# Most IT Leaders See AI Results, But Impact Remains Underwhelming
Two-thirds of IT leaders report measurable AI results from their deployments, yet the gap between perceived progress and genuine business impact reveals a sobering reality. When tech entrepreneur Azeem Azhar posed a follow-up question to 160 IT vice presidents at a Las Vegas conference, asking who achieved results significant enough to interrupt a CEO's vacation, only eight hands went up. That stark drop-off exposes the central tension in enterprise AI adoption today: deployment proliferation is outpacing genuine value creation.
The data reflects a wider pattern. According to analysis by Ramp, a financial management platform, companies have dramatically increased spending on AI applications. Ramp examined spending data from tens of thousands of companies and billions of transactions, documenting accelerating AI budgets across sectors. Yet widespread spending has not translated to transformative business outcomes that warrant executive escalation or strategic pivots.
This mismatch matters because it questions whether companies are getting adequate return on their AI investments. IT leaders deploying AI systems face mounting pressure to justify these expenditures to boards and finance teams. Measurable results sound promising in reports and presentations. Executives want concrete evidence that AI drives revenue, cuts costs, or creates competitive advantage. When only 12 percent of IT leaders possess results compelling enough to pull a CEO away from downtime, it signals that most AI initiatives remain confined to incremental improvements or pilot programs rather than delivering game-changing capabilities.
Several factors explain the gap. Many companies implemented AI without clear use cases or measurable success metrics. Teams deployed tools because competitors were doing so, not because internal analysis identified where AI would solve specific business problems. Integration challenges persist. Legacy systems resist modern AI infrastructure. Data quality remains poor at many organizations, limiting what machine learning models can accomplish. Talent shortages mean fewer people capable of moving projects from proof-of-concept to production at scale.
There's also the maturation curve issue. Early-stage AI implementations typically show promise in controlled environments but struggle during real-world rollout. Model drift, changing data distributions, and unexpected edge cases reduce performance once systems leave the lab. Organizations that moved quickly into production are now managing underperforming deployments that require rework.
The gap between "measurable results" and "CEO-vacation-interrupting results" deserves attention from procurement teams and CFOs. If two-thirds of IT leaders have results but only 12 percent deem them exceptional, the question shifts from "Is AI working?" to "Why isn't AI working better given the investment?" That diagnostic matters for 2025 budget planning.
Companies achieving genuine impact tend to share characteristics. They started with specific problems, not generic AI ambitions. They invested in data infrastructure before deploying models. They set realistic timelines and measured outcomes against defined baselines. They treated AI as a multi-year program requiring sustained engineering effort, not a quick fix.
The vacation test is blunt but useful. It separates genuine competitive advantage from organizational box-checking. As boards increasingly scrutinize AI spending, the 8 out of 160 statistic will likely intensify pressure on underperforming initiatives to either demonstrate value fast or get cut. That correction, though painful, could redirect resources toward projects where AI actually solves material business problems rather than simply modernizing operations for its own sake.