According to a study, 40% of projects with AI agents will be canceled before 2028

According to a prediction of the Gartner consultant, more than 40% of agricultural artificial intelligence projects will be canceled at the end of 2027. The main reasons are the Cost increasean unclear commercial value and the lack of adequate risk controls to manage this technology that has just arrived. Although large companies like Huawei are betting on AI -based agents, even for the connected house, this concept still has a long way ahead to generalize, according to this information.

Anushree Verma, analyst and senior director of Gartner, explains that most of the current initiatives are experiments or concept tests. These projects, often driven by high expectations (literally use the word “hype”), can hide organizations for organizations real cost and complexity which implies a large -scale implementation.

The “hype” for AI agents is causing an “image washing”

Gartner has also identified a practice that contributes to confusion, called “agent image washing.” It consists of what Many suppliers simply rename existing productssuch as chatbots or automation systems, without giving them true agricultural capacities. It is estimated that, of thousands of suppliers in the sector, only around 130 offer agricultural solutions that can be considered authentic.

Currently, most of these proposals lack a Investment return demonstrable, since the models do not have the necessary maturity to autonomous complex business objectives. In many cases, the problems trying to solve with agriculture do not actually require such advanced implementations.

Despite the challenges, the trend represents a market opportunity. The agricultural AI will allow optimizing resources and automating tasks of forms that exceed the current bots. The report also predicts that by 2028, the 15% of labor decisions Daily will be taken autonomously thanks to this technology and will be present in 33% of business software applications.

The recommendation for companies is to implement agricultural AI only when it offers a clear value (see the integration of goal AI in WhatsApp). Probably, the most sensible thing will be, instead of integrating agents into inherited systems, redesigning workflows from scratch.