The Corporate Illusion of AI Transformation

Despite the relentless hype surrounding generative artificial intelligence, few major organizations have actually redesigned their fundamental operations to center on the technology. During a recent Fortune Brainstorm AI event, four top executives argued that while companies are adopting AI at scale, they are mostly using it for peripheral tasks like drafting emails or summarizing meetings rather than reimagining how their businesses function. Marc Warner, the global chief technology officer at Accenture, noted that the number of enterprises that have truly reinvented their core processes is essentially close to zero.

The panel suggested that the primary barrier to success is not the software itself, but internal corporate culture. Warner described these efforts as politics projects rather than technical ones, noting that shifting the economic engine of an industry requires taking risks with core workflows that most leaders avoid. Greg Ulrich of Mastercard echoed this sentiment, explaining that simply speeding up one part of a lengthy manual process often creates bottlenecks elsewhere. True gains require remapping entire lifecycles and shrinking team structures, a feat that demands intense executive sponsorship and a willingness to disrupt established hierarchies.

This transition period has created a precarious environment where accountability remains murky and tools become obsolete almost instantly. When discussing what happens when an AI agent fails, Warner urged companies to abandon the blame game and instead encourage experimentation, provided there isn’t a catastrophic security breach. This need for agility is underscored by the sheer speed of innovation; Ulrich admitted that some of his firm’s custom AI tools from just a year ago already feel quaint and useless, forcing constant cycles of rebuilding.

As firms look toward the end of the decade, the strategy is shifting from broad enthusiasm to targeted value creation. For instance, AXA has moved away from hundreds of small test cases to focus on high-value goals involving billions of unstructured documents. While energy constraints may eventually slow progress in the early 2030s, these leaders agree that for now, the real challenge lies in bringing people along for the ride and overcoming the institutional inertia that keeps AI on the sidelines.