• As companies continue to spend sizeable sums on artificial intelligence, the question becomes whether they have the skills and infrastructure needed.
• Building a data infrastructure is one of the biggest challenges in supporting AI initiatives.
• When it comes to data infrastructure, many organizations take a Whac-A-Mole-type approach: tackling data on a project-by-project basis and solving issues as they come up.

As businesses continue to spend heavily on artificial intelligence, the question becomes whether they have the infrastructure and skills in place to get the most returns on these investments.

AI has become "the engine driving innovation across industries," said Paul Pallath, vice president of applied AI at technology consulting firm Searce. From boosting operational efficiency to delivering insights and discovering new opportunities, AI has the power to redefine how businesses operate. Yet, for many organizations, this potential remains out of reach, "because the road to AI adoption is strewn with challenges that often derail success," he said.

For AI to deliver real value, businesses need to get over these hurdles while building a solid foundation for sustainable AI integration, Pallath said.

One of the biggest challenges is building a data infrastructure to support AI initiatives. A survey of 500 U.S. senior business leaders conducted by consulting firm EY last year showed that 83% of respondents said their organization's AI adoption would be faster if they had stronger data infrastructure in place. Two-thirds admitted that a lack of infrastructure is holding back AI adoption at their companies.

"This highlights the crucial need for robust data hygiene practices to ensure the accuracy, consistency, and reliability of data used to train and operate generative AI models," said Dan Diasio, global AI consulting leader at EY.

When it comes to data infrastructure, Diasio said he's been seeing many organizations take a Whac-A-Mole-style approach, meaning they're tackling data on a project-by-project basis and solving issues as they come up.

However, to have a solid foundation for AI, companies need to build a cohesive enterprise strategy. "Although it can be difficult to stitch and connect everything together, it can have major benefits," Diasio said. "Once the data and knowledge foundation is in place, it becomes a flywheel. Having greater context benefits all initiatives, and each new initiative becomes easier to implement."

The lack of a solid data infrastructure can lead to poor data quality. "AI thrives on data, [but] most organizations are drowning in messy, siloed, and unreliable information." Pallath said. "Without clean, correct, and accessible data, even the most advanced AI models are destined to fail."

For a strong data foundation, companies need to implement robust data governance, Pallath said, including developing policies and standards to ensure data accuracy, consistency, and security across the organization. They should also break down silos by integrating disparate data sources into unified platforms, such as data lakehouse implementations leveraging data fabric architecture.

Investing in automation

Another good practice is to invest in automation, Pallath said. "Use tools to clean, deduplicate, and validate data continuously," he said. "AI is only as smart as the data it's trained on. Clean data isn't just a technical asset; it's the currency of trust in an AI-driven world."

IT leaders also continue to face the challenges of finding the AI-related skills they need and overcoming cultural barriers to AI adoption.

"AI may be cutting-edge, but it still needs people to build, manage, and guide it responsibly," Pallath said. "Equally important is a culture that supports innovation. Yet, many organizations struggle to find talent skilled in machine learning, data engineering, or AI ethics. Worse, cultural resistance to AI — driven by fear of automation and the potential displacement of jobs — further slows progress.

The EY research revealed an ongoing "AI fatigue," with half of the senior business leaders reporting declining company-wide enthusiasm for AI integration and adoption.

"While employees generally trust AI technologies, anxiety persists, particularly regarding job displacement and the rapid pace of adoption," Diasio said.

The EY research showed that business leaders have increased their focus on responsible AI over the last six months. The findings showed that organizations are working to build trust in order to maximize the positive impact of AI across the enterprise.

Companies need to upskill their workforce and foster AI literacy, Pallath said. "Invest in training programs to build AI literacy across teams, from technical skills to understanding ethical implications, and responsible use of AI," he said.

In addition, they should establish open communication channels.

"Talk about AI honestly because open dialog is critical to address fears, misconceptions, and resistance," Pallath said. "Foster transparency through channels that encourage information sharing, feedback, and ongoing collaboration. Talent isn't always just a skill gap, it's a mindset gap. Build a culture that embraces AI, not fears it."

It's important for leaders to take a step back and think holistically about AI adoption, rather than just thinking of it as a technology installation, Diasio said. "AI is powerful and disruptive, and it impacts all aspects of a business: people, processes, data, and technology. Leaders should make sure they're addressing all of these factors to fully harness its ability to power business transformation," he added.