Gartner: AI budgets are going to projects that demonstrate value, real-world impact

Many infrastructure and operations leaders aren’t able to dig out enough money from budgets to reallocate to AI projects, Gartner said in a survey released this week.

The research firm surveyed 253 IT leaders globally, and the budget issue plagued half the participants. As a result, 54% said they are focusing on AI projects with attainable results and foreseeable cost savings, Gartner said.

The projects are mostly in functions “where AI can quickly demonstrate measurable impact,” said Melanie Freeze, research director at Gartner. “We’re seeing the greatest AI momentum in IT service management and digital workplace functions, which are places where automation and generative AI can directly enhance productivity and reduce costs.”

Freeze cited a prominent healthcare company that used AI to improve productivity and reduce incidents by boosting the self-service capabilities of its IT service desk. “To do this, they created a portal where employees could type the issues they were facing, and a generative AI model would give them a solution based on the ITSM data and knowledge base within their organization,” Freeze said.

That helped reduce first-call resolution incidents and allowed service desk employees to move to more complex tasks, Freeze said.

However, AI integration is becoming difficult, as it is complex to put all the pieces together and make sure everything fits. The best way to address this is by starting small and winning big, she said —is a classic way of approaching IT.

“In other words, I&O (infrastructure and operations) really needs to prioritize AI use cases that can build that momentum with relatively easy returns,” Freeze said.

While cost optimization in AI could provide short-term benefits, the goal is to deliver long-term value and innovation by learning and building on these AI capabilities. “But they’re being held back by budget constraints and integration difficulties,” she said.

About 48% of the IT leaders surveyed reported integration difficulties.

The Gartner data follows controversial research from MIT; it found that 95% of AI projects failed.

That echoes Gartner research in June that said more than 40% of agentic AI projects would be canceled by the start of 2028. AI just can’t be in that exploratory “experimentation everywhere stage,” Freeze said. “They have to focus those efforts and target places where they can demonstrate the ROI and then build that momentum.”

Balancing the value and feasibility of AI projects comes down to determining whether the projects align with the company’s priorities. “The cost of integration might be the risks associated with it. It might be the risk of just the cost of the solution. Those things can branch out when it comes to value,” Freeze said.

It also comes down to using AI to provide insights and capabilities and enhancing the customer experience, Freeze said.

Many tech companies, including Microsoft, Google, Slack, Box, and Zoom, regularly talk about providing value with AI tools that unlock information and insights from existing and dormant datasets. That can help companies make better long-term decisions.

The survey did not look at long-term capital expenditures. It focused only on projects that directly affect worker productivity.

The participants surveyed expect to remain cautious about squeezing money from their budgets for AI projects. “AI has always been a challenge for infrastructure and operations leaders because they are tasked with keeping the lights on while still finding opportunities for innovation and helping enable business value,” Freeze said.Gartner: AI budgets are going to projects that demonstrate value, real-world impact – ComputerworldRead More