Over 40% of Agentic AI Projects Expected to Be Abandoned by 2027, Says Gartner
According to Gartner, Inc., more than 40% of agentic AI projects will be discontinued by the end of 2027 due to escalating costs, unclear business value, or inadequate risk management.

Agentic AI refers to advanced artificial intelligence systems capable of perceiving their environment, setting and reasoning about goals, making decisions, and taking autonomous actions—without ongoing human intervention. Unlike traditional AI, which typically operates based on fixed rules or predefined tasks, agentic AI dynamically adapts to changing conditions, learns from interactions, and pursues complex, multi-step objectives through contextual understanding.
“Most current agentic AI initiatives are in their early stages—often limited to experimental or proof-of-concept efforts driven more by hype than by strategic alignment,” said Anushree Verma, Senior Director Analyst at Gartner. “This can obscure the true costs and complexities involved in scaling such systems, ultimately hindering their transition to full production. Organizations must move beyond the hype and make deliberate, well-informed decisions about where and how to implement this emerging technology.”
A Gartner webinar poll conducted in January 2025, which surveyed 3,412 attendees, found that:

• 19% of respondents reported significant investments in agentic AI,
• 42% had made conservative investments,
• 8% had made no investments, and
• 31% were either undecided or adopting a wait-and-see approach.

Adding to the confusion, many vendors are engaging in "agent washing"—rebranding traditional AI tools such as chatbots, robotic process automation (RPA), and virtual assistants as agentic AI, despite lacking true agentic capabilities. Gartner estimates that of the thousands of vendors claiming to offer agentic AI solutions, only around 130 are delivering legitimate offerings.

Current Maturity and ROI Remain Limited
“Most agentic AI offerings today lack the maturity and capability to deliver meaningful return on investment (ROI),” Verma noted. “Many so-called agentic applications don’t actually require autonomous, agentic functionality. The models often struggle to meet complex, long-term business objectives or interpret nuanced instructions.”
Despite these limitations, agentic AI represents a significant step forward in AI evolution, offering potential to enhance resource efficiency, automate sophisticated tasks, and create entirely new business models—far beyond the capabilities of scripted automation tools.

The Path Forward for Agentic AI

Gartner predicts that by 2028:

• At least 15% of routine work decisions will be made autonomously via agentic AI—up from virtually none in 2024.
• 33% of enterprise software applications will integrate agentic AI—compared to less than 1% in 2024.

However, Gartner advises that organizations should only pursue agentic AI where there is clear value or measurable ROI. Integrating such systems into existing infrastructure can be complex and disruptive, often requiring significant re-engineering. In many cases, it is more effective to redesign workflows specifically with agentic AI in mind.

“To unlock real value from agentic AI, the focus should shift from augmenting individual tasks to enhancing enterprise-wide productivity,” Verma emphasized. “Use agents where decision-making is needed, automation for routine workflows, and assistants for basic information retrieval. Ultimately, it’s about achieving business value through improved cost efficiency, quality, speed, and scalability.”