As enterprises race to operationalize AI, many are discovering that the real challenge is no longer access to models. It is the harder, less glamorous work underneath: building trusted data foundations, aligning systems, governing information flows, and ensuring AI outputs can actually drive measurable business outcomes.

That reality is becoming especially visible as organizations move beyond experimentation into automation and agentic AI. While excitement around autonomous decision-making continues to grow, leaders are increasingly confronting a fundamental question: Can AI truly scale without trusted data ecosystems underneath it?

In this first installment of a three-part series, former Qlik CEO Mike Capone joins Dr. Adita Karkera, Chief Data Officer, Government and Public at Deloitte, to discuss the evolving relationship between AI, analytics, governance, and enterprise transformation. Drawing from his experience, Capone outlines why the next phase of AI success will depend less on models themselves and more on the surrounding data architecture, governance frameworks, and operational trust organizations build around them.

From programmer to CEO: A different kind of technology leadership

Capone describes himself as a “non-traditional CEO,” explaining that his path to leadership did not emerge through finance or sales, but through technology and operations.

“I started as a programmer, but I had an affinity for figuring out how to marry technology with business outcomes,” says Capone.

His career journey moved from software development into enterprise IT leadership, eventually overseeing a multi-billion-dollar technology organization at ADP before transitioning into operational leadership roles that positioned him for the CEO path.

While many technology leaders remain within technical functions, Capone believes the industry is entering a period where more CIOs and CTOs will eventually move into CEO positions.

That transition, he says, ultimately came down to one core capability: understanding how technology translates into business transformation.

“The key to all of that has always been the ability to synthesize what technology can do to transform a business.”

Today, that same principle increasingly sits at the center of enterprise AI strategies.

Why AI projects are falling short on outcomes

According to Capone, organizations across industries are aggressively investing in AI, yet many are still struggling to generate meaningful returns.

“The market is discovering that the barrier to AI success is not plugging these models. That’s the easy part. It’s the data and systems around the model.”

This is where Capone sees the partnership between organizations like Qlik and Deloitte becoming increasingly important over the next several years. Enterprises are not simply looking for AI tools. They are looking for trusted ecosystems that can help operationalize AI responsibly inside complex enterprise environments.

According to Capone, organizations must focus on foundational capabilities, including:

• Data integration across fragmented environments
• Governance and trust frameworks
• Lineage and transparency
• Indexing and discoverability
• Operational readiness for AI workflows
• Business alignment tied to measurable outcomes

“You have to do the hard work,” Capone says. “You have to be urgent about building the right platform.”

Why the “walled garden” AI approach fails

One of the strongest themes throughout the discussion between Capone and Karkera centered on openness and ecosystem collaboration. Capone argues that enterprises increasingly reject closed AI models that require organizations to hand over all of their data in exchange for outcomes.

“The walled garden approach, where you say, ‘Hey, give me all your data, and I’ll give you the outcome that you want,’ doesn’t work.”

Instead, he advocates for a more collaborative approach that starts with the customer’s existing reality: “It’s much more meeting customers where they are, understanding what their landscape is, how they can be successful in their world, and then actually working backwards.”

That approach is particularly important because organizations are operating at vastly different levels of data maturity. As Karkera points out during the conversation, there is no universal AI maturity model that works for everyone.

Rather than forcing enterprises into rigid architectures, Capone suggests successful AI partnerships will increasingly revolve around helping organizations evolve from their current state without dismantling what already works.

Agentic AI raises the stakes

The conversation also explored the growing momentum around agentic AI and autonomous decision-making.

While Capone acknowledges that agentic AI is currently “all the rage,” he emphasizes that the concept itself is not entirely new: “We’ve had a product in the market called Qlik Automate for quite some time now.”

What has changed, however, is the acceleration of capabilities over the past 18 months and the growing willingness of organizations to automate actions directly from AI-driven insights.

“People are now anxious to start automated decision-making and actioning off of their analytics and AI insights.”

But Capone returns to the same warning: automation amplifies the importance of trust.

The concern is straightforward. Once organizations remove humans from critical workflows, errors become operational rather than theoretical.

“You’re closer to agentic than you think”

Despite the caution, Capone remains optimistic about how quickly organizations can evolve toward agentic capabilities, particularly those that have already invested in analytics and trusted data workflows.

Explaining why he believes organizations are “closer to agentic than they think,” Capone says many enterprises already have the foundational pieces in place to support AI-driven automation. In many cases, the groundwork, trusted analytics workflows, operational data environments, and established governance practices already exists even if organizations have not yet recognized how close they are to building on top of it.

“We work with customers who’ve built a good foundation around analytics, and they’re running their business off of it.”

For many enterprises, the transition may not require rebuilding infrastructure from scratch. Instead, it may involve extending trusted workflows with AI and automation layers on top.

“Maybe they’re just running it on dashboards or on reports, but they’re getting it done, and they’re trusting that data.”

That existing trust, he says, becomes the launch point for agentic AI adoption: “You can build on top of that platform to actually get to Agentic faster.”

Avoiding expensive AI rebuilds

Capone also cautions against organizations abandoning existing ecosystems in pursuit of entirely new AI stacks. “The alternative is to tear down everything you’ve built, to rip apart your foundation and trusted partnerships with the vendors that you got and start something else, and that’s expensive and risky.”

Instead, he positions the future of enterprise AI around augmentation rather than replacement:

• Build on trusted analytics foundations
• Extend existing workflows with AI
• Layer automation gradually
• Preserve institutional trust in data
• Focus on measurable operational outcomes

For Capone, the next five years of AI will not be defined by who adopts models the fastest. It will be defined by who operationalizes trust, governance, and decision intelligence most effectively at scale.

And increasingly, that appears to be where strategic partnerships between technology providers and advisory leaders like Qlik and Deloitte are positioning themselves: helping enterprises move from AI experimentation to durable operational value.

CDO Magazine appreciates Mike Capone for sharing his insights with our global community.