Are Enterprise AI strategies collapsing prematurely under the weight of broken data and bad priorities?

Most enterprises can’t use the AI tools they’re buying. Broken data, scattered systems, and poor prioritization are sinking strategies before they start. This isn’t a technology problem, it’s a failure to fix the foundation before chasing the future.

Ali Shoukat is a Senior Manager of Enterprise Architecture at Hilton, where he focuses on aligning platforms, data, and teams to drive transformation at scale. He offers a clear view into why so many AI strategies are already falling apart.

All cents, no sense: "The biggest blocker isn't technology, it's a people problem," Shoukat says. "When leaders say fixing their foundational data costs too much, they reframe it as a budget problem. But it really isn’t. It’s a you problem, because you made the choice to spend millions on a TV ad that lives for two months instead of investing in a data solution that is everlasting." The issue isn’t financial, it’s a failure to prioritize what actually builds long-term capability.

Mine your own gold: Before chasing AI, Shoukat says businesses need to fix what they already own. “AI is only as good as the data you feed it,” he warns, and too many companies are sitting on a gold mine they refuse to clean up. Instead of investing in their first-party data, they panic, get overwhelmed, and buy less-effective lists from brokers. “You’re sitting on a mine that you refuse to fix.”

Hilton took a different path: it centralized standards and platform choices, then federated responsibility to domain experts across the organization. The result? A structured, scalable system where clean data is ready when it’s time to activate AI.

One fish, two fish: Shoukat suggests framing the investment using the "fish model." A "fat fish" represents a large, upfront investment that shortens the adoption time, while a "skinny fish" means spending less over a longer period. Finding the right balance, he notes, is the difficult but essential strategic choice that separates successful transformations from failures.

But why, though: There's a fundamental flaw in how businesses approach AI adoption: they jump straight to what the technology can do, without first aligning on why they’re pursuing it. That shortcut creates confusion and skepticism. “We are emotional and social creatures, so we care a lot about the story and how we get there,” he says. “That’s what’s missing and creating this economy of ‘should I or should I not?’” When companies fail to connect the purpose to the path, many employees remain stuck on the sidelines, either unsure of the value or too hesitant to engage.

Raising trouble: Shoukat likens AI to a child we’re all raising, but without rules. “If you’re a parent, you provide guardrails so your child can foster in a good environment and develop healthy core values,” he says. “We haven’t given AI that. The guardrails we’ve put on it are not the Ten Commandments. There is no moral compass.” Without a shared value system, he warns, AI will absorb our worst online behaviors and treat them as norms worth replicating.

The weakest link: A great customer experience isn’t marketing’s job or IT’s job; it’s everyone’s responsibility. Shoukat has seen companies pour millions into sales and branding, only to lose customers at the last mile. “When asking customers why they didn’t come back, they'd say, ‘Getting the product up and running was a nightmare,’” he recalls. “You did the marketing, the sale, the branding—but your supply chain still runs on a 1980s process. No one thought onboarding could be the thing that kills the business. That’s because we still treat CX like a function, not a cross-functional mandate. It’s time every department owns the outcome."