AI Adoption Surges But Data Readiness Lags: The Real Challenge for Enterprises

Enterprises are diving headfirst into artificial intelligence, with nearly all organizations now running some form of AI initiative. Yet a recent survey by Dun & Bradstreet reveals a striking disconnect: while 97% of businesses have active AI projects, only 5% believe their data is truly ready to support those efforts. This gap between ambition and infrastructure is creating a bottleneck as companies try to move from isolated pilots to large-scale operational AI. Below, we explore the key findings from the report, the returns early adopters are seeing, and the data hurdles that remain.

• What percentage of enterprises are investing in AI, and how many have data ready?
• Are enterprises seeing returns on their AI investments?
• How many organizations plan to increase AI spending in the near future?
• What are the main data-related challenges hindering AI scaling?
• Why is data readiness more critical for production AI compared to pilots?
• How confident are enterprises in identifying and mitigating AI-related risks?
• What does the future hold for AI operationalization?

What percentage of enterprises are investing in AI, and how many have data ready?

According to the Dun & Bradstreet AI Momentum Survey, a staggering 97% of organizations report that they have active AI initiatives underway. However, only a tiny fraction—just 5%—say their data is ready to support these initiatives effectively. This striking disparity highlights that most companies are forging ahead with AI despite significant data preparedness gaps. As Cayetano Gea-Carrasco, Dun & Bradstreet's chief strategy officer, puts it: “You do not need enterprise-wide AI-ready data to launch pilots or isolated AI use cases. But you do need it to scale AI reliably across mission-critical workflows and systems.” In other words, while experimentation is booming, true enterprise-wide deployment remains elusive for the vast majority.

Are enterprises seeing returns on their AI investments?

Yes, many organizations are beginning to see tangible returns. The survey found that 67% of businesses report “early signs or pockets” of return on investment from AI. A further 24% say they are achieving “broad or strong” returns. This represents a marked improvement from just a year ago, when such results were far less common. However, returns remain uneven across sectors and use cases. The report notes that as adoption accelerates, early successes are becoming more frequent, but the path to consistent, scalable ROI is still challenging. The key takeaway: initial gains are real, but they are not yet universal or stable enough to satisfy all stakeholders.

How many organizations plan to increase AI spending in the near future?

More than half—56%—of the 10,000 businesses polled by Dun & Bradstreet say they plan to increase their AI investment over the next 12 months. This indicates strong confidence in the technology’s potential, despite the data readiness hurdles. Additionally, 30% of organizations are currently scaling AI into production environments, and 26% are operationalizing the technology across multiple core processes. These numbers suggest that while many companies are still in early stages, a significant minority are already moving beyond pilots. The appetite for AI is clearly growing, and budgets are following suit.

What are the main data-related challenges hindering AI scaling?

Enterprises face a cluster of data-related obstacles that impede AI at scale. The survey identifies several key issues: access to data is a problem for 50% of respondents; privacy and compliance risks trouble 44%; data quality and integrity concerns affect 40%; lack of integration across systems is reported by 38%; and shortage of qualified AI professionals is cited by 37%. These challenges compound each other. For example, even if an organization has good data, it may be locked in silos (integration) or lack the talent to use it effectively. The result is a landscape where most firms struggle to move from controlled experiments to reliable production AI.

Why is data readiness more critical for production AI compared to pilots?

As Gea-Carrasco explains, it is relatively easy to launch a pilot or a departmental AI tool using general-purpose models and get impressive results in a controlled setting. But deploying AI into production workflows—such as onboarding, compliance, risk management, or customer operations—demands accuracy, accountability, explainability, interoperability, and consistency. These attributes directly impact business decisions. In production, a small error can cascade into significant financial or reputational damage. Clean, governed, and interoperable data becomes non-negotiable. Without it, AI cannot be trusted to operate autonomously or integrate seamlessly with existing systems. That is why only 5% of enterprises feel their data is ready for such wide-scale deployment.

How confident are enterprises in identifying and mitigating AI-related risks?

Confidence levels remain alarmingly low. Only 10% of enterprises say they are highly confident in their ability to identify and mitigate AI-related risks. This is a concerning finding, especially as AI moves into more critical, high-stakes areas. Without robust risk management frameworks, companies risk exposing themselves to regulatory penalties, biased outcomes, security vulnerabilities, and loss of customer trust. The survey suggests that most organizations are still in the early stages of building these capabilities. The lack of confidence pairs directly with the data readiness problem—poor data governance often leads to uncontrolled risks. Enterprises must invest in both data infrastructure and risk governance to close this gap.

What does the future hold for AI operationalization?

The Dun & Bradstreet survey indicates that the momentum behind AI is strong and likely to accelerate. With 56% planning to increase investment and a growing number moving into production, the direction is clear. However, the data readiness challenge is “even more profound” in 2026 than in the previous year, according to the report. This means that while more companies are advancing, the underlying foundation is not keeping pace. The key question, as Gea-Carrasco notes, is no longer whether organizations are experimenting with AI, but whether they have the data and infrastructure to deploy it reliably at enterprise scale. The near future will likely see a greater emphasis on data governance, integration, and talent development as prerequisites for successful AI at scale.