The AI industry is moving quickly toward autonomous agents that can make decisions, use software and execute tasks with limited human intervention. But Gartner is warning that more than 40% of agentic AI projects could be canceled by the end of 2027.
The problem may not be the technology itself. Gartner points to rising costs, unclear business value and inadequate risk controls as major reasons projects fail to move beyond experimentation and into production. Many initiatives are still proof-of-concepts driven more by the excitement around AI than by a clearly defined business problem.
There is also a growing problem of “agent washing”—companies labeling existing assistants, chatbots or automation tools as agentic AI without providing meaningful autonomous capabilities. Gartner estimates that only around 130 of thousands of vendors marketing agentic AI actually offer genuine agentic capabilities.
Yet Gartner does not expect the technology to disappear. Quite the opposite. By 2028, the firm expects 33% of enterprise software applications to include agentic AI, while at least 15% of day-to-day work decisions could be made autonomously by AI agents.
The message for businesses is simple: AI agents may become one of the biggest transformations in enterprise software—but deploying an agent is not the same as creating business value.
The companies that win will likely be those that know where autonomy actually makes economic sense.
