13th November,2025:-Businesses worldwide have invested billions of dollars into deploying generative AI. Recent headlines have raised doubts about whether organizations are getting their money’s worth.

Despite the skepticism, the trend is not retreat but refinement: fewer firms expect to be “just starting” or stuck in pilots in 2026, and more expect to use AI regularly or integrating it across operations, according to a recent survey of global technology leaders conducted by IEEE.

The report, “The Impact of Technology in 2026 and Beyond: an IEEE Global Study,” found that technology leaders expect more intensive generative AI deployment at their firms.

Respondents were asked to select one of six stages that best described their organization’s adoption of generative AI as of 2026. In the prior survey, the largest number of respondents described their organization’s adoption level as having “high expectations or trying small projects.” In this year’s survey, the largest number of respondents (39%) reported that their organization would be “using regularly, but selectively,” followed by “rapidly integrating, expecting bottom line results” (35%).

• Just Starting: 1% (down 4% from 2025).
• High expectations and Trying Small Projects: 8% (down 25 points from 33%).
• Challenged and Rethinking the Approach: 4% (down 14 points from 18%).
• Learning and Seeing Some Benefits: 13% (down 11 points from 24%).
• Using Regularly, but Selectively: 39% (up 19 points from 20%).
• Rapidly Integrating, Expecting Bottom-Line Results: 35% (new in 2026; no 2025 baseline).

A Period of Skepticism? 

New technology frequently goes through what observers refer to as “the hype cycle.” The concept is useful for investors who need to decide where to put their money. Information technology managers use it to plan organizational tech roadmaps. It describes how new tech rockets from buzz to a peak of inflated expectations, falls into a trough of disillusionment, then climbs toward practical, mainstream value. The internet followed that path: the late-1990s dot-com boom peaked, crashed around 2000, and then matured into today’s everyday platform for commerce, media and work.

Generative AI has certainly been the focus of sky-high expectations. But nearly one-in-three projects will likely be scrapped. The consulting firm McKinsey reports that nearly eight in 10 companies use generative AI, and nearly the same proportion reports no measurable impact to the bottom line.

That’s led to something of a skeptical stance at many firms as companies plan their implementations.

“We’re entering a period of healthy skepticism that follows the natural progression of technology adoption cycles,” said IEEE Senior Member Santhosh Sivasubraman. “This isn’t necessarily negative; it represents a maturation of understanding about AI capabilities and limitations. Organizations that rushed into AI deployments without proper planning are now reassessing their approaches, leading to more thoughtful implementation strategies.”

Why Do So Many AI Projects Fail? 

The survey identified a key data point that may indicate where companies are struggling.  Projects often fail because teams assume the models are more reliable than they are. The confidence with which chatbots deliver results often leads to an overestimation of their capabilities, and 50% of respondents flagged over-reliance on AI and potential inaccuracies as a top concern.

“Inflated expectations have led leaders to imagine digital workers instead of carefully-designed systems for well-defined outcomes,” said IEEE Senior Member Eleanor Watson.

She notes that companies frequently apply advanced models to contexts where simpler analytics would have sufficed. At the same time, the quality of the data they use to feed the AI is often poor.

Another misalignment comes from confusing emotional aspects of generative AI with its functional benefits, according to IEEE Member Ning Hu. Too often, organizations invest in generative AI chatbots or virtual assistants that sound cheerful and supportive, but add little measurable value. Instead of the efficient manager that quietly handles tasks and keeps operations on track, companies end up with a companion that tries to ingratiate itself with a friendly tone and empathetic phrasing. Those don’t deliver real results.

Generating Real Results 

Generative AI’s most reliable returns are emerging in the background, where AI runs the pipes of the business, Hu said. Examples include automated order‑to‑cash workflows, inventory replenishment driven by predictive analytics, robotic process automation (RPA) for finance and HR and precision control in manufacturing or medical imaging.

The survey found that cybersecurity, supply chain and warehouse automation, along with software development are expected to be the top uses for AI in 2026.

Another resilient category is sandboxed experimentation: organizations create isolated environments where generative models can explore creative solutions or optimize designs, then transfer the validated outputs into production, Hu said.

“In all these scenarios, the AI’s value is measured in concrete, business‑critical metrics rather than in user sentiment alone,” Hu said.