Despite the global hype surrounding artificial intelligence, the number of large enterprises or government processes that have been truly reinvented with AI at their core is "basically close to zero." This assessment comes from Marc Warner, Accenture’s global chief technology officer, speaking at a Fortune Brainstorm AI virtual event.

From Periphery to Core

According to Warner and executives from AXA, Mastercard, and Mistral AI, most companies are currently deploying AI in peripheral functions such as marketing, meeting transcription, and note-taking. Warner compared this moment to early factories that used electricity only for lighting, before eventually redesigning production around electric motors.

Greg Ulrich of Mastercard emphasized that the most significant returns come from redesigning entire workflows rather than layering AI onto existing ones. He warned of a common trap: accelerating one step in an 80-step process only to create a backlog at the next stage.

"Politics Projects" Over Technology

Warner described AI transformations as "politics projects," noting that the primary challenge is organizational change rather than software. Matthieu Caillat of AXA echoed this sentiment, stating that as technology improves, the work becomes less about the tools and more about "how do you bring people along."

When asked about accountability for AI failures, Warner suggested a tripartite division: the business owner is responsible for value, the technologist for functionality, and the CISO for security. However, he urged organizations to avoid a "blame game" for now, encouraging experimentation even when it leads to failure.

Short Shelf Life and Future Targets

  • Mastercard reported that consulting tools built just a year ago are already "quaint" and require rebuilding.
  • AXA is targeting between €500 million and €700 million in recurring annual pre-tax value from AI by 2029.
  • Mistral AI focuses on providing open-weight models to allow customers to maintain control and set their own guardrails.

Warner anticipates that the current pace of AI development will continue until approximately 2030, at which point energy constraints in the West may begin to impact growth.