{"id":23688,"date":"2026-05-05T14:03:16","date_gmt":"2026-05-05T19:03:16","guid":{"rendered":"https:\/\/tomilli.com\/en\/?p=23688"},"modified":"2026-05-05T14:03:17","modified_gmt":"2026-05-05T19:03:17","slug":"enterprise-ai-agents-2026-management","status":"publish","type":"post","link":"https:\/\/tomilli.com\/en\/enterprise-ai-agents-2026-management\/","title":{"rendered":"2026, a Year for “Managing” Enterprise AI Agents"},"content":{"rendered":"\n

Mexico City, April 28, 2026<\/strong> \u2015 This year is positioned to be a watershed moment for any corporate strategy. More than a trend, artificial intelligence agents – systems capable of executing tasks, making operational decisions, and automating entire processes without constant intervention \u2013 are changing the way we work. But, according to Elisa Garc\u00eda Barrag\u00e1n, CEO at Netsoft, there is a question behind the enthusiasm that is starting to bother organizations: Who is managing these agents? And what processes should they actually be operating?<\/strong><\/p>\n\n\n\n

Data shows that adoption is moving fast, but maturity is not. According to Gartner<\/a>, by 2026, 40% of enterprise applications will incorporate task-specific AI agents, up from less than 5% in 2025. However, the same firm warns that operational reality is still far from that percentage: only 15% of organizations are piloting or deploying fully autonomous agents.<\/p>\n\n\n\n

The gap is evident: the market is moving faster than the capacity of companies to adapt to it. According to McKinsey, the true impact of artificial intelligence on organizations does not depend on the tool itself, but on its ability to redesign entire processes and workflows. Nevertheless, only 21% of companies have made structural changes in the way they operate their business, despite this factor correlating with improvements in financial results.<\/p>\n\n\n\n

\n

\u201cThe challenge is not in putting agents to work, it is in preventing them from working without controls. If there are no clear processes and a reliable data source, the only thing that is automated is disorder,\u201d says Elisa Garc\u00eda Barrag\u00e1n.<\/p>\n<\/blockquote>\n\n\n\n

In this context, Garc\u00eda Barrag\u00e1n points out that artificial intelligence is not solving inefficiencies: it is highlighting existing issues. In fact, she warns of a new emerging risk: the uncontrolled proliferation of agents that replicate distorted functions, operate on inconsistent data, and generate more complexity rather than efficiency.\u00a0<\/p>\n\n\n\n

\n

\u201cWhen agents begin to multiply without a clear architecture, they stop being an advantage and become a new layer of complexity. It is the same problem of fragmented SaaS that AI seeks to solve, but now it\u2019s automated,\u201d adds the subject matter specialist. “It’s a natural evolution of a problem that already existed with SaaS: too many tools, now replaced by too many agents.\u201d<\/p>\n<\/blockquote>\n\n\n\n

The paradox is that the economic potential is enormous. Capgemini<\/a> estimates that AI Agents could generate up to USD $450 billion in economic value, and that 93% of executives believe that their adoption will give their company a competitive advantage in the short term.<\/p>\n\n\n\n

But this expectation coexists with a more conservative reality: 85% of the business processes continue to operate with low levels of autonomy. The message is clear: The challenge is not to implement agents, but to integrate them.<\/p>\n\n\n\n

From Elisa Garc\u00eda Barrag\u00e1n\u2019s technical point of view, this involves something deeper than adopting new tools. It involves defining:<\/p>\n\n\n\n