A new survey from Genpact and HFS Research reveals a stark gap between corporate AI ambition and operational reality. The study, which polled over 2,000 enterprise executives across industries, found that 85% of leaders believe their underlying foundations—fragmented data, ungoverned processes, aging systems, and undertrained talent—are actively working against their AI investments. The report warns that layering AI on top of processes never designed for it does not unlock value but rather "locks in the cost of the status quo at machine speed."
Enterprises are currently directing an average of 13% of function spend toward AI. Yet the very systems supporting that spending remain unprepared, according to the findings. The research identifies four compounding "enterprise debts"—structural liabilities that do not appear on balance sheets but accumulate quietly in systems held together by tribal knowledge and workflows layered with workarounds. These debts, the report argues, threaten to negate the returns of even well-funded AI programs.
The survey underscores a widening chasm between executive confidence and actual readiness. While AI adoption is nearly universal among large enterprises, the infrastructure required to support it—clean data, modernized processes, and skilled talent—remains fragmented or absent. Genpact suggests that without addressing these debts, companies risk scaling inefficiency rather than innovation.
This finding carries significant implications for the broader enterprise software market. It suggests that a large portion of current AI spending may fail to deliver anticipated productivity gains. The report calls for a fundamental rethinking of enterprise architecture before further AI investment, positioning foundational modernization as a prerequisite rather than an afterthought.
Notably, the data is drawn from a self-reported executive survey, which may overstate both ambition and perceived challenges. The report was produced in partnership with Genpact, a vendor of business process services, which could introduce a bias toward emphasizing the need for external transformation support.