40%+ of Agentic AI To Be Canceled by 2027, Study Finds

AI research and consultancy firm Gartner recently published a study predicting that over 40% of agentic AI projects will be scrapped by 2027. The June 25 study cites rising costs, unclear business value, and “inadequate risk controls” as the top reasons behind the expected cancellations.

Projects to develop or partner with agentic AI (also referred to as AI personal assistants) are becoming increasingly popular among top technology companies and are seen as the next big thing in AI development. Agentic AI is designed to leverage the power of popular large language models (LLMs) and, given access to critical information, perform crucial tasks for a person or company.

Although commercially viable agentic AI remains speculative, major companies have already begun investing in it. In May, both Visa and Mastercard announced collaborations with some of the largest companies in the AI space to create “virtual shopping assistants” that would have access to key banking information.

However, Gartner researchers predict a lowering of enthusiasm in the field. “Most agentic AI projects right now are early-stage experiments or proof of concepts that are mostly driven by hype and are often misapplied,” said Gartner’s Senior Director Analyst, Anushree Verma.

“Most agentic AI propositions lack significant value or return on investment (ROI), as current models don’t have the maturity and agency to autonomously achieve complex business goals or follow nuanced instructions over time,” said Verma. “Many use cases positioned as agentic today don’t require agentic implementations.”

On top of the complications associated with developing functional AI agents, Gartner also highlights that most current and prospective products being branded as “AI agents” don’t actually have any “agentic” capabilities, and are instead simply chatbots, automation processes, or prompted LLMs.

In one case, an app that promised universal, AI-powered shopping assistants was sued after an investigation revealed the so-called “AI” was really human workers in the Philippines and Romania.

That said, Gartner still predicts that at least 15% of day-to-day work operations will go through AI agents by 2028.