Artificial intelligence (AI) agents may be inevitable in the business world, but they are far from a sure thing. More than 40% of agentic AI projects will be deep-sixed by the end of 2027 because of escalating costs, unclear business value, or inadequate risk controls, according to market researcher Gartner.
The January poll of 3,412 webinar attendees found just 19% said their organization had made significant investments in agentic AI, 42% had made conservative investments, and 8% no investments at all. The remaining 31% are taking a wait-and-see approach or are unsure.
Indeed, many companies are engaging in “agent washing,” the rebranding of AI assistants and chatbots without significant agentic capabilities, Gartner found. It estimates only 130 of thousands of agentic AI vendors are real.
“Most agentic AI projects right now are early stage experiments or proofs of concept that are mostly driven by hype and are often misapplied,” Anushree Verma, a senior analyst at Gartner, said in a statement. “Most agentic AI propositions lack significant value or return on investment (ROI), as current models do not have the maturity and agency to autonomously achieve complex business goals or follow nuanced instructions over time.”
Gartner’s poll reflects the early-early stages of AI agents, and where they stand within corporate environments that are assessing the benefits and shortcomings of autonomous work assistants. ROI, in particular, has become a mantra of chief information officers tasked with putting in place AI agents to augment their workers while boosting productivity and efficiency.
Plans to scrap AI agent projects by 2027 isn’t entirely unexpected. Many analysts have speculated that nascent projects of the emerging technology are likely to morph into different projects or outright removal based on their results after the first year or 18 months. It remains to be seen how agents will be integrated into current operations, where they will be most effective, and whether the investment will pan out.
“In the midst of intense competition and hype, some AI agent companies have begun tackling problems not because these tools are the most suitable or effective solutions, but simply because they represent the ‘coolest’ technology,” AI innovation and investment strategist Songyee Yoon said in an email. “This approach is fundamentally flawed. When solutions are driven by technological novelty rather than by real-world utility or efficiency, they inevitably fall short. True innovation lies not in chasing trends, but in delivering meaningful outcomes through the most appropriate tools.”
Unpredictability over the initial wave of AI agents and their success should come as ominous news to the scores of tech giants — including Salesforce Inc., ServiceNow Inc., IBM Corp., Alphabet Inc.’s Google, and Microsoft Corp. — that have poured billions of dollars into creating AI agents, systems that can autonomously complete goals and take actions as a digital workforce with minimal human interaction.
While 53% of employees with access use AI daily, a whopping 74% said though their company treats data like gold most aren’t managing it well enough for AI to do anything useful. But they do point to a need for agentic AI: 41% of workers want AI to handle “soul-sucking tasks like manual data entry,” the study said.
