Why Data is Like Fire: The AI Accountability Crisis No One Talks About

Executive Editor Tim King explores why enterprise AI initiatives are failing without modern data accountability, why speed has become the defining competitive variable in the AI era, and why organizations that continue treating data as a byproduct risk falling permanently behind.

AI Has Exposed the Enterprise Data Problem

Artificial intelligence did not create the enterprise data problem.

It exposed it.

During a recent episode of Insight Jam, Solutions Review President Doug Atkinson sat down with Sujay Dutta to discuss one of the least understood realities of the AI era: most organizations are attempting to scale artificial intelligence on top of data foundations that were never designed for real-time AI execution.

The conversation cut directly through the hype surrounding enterprise AI adoption and focused instead on a harder truth many executives still avoid confronting.

Most organizations do not actually have their data house in order.

For years, companies were able to survive despite fragmented systems, inconsistent governance, poor data quality, siloed ownership structures, and incomplete visibility into how information moved across the enterprise. Organizations compensated with time, manual intervention, and human reconciliation. Teams cleaned reports manually. Analysts fixed errors downstream. Finance departments spent weeks validating numbers before quarterly reporting cycles closed.

AI changes that equation completely.

As Dutta explained throughout the discussion, artificial intelligence has fundamentally compressed the speed dimension of business. AI systems no longer wait for organizations to reconcile messy data environments over weeks or months. Agents, copilots, automation systems, and real-time AI workflows require accurate, compliant, high-speed data immediately.

That is where many enterprise AI strategies begin breaking down.

AI Is Turning Data Into a Competitive Survival Issue

One of the most important ideas from the discussion centered on what Dutta describes as “data intensity.” Historically, organizations focused heavily on data quality and, more recently, compliance. AI introduces a third variable that many enterprises still underestimate: speed.

In the pre-AI era, organizations could often compensate for poor data systems through human effort. A reporting process might take weeks. Teams could manually validate outputs. Operational delays were frustrating but survivable.

AI-native environments operate differently.

An AI system executing customer support workflows, analytics recommendations, operational decisions, or agentic automation requires trusted data in real time. Organizations no longer have the luxury of “fixing the data later.” AI systems accelerate both good and bad decisions simultaneously.

That distinction matters because many executives still treat AI primarily as a technology conversation when it is increasingly a data accountability conversation instead.

Dutta repeatedly argued that organizations continue making a dangerous mistake by treating data as a byproduct of applications, processes, or departments rather than as an operational pillar of the business itself. That mindset may have been survivable in slower-moving operating environments. It becomes increasingly dangerous in AI-native environments where execution speed itself becomes a competitive differentiator.

As Atkinson noted during the conversation, data is no longer simply operational infrastructure.

It is table stakes for survival in the next decade.

Why Data Is Like Fire

Perhaps the strongest concept introduced during the discussion was Dutta’s analogy that “data is like fire.”

The comparison works because data now carries the same duality fire has always represented.

Handled properly, it powers organizations.

Handled poorly, it destroys them.

High-quality enterprise data can accelerate productivity, improve decision-making, unlock AI capability, increase operational speed, strengthen customer experience, and create entirely new business models. But bad data — especially when amplified through AI systems — creates the opposite effect.

AI does not simply automate work.

It automates mistakes at scale.

A flawed recommendation engine, inaccurate operational model, hallucinated AI workflow, or corrupted decision process can now propagate across an organization far faster than traditional human systems ever allowed.

That is why Dutta argues enterprise data accountability must move out of isolated technical silos and into executive leadership itself. In the book Data as the Fourth Pillar, he proposes that organizations treat data as an operational pillar equal to people, process, and technology.

That framing becomes increasingly important because many organizations still position data leadership as primarily an IT responsibility rather than an enterprise transformation function.

The Chief Data Officer May Become One of the Most Important Roles in Business

Another major theme throughout the discussion centered on organizational structure.

Dutta argued that enterprises serious about AI transformation may need to fundamentally rethink where data leadership sits within the business. Specifically, he proposed that Chief Data Officers should increasingly report directly to the CEO rather than operating under traditional IT structures.

The reasoning is strategic.

