The overlooked leadership skill holding back AI value

AI has dominated the executive agenda for the past two years. The promise of productivity gains, the opportunity to orchestrate data across entire organizations, to improve employee and customer experiences, and to ultimately increase revenue is driving enterprises to make significant investments with high expectations for returns.

But those expectations are now being questioned as conversations turn from experimentation to results. Research from PwC found that 56% of CEOs reported neither increased revenue nor reduced costs from AI over the past 12 months, highlighting a growing gap between ambition and realized impact. Similarly, McKinsey found that while more than 60% of organizations are using AI to enable innovation, less than 40% are seeing meaningful enterprise-level financial impact. 

While those statistics get attention, they often lead to the wrong conclusion. The issue is not that AI won’t deliver meaningful business impact. It is that many organizations are missing a foundational leadership capability required to unlock it: Data curiosity.

Data curiosity emerges as a critical leadership skill

In my work partnering with leaders across Genesys on AI transformation, I’ve seen AI act as a powerful forcing function, not as a corrective, but as an accelerator of understanding. In conversations with other CIOs and technology leaders, I am seeing a clear pattern: The organizations seeing value the fastest are not the ones moving the quickest, but instead those who are asking better questions.

Over the years, I have seen technology advancements reveal leaders’ blind spots in how they enable their teams and deploy next-generation capabilities. AI is essentially having the opposite effect. Rather than creating new blind spots, it’s revealing existing ones. Now, I’m seeing leaders confronting the reality of their data strategies as previously hidden gaps are exposed. Inconsistencies, poor governance and misplaced confidence in data quality are being surfaced and what once went unnoticed in dashboards and reports is quickly becoming visible. As teams apply AI to real workflows, they have to examine more deeply how data is sourced, governed and connected to business decisions.

That’s why data curiosity is quickly becoming one of the most important leadership traits you can have in the AI era. Leaders who lean into asking the questions and foster curiosity amongst their teams to do the same begin to uncover the real drivers and roadblocks for AI results.

The challenge is that AI is too often treated as a technology deployment. The most successful transformations are not tool-led. Some of the most critical works sit with the CIO and IT leaders: redesigning workflows, strengthening data foundations, embedding AI into how the business operates. Those changes scale when leaders also reshape how they engage with the data.

When leaders cultivate curiosity about their data, they create an environment where employees feel empowered to question outputs, challenge assumptions and continuously improve results. That mindset becomes the foundation for scalable AI value.

This requires going beyond mindset alone. Data literacy programs should be built alongside AI literacy, ensuring employees across the organization and at every level understand how data is structured, governed and used to drive decisions. When people know how to interrogate the data behind the systems, they are more likely to trust, challenge and improve the outcomes that those systems produce.

Data curiosity changes the trajectory of AI transformation

A common misstep I am seeing is organizations shoving poor-quality data into systems and hoping AI will fix it. That’s just not going to happen. Bad data leads directly to bad outputs. That’s why so many early AI initiatives fell short. As generative AI really took off, many organizations got a sense of FOMO, and hesitancy in adopting AI quickly shifted to “we need to do this now.”  But organizations overestimated how good their data was. When the outputs didn’t meet expectations, their instincts were to question the tool or the models. Data curiosity can flip that response.

Instead of asking “why is the AI wrong,” leaders should ask more productive questions:

• Who owns this data?
• How current is it?
• Where are the gaps?
• Does it accurately represent the business? 

Asking those questions can change the trajectory of the entire transformation. Most importantly, they lead to action. Instead of assuming data is “good enough”, teams will quickly be able to identify inconsistencies across systems, gaps in governance or outdated or incomplete datasets, and course correct. The conversation moves away from “which model should we use?” to “Do we actually have the right data and processes in place?” From there, AI results become more reliable, insights become more actionable and trust in the system increases.   

Equally important, these questions help prevent teams from switching off critical thinking once the AI works. By modeling this behavior, teams are more likely to challenge outputs instead of blindly accepting them, apply human judgment more rigorously and catch errors, bias or drift earlier. Monitoring becomes continuous, and that results in fewer failed pilots, faster paths to ROI and stronger alignment to business outcomes.

Data curiosity in action: A customer experience lens

The use of virtual agents and copilots in the customer experience industry is a clear example of the value of data curiosity. As more organizations adopt AI to power customer engagement, there is increasing awareness that the quality of consumers’ experiences is tied to the data quality, completeness and ability to orchestrate it across systems.

Consider a virtual agent handling a refund. If it’s relying on knowledge base articles that have outdated policies, it may confidently provide guidance that is no longer accurate. Instead of improving efficiency, the experience introduces friction. That miss can quickly compound as customers encounter additional friction trying to resolve their issue. What often follows is escalation to a human agent for something that could have easily been a self-service interaction, along with a frustrated customer whose trust has been eroded. And that single moment can have long-term impact. My company’s research found that more than half of consumers will abandon a brand after as few as two poor experiences.

While there have been instances where bots provided inaccurate information, those moments serve as meaningful reminders that AI data quality, context and governance are essential in delivering loyalty-building experiences. Whether supporting human agents or engaging directly with customers, AI systems rely on the information they are trained to provide personalized and helpful responses and efficiently solve customers’ inquiries. When responses fall short, the root cause is rarely the interface; it’s the data foundation behind it.

This is where data curiosity can drive business results. Asking the right questions puts organizations in a better position to improve both AI performance and customer outcomes, driving consistent, personalized and trustworthy experiences that ultimately result in better business outcomes.  

Where AI value is truly built

As organizations continue to invest in AI, realizing that unlocking the real value behind it may mean restarting their master data management and governance. The reality is, AI doesn’t introduce new problems as much as it brings existing ones into sharper focus. It exposes the fundamentals leaders need to understand—about their data, their processes and their decision-making—in order to scale AI responsibly.

That’s what makes data curiosity so critical. AI will scale whatever it’s given. If the inputs are incomplete, outdated or disconnected, those issues will be amplified. But when the data foundation is strong, AI can unlock new levels of efficiency, insight and experience across the business.

For leaders, this requires a shift in focus. It’s not enough to evaluate outputs or chase the next model advancement. The real work is in understanding and continuously improving the inputs—how data is sourced, governed and connected to outcomes.

And, while data curiosity is foundational, it doesn’t operate in isolation. CIOs are encountering other familiar challenges in transformation, like workflows that we’re redesigned, lagging change management, talent and skill gaps, and complex integrations. Add in vendor lock-ins, unclear ownership and competing priorities, and it becomes clear why so many initiatives struggle to scale.

What will separate organizations that break through is how they respond to those challenges. Leaders who are deeply curious about their data and better equipped to tackle each of them. Rethink workflows, identify gaps, question assumptions and continuously seek ways to improve your business’s operating model.

For leaders asking what to do next, initiating a clear operating rhythm, where AI initiative is treated like a production, with clear checkpoints, can make this real:

• Data lineage: Confirm where the data comes from and how it flows through the system
• Freshness: Audit how current the data is and how often it’s updated
• Quality: Ensure controls are in place to detect and correct errors, bias and drift
• Ownership: Assign who is accountable for the data and its outcomes

Organizations that embrace this mindset will move beyond experimentation and into impact. Not because they adopted AI faster, but because they built the conditions for it to deliver value.

This article is published as part of the Foundry Expert Contributor Network.
Want to join?