true probability this lead converts? Which message resonates with this specific audience?

The Execution Gap

The challenge isn't technology. The tools exist. The challenge is organizational. Most marketing departments were built around a model of content production. Inserting AI into this workflow creates friction, not transformation.

The companies winning in 2026 are restructuring around decision intelligence. They're asking their AI to sit upstream, in the planning phase, in the targeting phase, in the channel selection phase.