The bottom line: 2025 acted as a ruthless filter for companies considering AI adoption. As predicted by Gartner, around a third of generative AI projects were abandoned because they remained at the stage of mere purchase intent without ever reaching actual adoption by teams.
At the start of 2026, the difference between leaders and followers is no longer based on access to technology (which has become a commodity), but on mastery of "human-machine collaboration." The organizations that are performing well today are those that understood as early as 2024 that AI is not just a simple IT tool, but involves a profound cultural transformation.
AI review in 2025: the reasons why technology projects disconnected from people fail
Last year marked the end of the "gadget" era. We have moved from the euphoria of discovery to the ruthless rigor of return on investment (ROI).
Gartner's prophecy has come true
At the end of 2024, Gartner announced that 30% of generative AI projects would be abandoned after the testing phase. The 2025 review confirms this trend: companies that settled for a strategic intention (deployment of massive "top-down" tools) without supporting the change saw their licenses go unused and their budgets skyrocket. Without real adoption, the project remains an expense; with human involvement, it becomes an investment.
McKinsey's finding: the performance gap is widening
The latest McKinsey report (early 2026) shows that while 80% of companies have now integrated AI into their processes, only 15% of them are deriving a real competitive advantage from it. What do these leaders have in common? They invest twice as much as the market average in training and acculturation. These companies have not only stated their intention to innovate, they have also enabled their employees to embrace these tools in their daily work.
Source: McKinsey – The state of AI in 2026 / Gartner – Gen AI Maturity Report 2026.
The operational challenges of AI in 2026: governance, European compliance, and managing team fatigue
The problem is no longer accessing AI, but transforming the intention in a sustainable way without exhausting your teams or compromising security.
The glass ceiling of skills: beyond the simple use of "prompts"
Even in 2026, the skills gap remains the primary obstacle to adoption. The World Economic Forum highlights that the gap between "passive users" and "orchestrators" is widening. It is no longer enough to know how to ask an AI a question; you need to know how to orchestrate autonomous agents, verify the accuracy of the data produced, and integrate these outputs into a complex workflow. At Boeckli & Gomes, we have noticed that anxiety about skills becoming obsolete has turned into a pressing demand for meaning.
The AI Act and compliance as a driver of trust for businesses
Since 2025, the European AI Act has entered its full implementation phase. Companies that anticipated compliance (model transparency, bias management, personal data protection) now have a major advantage: trust. Conversely, those that remained vague in their intentions and lacked governance now face legal risks and growing mistrust from both employees and customers.
"AI fatigue": a new managerial risk linked to constant change
A new phenomenon emerged at the end of 2025: weariness in the face of constant change. Teams, asked to test new tools every month, are showing signs of exhaustion. Without a deep acculturation that provides rest and clear methods, AI is perceived as a supervisor or an unbearable pace accelerator, which blocks any real adoption.
Performance strategies in 2026: concrete examples of AI serving human expertise
Today, AI does not replace humans: it amplifies them for those who have been able to transform intention into professional reflexes.
Practical cases of professional application in 2026
- Management and HR: automation of 360° feedback summaries. AI processes raw data, but managers focus exclusively on face-to-face interviews, freeing up time for empathetic listening and employee development.
- Marketing and sales: use of specialized agents to personalize commercial proposals based on the customer's complete history and current context (news, financial reports). Preparation time is reduced by 70%, but the final decision and negotiation remain 100% human.
- Technical operations: computer vision-assisted quality control where AI suggests anomalies. The operator is no longer the one who "looks for" the error, but the one who "understands" why it occurred and how to improve the production process.