AI transformation is no longer isolated to technology departments. It impacts:

• Operations
• Workforce design
• Customer experience
• Analytics
• Compliance
• Product development
• Supply chains
• Decision-making
• Competitive strategy

This creates a governance challenge many organizations are still struggling to navigate.

Traditional enterprise structures evolved during slower operational eras. Departments optimized locally. Data ownership remained fragmented. Technology teams controlled systems while business teams controlled outcomes. AI increasingly collapses those distinctions.

The organizations moving fastest are often those treating data leadership as a cross-functional business capability rather than a backend infrastructure role.

That shift may become increasingly necessary as enterprises move toward agentic AI systems capable of executing workflows autonomously across multiple business functions.

AI Is Lowering the Barrier to Disruption

One of the most compelling parts of the conversation focused on competitive disruption.

Atkinson repeatedly challenged the idea that large organizations can safely assume their existing market position protects them from AI-native competitors. Dutta largely agreed, arguing that AI dramatically lowers the cost and speed required to launch new products, test new ideas, scale services, and compete operationally.

This may become one of the defining business realities of the next decade.

Historically, large enterprises benefited from scale advantages, infrastructure ownership, capital access, and operational complexity that created meaningful barriers to entry. AI compresses many of those advantages.

A small company equipped with AI-native tooling, cloud infrastructure, AI-assisted development, and high-quality data can now build, iterate, and scale dramatically faster than previous generations of startups could.

That creates enormous pressure on legacy organizations still operating under slower decision-making structures and fragmented data environments.

As Dutta noted, organizations increasingly risk a “Kodak moment” if they continue assuming their current market position guarantees long-term survival. AI-native competitors may emerge faster, operate leaner, and scale more efficiently than many established enterprises expect.

The AI Era Will Create Winners and Losers

The conversation also confronted a reality many executives still discuss cautiously: AI-driven efficiency gains will almost certainly reduce portions of the human workforce over time.

Dutta acknowledged the tension directly.

Organizations pursuing AI transformation are ultimately pursuing speed, efficiency, scalability, and productivity gains. In many cases, those gains reduce the need for certain categories of labor. At the same time, AI may create entirely new forms of work, new organizational models, and new management structures centered around supervising AI systems themselves.

The broader point was not that AI eliminates humans entirely.

It changes the abstraction layer at which humans operate.

Manufacturing already demonstrated this transition decades ago. Modern factory floors rely heavily on automation while humans increasingly supervise, manage, and optimize machine-driven workflows. Dutta believes knowledge work may follow a similar path as AI agents become more capable inside enterprise environments.

That transition may not happen evenly.

Some organizations will adapt successfully. Others will fail to modernize quickly enough. Some workers will develop AI-native skills early. Others may struggle to adjust.

The uncomfortable reality is that AI is unlikely to distribute outcomes equally.

The Organizations That Win Will Treat Data as a Business Asset

One of the strongest takeaways from the discussion was that organizations can no longer afford to treat data purely as technical plumbing hidden beneath the business.

The companies most likely to succeed in the AI era may be those that:

• Understand the business demand for data
• Align data strategy with operational outcomes
• Elevate data accountability into executive leadership
• Build AI-ready governance structures
• Modernize fragmented systems
• Improve organizational speed
• Treat data as a strategic business asset rather than a backend IT concern

Dutta repeatedly emphasized that data leaders must stop positioning themselves purely as technology providers and instead become translators of business value. That means connecting data investments directly to operational outcomes, productivity improvements, customer impact, and competitive differentiation.

Because ultimately, the organizations that win the AI era may not simply be those deploying the most advanced models.

They may be the organizations most capable of delivering trusted, high-speed, accountable data into the systems driving modern business execution.

The Enterprise AI Accountability Gap

• AI has compressed the competitive speed dimension in business.
• Most enterprise AI failures are fundamentally data failures.
• Organizations can no longer rely on “fix it later” data operating models.
• Data quality, compliance, and speed now function together as a unified business requirement.
• AI lowers entry barriers for startups while increasing disruption risk for legacy enterprises.
• Organizations that fail to elevate data leadership risk long-term competitive decline.