A Gartner survey released on October 19 found that only 13 per cent of chief information officers said their organisations had achieved significant or transformational value from AI. The findings, presented at the Gartner IT Symposium in Orlando, underline growing pressure on executives to justify spending as experimental projects move into everyday operations.
The survey covered 2,002 chief information officers and technology executives between May and July 2026. Gartner found that AI use cases had multiplied sharply, while their probability of success remained around one in five. The results suggest that expanding deployment alone does not guarantee commercial returns.
“Volume is not value,” said Alicia Mullery, a Gartner vice-president and analyst. She identified inadequate assessment of potential benefits and reluctance to discontinue unsuccessful projects as important weaknesses in corporate AI programmes.
Only 31 per cent of respondents had established criteria for preventing unsuccessful pilots from entering production, while 17 per cent had rules for withdrawing underperforming applications already operating. Organisations applying those respective controls were 2.3 and 2.6 times more likely to belong to Gartner’s higher-performing group.
Financial pressures are adding to the challenge. Although 58 per cent of technology leaders faced greater demands to reduce costs through AI, 51 per cent expected the technology to increase their overall cost of owning and operating technology systems. Computing requirements, model usage and integration expenses complicate forecasts that assume automation will produce immediate savings.
A separate McKinsey global survey published in August found that 37 per cent of respondents attributed some improvement in earnings before interest and tax to AI, broadly unchanged from the previous year. Only 6 per cent qualified as high performers, reporting both significant value and an earnings contribution of at least 5 per cent.
Yet the same survey, involving 1,719 participants across 97 countries, found that 80 per cent believed AI had improved their individual productivity. Half said it helped them make better decisions. The contrast indicates that faster completion of individual tasks does not necessarily translate into lower corporate expenditure or higher revenue.
Savings can remain theoretical when employees use AI to draft documents or analyse information without changes to staffing, service delivery or business processes. Organisations may also incur additional expenses checking automated outputs, managing sensitive data and connecting tools to existing systems. Productivity measures therefore require comparison with the full costs of implementation.
McKinsey found that nearly three-quarters of its AI high performers had fundamentally redesigned workflows, compared with approximately one-quarter of other respondents. Higher-performing organisations were also more likely to establish formal procedures for measuring results and secure active support from senior management.
Scaling remains another obstacle. Gartner reported in September that only 22 per cent of organisations surveyed had successfully expanded AI across multiple business units or adopted an AI-first approach. Nevertheless, 85 per cent of functional leaders planned to increase spending during 2026, despite limited evidence of organisation-wide success.
Cost visibility is particularly uneven. About 11 per cent of organisations in that Gartner survey did not know what their function had spent on AI during 2025. Without reliable expenditure records, companies face difficulties calculating returns or comparing competing projects.
KPMG’s third-quarter global AI survey likewise found that only 12 per cent of organisations consistently assessed AI value against its cost. Among organisations already reporting established returns, that proportion rose to 48 per cent, indicating a relationship between financial discipline and measurable outcomes rather than proof of causation.
The KPMG survey also found that 86 per cent of organisations with established returns operated a formal cross-functional or enterprise-wide AI management structure, compared with 31 per cent of those still experimenting. Such arrangements coordinate governance, security and operational accountability across departments.
McKinsey also found that 20 per cent of respondents had restricted AI usage because of operating costs, including charges linked to processing requests. Such constraints affect the economics of deployment even where employees consider the tools useful.
Workforce planning remains underdeveloped. Gartner found that just 4 per cent of technology executives had explicit plans to redesign jobs around cooperation between employees and AI agents, although organisations with such plans were substantially more likely to report stronger business outcomes.
Follow Arabian Post
Select Arabian Post as your preferred source on Google and MSN News for trusted business news and Arab politics and updates.
