Why Marketing AI Tools Are Solving Problems Nobody Actually Has – News Round Up: 09/26-10/03

by Jeff Licciardi

AI Marketing Tools Are Building Solutions for Yesterday’s Problems – News Round Up: 09/26-10/03

The marketing technology industry has perfected the art of solving problems that don’t exist anymore. While marketers struggle with basic attribution, creative fatigue, and proving incremental value, AI vendors are racing to build tools that automate tasks nobody actually needs automated—or worse, that can’t be automated without breaking the thing that made it work in the first place.

This week’s industry news crystallizes a growing disconnect: the problems AI marketing tools claim to solve versus the problems marketers actually face in their day-to-day work. The gap isn’t just wide—it’s getting wider, and it’s costing brands real money.

The AI ROI Mirage: Where’s the Actual Value?

According to Ad Age’s deep dive on AI advertising ROI, marketers are finding returns in surprisingly mundane places—not the flashy creative generation tools dominating headlines. The real wins? Data analysis, audience segmentation, and campaign optimization. In other words, the boring stuff that’s been algorithmically driven for years anyway.

Meanwhile, Digiday reports that creative challenges and legal risks continue to loom large for AI adoption. Marketers are “warming” to AI—which is marketing speak for “we bought the tools because our CEO read about them, but we’re not sure what to do with them yet.”

The disconnect is stark: AI vendors are pitching creative automation and content generation at scale, while actual marketing teams are dealing with brand safety concerns, copyright uncertainty, and the realization that AI-generated creative performs marginally at best. The problems AI tools are designed to solve—”What if you could generate 10,000 ad variations?”—aren’t the problems keeping CMOs up at night. The actual problems are: “How do I prove my marketing spend drove incremental revenue?” and “Why does my brand feel increasingly generic?”

The Retail Media Reality Check Nobody Wants to Hear

If you want to see what happens when an industry collectively decides to solve the wrong problem, look at retail media networks. Digiday’s analysis shows that “the easy dollars are gone” as retail media approaches maturity. Translation: brands have already shifted their bottom-funnel search budgets to Amazon and Walmart, and now RMNs need to actually prove incremental value.

The problem? Retail media networks spent the past three years building infrastructure to capture search demand that already existed, rather than creating new demand. They solved for “where else can we put search ads” when they should have been solving for “how do we help brands reach customers in moments that actually influence purchase decisions.”

As one Digiday source puts it, “there’s a point of diminishing returns” with retail media—and we’re hitting it faster than anyone expected. Brands are now reshaping their entire organizational structures to manage retail media, which is corporate speak for “this is way more complicated than anyone promised, and we don’t have the headcount to manage 47 different retail media dashboards.”

The reckoning coming for retail media isn’t about consolidation or better measurement—it’s about the fundamental realization that copying Google’s 2010 playbook isn’t innovation, it’s just arbitrage with better first-party data.

When AI Vendors Discover Advertising (And Why That’s Awkward)

Perhaps nothing better illustrates the “solutions looking for problems” phenomenon than OpenAI’s pivot from “hatred to hiring” when it comes to advertising. The company that positioned itself as above the dirty business of ads suddenly discovered that burning billions in compute costs requires, well, revenue.

And then there’s Elon Musk outlining an “AI-led Grok future” for advertising on X, which would be more compelling if X’s advertising business wasn’t in freefall. The pitch seems to be: “What if the platform you’ve already abandoned had AI?” It’s solving a problem—”how do we make X advertising appealing again”—that exists only because of self-inflicted wounds, not market evolution.

These aren’t examples of AI identifying unmet marketing needs. They’re examples of AI companies realizing they need advertising revenue and hastily building tools to monetize their existing technology, regardless of whether those tools address actual marketing challenges.

The CTV Gold Rush Built on Shaky Attribution

AdExchanger reports that CTV will keep growing in 2025, which is probably true—but growth isn’t the same as solving problems. The CTV industry has spent years perfecting programmatic delivery and measurement infrastructure while largely ignoring the elephant in the room: most CTV attribution models are sophisticated lies.

Political advertisers are tripling down on CTV, which tells you everything you need to know. Political campaigns have notoriously poor measurement—they run for weeks, end on a fixed date, and rarely do serious post-campaign analysis. If CTV’s core growth segment is advertisers who can’t or won’t measure effectiveness rigorously, that should concern everyone else.

The problem CTV needs to solve isn’t “how do we get more ad impressions on smart TVs”—it’s “how do we prove incremental reach and actual business impact beyond last-click attribution models borrowed from digital.” Instead, the industry keeps building better pipes while the fundamental value proposition remains unproven.

What Marketers Actually Need (And Aren’t Getting)

Strip away the hype cycles, and most marketers are dealing with three core challenges: proving incremental value in an increasingly fragmented media landscape, maintaining brand distinctiveness amid platform homogenization, and doing all of this with flat or declining budgets.

The tools being built? They’re optimizing click-through rates, automating creative production, and adding new media channels to manage. These aren’t solutions to the core problems—they’re incremental improvements to tactics that may not matter in the first place.

The AI and ad tech industries are trapped in a cycle of building solutions for problems that made sense in 2019. Meanwhile, the actual challenges of 2025—signal loss, creative commodification, channel proliferation, and demonstrating true incrementality—remain largely unaddressed. The tools we have are impressive. They’re just impressively beside the point.

Until vendors start building for the problems marketers actually have—rather than the problems that are technically interesting to solve—we’ll keep seeing this pattern: cool technology, modest adoption, disappointing ROI, and a persistent feeling that we’re all running faster to stay in the same place.

Sources & References