DIGITAL LIFE

Almost half of projects using AI agents in companies are unlikely to survive until 2027, Gartner predicts

By 2027, around 40% of agent-based artificial intelligence projects will be cancelled, Gartner predicts, citing issues such as uncertainty about their real value to the business, the costs associated with the generalisation of the technology or poor risk management as the main reasons.

“Most AI projects with agents are currently early-stage experiments or proofs of concept driven by the hype around the topic and often poorly implemented”, highlights Anushree Verma, senior analyst at Gartner, quoted in a statement.

“This can blind organisations to the real costs and complexity of implementing AI agents at scale, preventing projects from going into production”, adds the same person, advising organisations to think strategically about the topic and how and where to apply the technology, without being influenced by trends.

Suppliers take advantage of market enthusiasm...Gartner even warns that suppliers are somewhat “blame” for the high expectations surrounding AI agents, which end up influencing less studied investments. It says that many companies have taken advantage of the moment to rebrand existing solutions, repositioning them as AI agents that they are not.

Among true AI agents, chatbots and robotic process automation (RPA) solutions, Gartner estimates that only 130 of the thousands of AI agent suppliers can claim to be such.

Even so, “most AI agent proposals lack significant value or return on investment (ROI), because current models are not mature enough to autonomously achieve complex business objectives or follow differentiated instructions over time”. 

Bad experiences do not erase great advantages...In January of this year, a survey by the consultancy firm among 3,412 participants in a webinar organized by the consultancy concluded that 19% of organizations had already made significant investments in AI agents, 42% classified the investments made as conservative and 8% had not yet invested. Meanwhile, 31% were awaiting developments or were still unsure whether they wanted to invest.

Even with the less than optimistic scenario, Gartner emphasizes that AI agents “represent a leap forward in AI capabilities and market opportunities”, which will help organizations improve resource efficiency, automate complex tasks and introduce business innovations.

By 2028, the consultancy firm estimates that at least 15% of day-to-day work decisions will already be made through agentic AI, compared to 0% in 2024. By the same time, 33% of business software applications will already include agent technologies. Last year this was only the case in 1% of applications.

mundophone