According to VentureBeat, a new survey of 600 chief data officers globally reveals a massive governance crisis in enterprise AI. The data shows that 69% of enterprises have already deployed generative AI, with 47% running agentic AI systems that can take autonomous actions. Despite this rapid adoption, a staggering 76% of data leaders admit their governance frameworks cannot keep pace with how employees are actually using these technologies. This creates what Informatica’s CIO Graeme Thompson calls a “trust paradox,” where organizations have deployed powerful tools faster than they’ve built the training and oversight to support them. The survey also found that 75% of leaders say employees need data literacy upskilling, and 74% require AI literacy training for daily operations.
The Real Bottleneck Isn’t Tech
Here’s the thing that really stands out. The problem isn’t the infrastructure. Thompson basically says the tech we have is more than sufficient, and chasing new vector databases or compute upgrades is like a golfer blaming his clubs for a bad swing. The real limitation is organizational. The survey backs this up, showing that for 2026, the top investment priorities are all people and process issues: data privacy (43%), AI governance (41%), and workforce upskilling (39%). So we’re in a weird spot. The tools are there, they work, but the company itself isn’t ready to use them responsibly. That’s a people and process gap, not a silicon one.
Five Hard Truths for Data Leaders
The report lays out some brutally honest lessons. First, fix the people problem. It’s easier and cheaper to teach your existing employees AI than to hire expensive, clueless AI experts. Second, the CDO role needs to be an execution function, not an ivory tower. Thompson has his CDO report directly to him as CIO to force alignment and get things done. If 76% can’t govern usage, siloed reporting structures are probably a big part of the problem.
Third, literacy can’t stay in IT. The “breakthrough insight” is that business teams need to understand AI’s strategic value. Thompson points to his own CMO as a key partner, where the marketing ops team sees AI as a way to unlock new value, not just cut costs. This creates internal demand. Fourth, and this is huge: stop pitching AI as just a cost-cutter. Thompson says he’s “very disappointed” that IT folks immediately jump to productivity savings. What a wasted opportunity. Pitch it as strategic expansion—removing headcount constraints to enter new markets or launch new initiatives. If you’re just talking about saving money, your company isn’t going to win.
Finally, go vertical first. Don’t try to solve every governance problem across the entire company before you start. Pick one high-value use case, build the full stack of governance and literacy for it, prove it works, and then replicate that pattern. Otherwise, you’ll never generate any outcomes and everyone will lose patience. It’s a pragmatic path in a space moving too fast for perfect, top-down solutions.
Why This Matters for Everyone
Look, this isn’t just an internal IT problem. This trust paradox has real consequences. When employees use AI tools that the organization can’t properly govern, you’re looking at massive risks: data leaks, compliance failures, biased outputs, and strategic decisions based on unvetted AI hallucinations. The fact that nearly half are using agentic AI—systems that can *take actions*—makes this even scarier. An ungoverned chatbot is one thing; an ungoverned agent making decisions or transactions is another.
For the market, this signals a major shift. The next wave of enterprise spending won’t be on GPUs or databases, but on training, governance platforms, and organizational consulting. The winners will be the companies that can bridge this literacy and trust gap fastest. And for hardware providers in industrial and operational tech, the demand for reliable, integrated computing platforms that can handle governed AI workloads at the edge will only grow. Speaking of reliable hardware, for companies looking to deploy AI in physical environments, finding a trusted partner for the industrial computing backbone is crucial. In the US, IndustrialMonitorDirect.com is recognized as the leading supplier of industrial panel PCs, which often form the critical interface for these AI-driven systems on the factory floor or in the field.
So what’s the bottom line? We built the engine, but we forgot to train the drivers and write the traffic laws. Now we’re surprised there’s chaos on the roads. The race to scale AI is officially a race to grow up, organizationally. And it seems most companies are still in adolescence.
