The focus: why 40 % of the projects could fail by 2027!

The focus: why 40 % of the projects could fail by 2027!

artificial intelligence, and yet many companies face major challenges when using AI agents. From the perspective of Annette Zimmermann, VP Analyst at Gartner, it becomes clear that the term “AI” can vary widely, depending on who you speak to. Misunderstandings are not uncommon and often lead to misjudgments and excessive expectations. In conversation about agent -based AI, it becomes clear that these technologies are already tested in many companies, but the chances of success are anything but secured. According to All about security forecast gartner that more than 40 % of the AI projects are canceled or hired by 2027. The reasons for this are primarily high operating costs and a lack of economic benefit.

companies are faced with the question: How do you recognize serious solutions in the jungle of the providers? Unfortunately, there are significant differences between classic AI applications such as chatbots and the acting AI agents who can actually make decisions autonomously. Gartner estimates that just 130 providers provide real agent-based technologies, while many simply renaming existing technologies-from robot processes to AI assistants. This makes it difficult to separate the chaff from wheat.

The reality of the Agentic ai

Agentic AI, which enables autonomous decision -making, harbors enormous potential, but also has a high risk. According to an article by Forbes Agentic Ai can do without rethinking the required corporate structure. A survey from January 2025 shows that only 19 % of the over 3,400 experts surveyed have made significant investments in this area. While 42 % are conservative, a third of the survey -participants are undecided.

High costs for compliance, infrastructure, training and comprehensive workflow designs represent considerable hurdles. Integration into existing systems is often difficult by outdated technologies and forces companies to fundamentally revise existing workflows. But the effort could be worthwhile: Gartner predicts that 15 % of all routine business decisions by AI agents could be made autonomously by 2028.

The future of AI agents

The potential for agent AI in specific areas of application is clear, but the way there is rocky. In order to be able to survive in the competition, companies must define clear business results and a prudent use of these technologies is essential. A future-oriented and corporate-centered approach could increase the likelihood that AI agents are not only perceived as hype, but actually generate added value. The challenges remain high, but it is crucial to ask the right questions and make sound decisions in order to benefit from the advantages of AI agents.

In this dynamic environment, it will be essential for decision -makers to work closely with their teams. This is the only way to make the balancing act between human supervision and autonomous decisions. A single failure could have fatal consequences for the whole company. So it remains exciting how the landscape of AI agents will develop in the coming years and whether we are really ready for the profound changes that they bring with them.