A CIO's Guide To Closing The Gap Between AI Headlines And Operational Reality

Headlines about enterprise AI tell a tale of two worlds: the best of times, and the worst of times. Tech leaders promise sweeping automation on an impossibly short timeline, a narrative fueled by high-profile examples of AI replacing large volumes of labor. But the more common story is one of corporations running endless experiments that never seem to leave the lab, a struggle validated by recent industry surveys. Behind the layoff headlines, the reality is often simpler: companies are just shifting money from payroll to processing power.

Ariel Spak is a global executive who has served as both CFO and CIO across multinational corporations, including senior roles at Microsoft and Toyota. Most recently, he served as Vice President and CIO at Aramark, where he led digital transformation and AI strategy across 13 countries. He believes the gap between AI's promise and what organizations are actually delivering is one of the defining leadership challenges of 2026.

"This moment in AI is defined by two completely different views of the world. On one hand, tech leaders are saying every white-collar job will be automated in eighteen months," said Spak. "On the other, you have companies that have been trying with pilots but were never able to get anything concrete." Spak described the layoff headlines driving this perception as largely a budgetary shift, with companies moving money from labor to capital expenditure to build AI infrastructure, rather than evidence that automation had actually arrived.

Spak called this disconnect what it often is: a "marketing play." He noted that even the most technologically advanced organizations had fallen behind in rewiring their core operations, despite headlines touting AI supremacy. This gap between AI's potential and what companies have actually built was, in his view, the defining challenge for executives in 2026.

• Even giants are behind: "I have been talking with a couple of CFOs from Big Tech in the last few weeks, and also from major payment network companies. If you really ask around, they haven't implemented much AI into their processes, so a lot of companies are really, really behind," said Spak. Core workflows like forecasting, budgeting, HR processes, and first-line customer interactions were among the areas he identified as most lagging.

• The automation hype: "I don't believe the claim that every white-collar job will be automated in eighteen months is true. Even if the AI technology evolves, the culture change, the training, and capturing the data is a massive amount of work," said Spak. "Just that would take eighteen months." The observation pointed to a deeper issue: the bottleneck in enterprise AI is never the technology itself, but the organizational change required to deploy it at scale.

To cut through this noise, Spak proposed a roadmap that begins with a decisive strategic shift. It requires moving beyond the safety of isolated experiments and committing to transforming core business processes, a move that mirrors a larger trend in the industry as consulting giants partner with OpenAI and system integrators like Capgemini work to scale AI into operations.

• Pilot purgatory: "If you want to be successful, you need to make a bet. You cannot continue with small pilots. The company needs to think about its core processes, identify the two or three most important ones and completely redesign them from the ground up," said Spak. "It's not about running another pilot. It's about making a real bet." For a manufacturing company, that meant supply chain and production, he noted. For a services company, it meant the white-collar workflows that drove the business.

• The literacy gap: Spak said this transformation is fundamentally a leadership and governance challenge, and that treating AI as a purely technical issue to be delegated is a common pitfall. In his view, success often depends on C-suite leaders personally investing in understanding AI's capabilities and limitations far beyond consumer-grade tools. "Any CEO, any CFO, or any HR lead today needs to understand the fundamentals of AI: the different types, what is real versus what is hype. This goes far beyond using tools like ChatGPT or Copilot," said Spak. "They don't need to be an expert, but they need to understand these concepts much better than they do today to lead a successful transformation." He pointed to a common misconception: that AI was simply a tool for answering questions or summarizing emails, when its real value lay in autonomous agents capable of taking action on behalf of the organization.

Once leaders make that bet, he said the focus should then turn to cultural implementation. He advocated for a third, more balanced approach rooted in empowerment and trust, noting that companies often take one of two common approaches: ignoring AI entirely, or trying to force adoption through invasive oversight.

• Results over rules: "Manage by objectives. As a manager, I don't care if people are at their desk nine hours per day or at the beach, as long as they accomplish their goals," said Spak. "It's the same with AI. I give you the tool. If that tool helps you be more productive and reach the same goals in less time, great for you. If not, and you still achieve the goals, that's fine." He contrasted this with companies that tied AI adoption to performance reviews and tracked employee tool usage, an approach he viewed as counterproductive and invasive.

• Permission to play: Spak was direct about the data risk concern that leads many organizations to restrict employee AI access. "I'm not in the camp that says we should ban people from using these tools due to risk. The concern that putting company information into an enterprise tool like ChatGPT would expose company data? I don't think that's true," he said. He noted that organizations with enterprise contracts with major AI suppliers had data protection built in, and that some guardrails were reasonable, but blanket bans were not.

Spak described two separate playbooks for navigating the speed-versus-safety question: a "fast" track of individual productivity tools to build cultural readiness, and a "slow" track of core process redesign requiring rigorous governance, an understanding of changing AI regulations, and mature operational frameworks. "You need to separate AI into two pieces. One is the AI that any employee can use every day, which has been shown to increase productivity maybe 5% or 7%, but it's not the real gain here. I would do that for the culture change and for people to understand AI," said Spak. "The true benefits come when the company as a whole completely changes a process leveraging AI. That's a project that will probably take twelve months to land."

Looking ahead, Spak said AI transformation should be a recurring board-level agenda item. He suggested the motivation isn't about avoiding a sci-fi future of job automation, but the practical risk of competitive disadvantage, already compounding for companies still caught between the two worlds. "Even from my own experience applying to CFO roles, it's clear that hiring managers are now seeking leaders with a background in technology. They understand that even for a CFO, you need someone who knows how to take advantage of AI in the next few years," Spak concluded.