AI's Scale Problem: 76% of Leaders Can't Govern Employee AI Use

The promise of enterprise AI is hitting a stark reality: while generative and agentic AI systems are rapidly being deployed, a staggering 76% of data leaders admit their organizations lack the governance frameworks to manage how employees actually use these powerful tools. This "trust paradox," identified in a new survey by Informatica, highlights a critical disconnect between technological adoption and the essential human and procedural infrastructure required to harness AI responsibly and at scale.

This inability to govern employee AI usage is a primary roadblock preventing many companies from moving beyond experimental pilots to widespread production deployment. Chief Data Officers (CDOs), now pivotal figures at the intersection of data governance, AI strategy, and workforce readiness, find their decisions directly impacting whether AI initiatives succeed or falter. The findings underscore that the bottleneck isn't infrastructure, but rather a profound gap in data and AI literacy.

The Widening AI Literacy Chasm

The adoption of AI technologies has been explosive. Generative AI deployment has surged from 48% to 69% of enterprises in just one year, with nearly half (47%) now running agentic AI systems capable of autonomous action. This rapid influx of sophisticated tools into the workforce, however, has outpaced the development of necessary oversight and training.

"The gap now is just, can you trust the data to set an agent loose on it?" Graeme Thompson, CIO at Informatica, told VentureBeat. "The agents do what they're supposed to do if you give them the right information. There's just such a lack of trust in the data that I think that's the gap." This lack of trust stems not from inherent flaws in AI, but from an organizational inability to ensure the data feeding these systems is accurate, secure, and used ethically.

Graeme Thompson dismisses the notion that hardware or software infrastructure is the primary constraint. The technology is available and functional; the challenge lies in the human element and organizational processes. He draws a parallel to amateur athletes blaming their equipment for poor performance, suggesting that most organizations are far from hitting the limits of their current technological capabilities.

The survey data strongly supports this view. When asked about investment priorities for 2026, the top three concerns are people and process-oriented: data privacy and security (43%), AI governance (41%), and workforce upskilling (39%). This indicates a clear strategic pivot away from chasing new technologies towards building the foundational human capital and governance structures needed to support them.

Five Lessons for Navigating the Trust Paradox

Informatica's research, combined with Thompson's practical experience, offers actionable insights for data leaders grappling with these challenges. The core issue isn't a lack of AI tools, but a deficit in the people who use them and the processes that govern their deployment.

Firstly, organizations must stop chasing infrastructure and fix the people problem. The overwhelming majority of data leaders (75%) report that employees need upskilling in data literacy, and 74% require AI literacy training. Thompson emphasizes that it's more efficient to train existing employees, who understand company processes and data, in AI rather than hiring expensive AI experts unfamiliar with the business context. This "people gap" is the true bottleneck.

Secondly, the Chief Data Officer role must be an execution function, not an ivory tower. Thompson's structuring of Informatica, with the CDO reporting directly to the CIO, ensures data governance is embedded as a proactive, "get things done" function. This alignment under a common leadership structure fosters shared priorities between data teams and application owners, mitigating the silos that often prevent pilots from scaling.

Thirdly, literacy must be built outside IT teams. AI competency needs to permeate all business functions, not just technical departments. Thompson points to the marketing operations team as a key partner, understanding how AI can strategically enhance their efforts, such as optimizing marketing spend. This business-side literacy creates a demand pull for AI, driven by perceived strategic value rather than just efficiency gains.

Fourthly, AI should be pitched as strategic expansion, not cost reduction. Data leaders have an opportunity to shift the perception of IT from a cost center to a strategic enabler. Thompson expresses disappointment that many are immediately framing AI in terms of productivity savings. Instead, he advocates for pitching AI's ability to remove headcount constraints and enable new market reach or initiatives that were previously cost-prohibitive. This reframes AI from operational efficiency to a driver of strategic growth.

Finally, organizations should go vertical first, then scale the pattern. Rather than waiting for perfect horizontal data governance, companies should identify a high-value use case, build the complete governance, data quality, and literacy stack for that specific workflow, validate results, and then replicate the successful pattern. This approach delivers tangible production value incrementally while building organizational capability.

The Path Forward: From Pilots to Production

The "trust paradox" is a critical hurdle for enterprise AI. As organizations continue to embrace generative and agentic AI, the focus must shift from the deployment of new technologies to the cultivation of a data-literate, AI-aware workforce operating within robust governance frameworks. The infrastructure is largely in place; the real work lies in empowering people and refining processes. By addressing the human element and implementing strategic governance, companies can finally bridge the gap between AI potential and scalable, trustworthy deployment, ensuring that AI truly drives strategic expansion rather than becoming a compliance headache.