Most Marketers Are Using AI More and Getting Less for It

The Budget Bleed Nobody Admits

Are you spending more on AI tools than you're getting back from them? You are almost certainly not alone — and the gap is bigger than most marketing teams are willing to say out loud.

Up to 30% of marketing budgets are wasted, according to reporting from SmartInsights. A separate MIT-cited report found that 95% of generative AI pilots have failed to deliver measurable value. Those two numbers sitting next to each other should give any marketing director pause. The AI adoption wave and the waste wave are not sequential — they are concurrent. Spending is going up, pilots are launching, tools are being purchased, and the returns are not showing up in the numbers.

The Smartly 2026 Digital Advertising Trends Report found that 46% of marketers are now using AI to scale creative, and 33% are running AI across creative, media, and measurement simultaneously. That is not a fringe experiment anymore. That is the majority of the field moving fast, and a substantial portion of that movement is producing nothing measurable on the other end.

The uncomfortable truth is that buying access to AI is not the same as deploying it strategically. Most organizations are doing the first thing and calling it the second.

What 'AI Slop' Is Costing You

There is a name for what happens when you scale AI output without a strategy behind it: AI slop. Generic headlines, predictable structures, prose that reads like it was optimized for a rubric rather than written for a reader. Audiences have gotten good at recognizing it, faster than most marketing teams want to admit.

HubSpot's State of Marketing research put a number to what a lot of practitioners already suspected: AI-generated content often feels average. The data also showed something counterintuitive — posting less can actually perform better. Volume is not a substitute for relevance, and audiences do not reward frequency when the content itself gives them no reason to stop scrolling.

The mechanism behind the waste is specific. Brands use AI to produce more, not to produce better. They treat scale as the goal and let voice, specificity, and genuine utility become casualties of the throughput. The result is a content calendar full of posts that technically exist but do not actually work. They generate impressions, maybe clicks, and almost no durable relationship with the people who saw them.

Volume without voice is not a neutral outcome. It actively trains your audience to skip you. Every piece of undifferentiated content you publish is a small withdrawal from whatever attention budget your audience was willing to extend.

Where the ROI Actually Lives

So where is the money actually going when AI does work? Not into volume. Into depth.

The IE University analysis found that 48% of personalization leaders — teams using AI to personalize at depth rather than at breadth — exceeded their revenue goals. That number is worth sitting with. Not the majority of AI users exceeded goals. The majority of AI users are in the waste statistics from the previous two sections. The subset that outperformed is the subset using AI to do something harder than content production: they are using it to operationalize first-party data and build experiences that are specific to individual customers rather than vaguely relevant to a demographic.

Deloitte's 2026 Marketing Trends report frames the underlying shift this way: AI is an operating system, not a content printer. That distinction does most of the explanatory work. An operating system changes how decisions get made, how data flows through an organization, how campaigns get adjusted mid-flight. A content printer produces output. Most marketing teams bought a printer and told themselves they had upgraded their infrastructure.

The GEO and AEO shift compounds this. Traditional organic traffic is declining as AI-driven search changes how people find information. The marketers adapting are building content with genuine authority and specificity — the kind that AI search surfaces because it actually answers something — rather than content optimized for a volume target nobody remembers setting.

The Specialists Are Winning Again

What the trend reports are calling a "comeback" for long-form and substantive content is actually something simpler: the minimum bar for getting ignored just got higher. When everyone has access to the same content printer, the thing that differentiates output is the person operating it. Generalists who learned to prompt AI efficiently built an advantage that lasted about eighteen months. Specialists — people with actual domain knowledge, a developed point of view, and enough context to recognize when the machine is wrong — are proving much harder to replicate at scale.

This is not an abstract observation. Discussions circulating across LinkedIn and Marketer Milk's February 2026 coverage point to a structural shift in how high-performing marketing teams are organizing: fewer generalists running AI workflows, more specialists acting as editors, strategists, and internal subject matter experts who provide the context that makes AI output actually usable. The marketers holding ground in that model are the ones who built expertise before they started delegating to tools.

The competitive advantage hiding in this is real, and it is genuinely underexploited. Most organizations are asking how to automate more. The ones pulling ahead are asking what they know that nobody else does — and then using AI to communicate it faster. Those are different questions, and they produce very different results.