Insights

Artificial intelligence has become the defining business priority of this decade. Corporate investment is surging, boardroom conversations are consumed by it, and the competitive pressure to act is real across every major industry — from financial services to consumer goods to manufacturing. For large, established organizations, the strategic rationale is clear: AI promises faster decisions, leaner operations, and new revenue models that simply did not exist five years ago. The urgency is not manufactured. Companies that move too slowly risk ceding ground to leaner, faster competitors who were built with AI at their core.

Yet the gap between ambition and execution is wide — and widening. Most large corporations are not failing because the technology does not work. They are failing because they underestimate what it takes to absorb it. From my work advising organizations across North America, four challenges come up consistently:

CHALLENGE 1  ·  LEGACY INFRASTRUCTURE

Decades of siloed data, incompatible systems, and fragmented technology stacks make it extraordinarily difficult to deploy AI at scale. You cannot build a modern intelligence layer on a foundation that was never designed to share information across functions. Most enterprises discover this only after a costly pilot that works in isolation and dies on contact with the real operating environment.

CHALLENGE 2  ·  TALENT AND SKILLS GAPS

There is a meaningful shortage of people who can bridge business strategy and AI execution — not just data scientists, but leaders who understand both the technology and the organizational change required to deploy it effectively. Training programs exist, but most large organizations have not committed the time, budget, or protected space for employees to actually develop new capabilities.

CHALLENGE 3  ·  CULTURAL RESISTANCE

Fear of job displacement is real, and it suppresses adoption at every level. Frontline employees default to familiar routines. Middle managers who feel threatened by AI quietly undermine initiatives. And without senior leaders visibly using and championing these tools in their own work, the cultural signal never reaches the rest of the organization.

CHALLENGE 4  ·  GOVERNANCE AND OWNERSHIP GAPS

AI initiatives that live between the IT department and the strategy team tend to die in committee. Without clear accountability — who owns the outcome, who has budget authority, who can make the call to scale or kill a pilot — even well-funded programs stall. This is particularly acute in regulated industries like insurance, banking, and healthcare, where compliance layers add friction at every decision point.

"To truly implement AI in a legacy industry like insurance, you have to set everything aside and build the company from scratch. The existing structure simply will not allow it."

— Senior executive, Salles Consulting client (insurance sector)

It is a view I hear more often than you might expect — and one that reflects genuine, hard-won frustration. Legacy constraints are real. But I do not fully agree with the conclusion. The organizations getting this right are not the ones that burned the existing structure down. They are the ones that chose a small number of high-value use cases, built visible wins, and used those wins to earn the organizational trust needed to go deeper. As BCG's 2025 Build for the Future study noted, only 5% of companies are generating substantial financial value from AI — and what separates them is not a blank slate, but disciplined execution and sustained leadership commitment. The path forward is harder than a fresh start sounds — and also more achievable.

Source referenced: BCG Build for the Future 2025 Global Study (n = 1,250 C-suite executives).