# Resistance to AI at work: why it happens and how to avoid it Author: Quadra URL: https://quadra-website.matt-madd.chatgpt.site/risorse/articoli/resistenza-ai-in-azienda Published: 2026-09-18 Updated: 2026-09-18 Language: en Reading time: 5 min AI almost always arrives from the top. That is why people reject it, and almost nobody says so. ## At a glance - Resistance to AI often begins with how it is introduced: a change imposed without involving the people doing the work. - Concerns about roles, skills and quality make sense. Listening helps identify more useful processes and tools. - Involve people in trials, clarify responsibilities and assess what really improves everyday work. There is a scene that repeats in many companies. A meeting, an announcement, a new platform. Then, a month later, someone checks the logs and discovers that four people use it. The official explanation is always the same: resistance to change, people fear technology, more training is needed. It is a convenient explanation because it shifts the problem onto those who made none of the decisions. There is a less comfortable reading: people are not resisting the technology. They are responding to the way it arrived. ## What the data says A survey conducted in early 2026 among 2,400 people in the United States and Europe, half executives and half employees, all already using AI tools at work, describes a company split in two. 79% of executives report difficulties with adoption: returns that do not materialise, unclear strategy and internal tensions. 75% admit their company's AI strategy is more about appearances than direction. 56% report internal power struggles. 92% are cultivating an “AI elite” among employees, and 60% plan to dismiss those who do not adapt. On the other side, 29% of employees admit to having sabotaged AI adoption at their company in some form. Among younger employees, the figure reaches 44%. A necessary qualification: that research was commissioned by a company selling AI platforms, so it has an interest in describing a problem it can then solve. But the figures align with independent findings: according to Gallup, which surveyed more than 23,000 American workers in February 2026, only around one in ten in companies that have adopted AI believes it has truly changed how work is done. Together, the two findings say one thing. **The technology has arrived almost everywhere; the work has changed almost nowhere.** ## Why resistance makes sense If one employee in four is working against the project, the problem is not a personality trait. Someone who has worked in a process for eight years knows things that are not written anywhere. They know that client needs a call before the confirmation is sent, that the item code has been wrong in the system since 2019, that logistics accepts no changes on Fridays. When a tool arrives promising to do their job and nobody has asked them anything, they understand two things at once: their experience has been considered irrelevant, and the result will be worse than their own. They are right on both counts. At that moment, the project is already over, months before anyone looks at usage figures. The 75% of executives who admit to a strategy “for appearances” completes the picture. People can tell. It is hard to commit to a change that the people announcing it are not taking seriously. ## Where small businesses have an advantage This is one of the few cases where being small really helps. In a multinational, the AI project starts three thousand kilometres from the people who will use it. In an eighty-person company, decision-makers and those doing the work meet in the corridor. The person who knows the process is ten minutes away, and implicit rules can be uncovered through conversation rather than six months of analysis. That advantage is not automatic, though. It is lost the moment the announcement comes before the conversation. ## Where we are heading The risk visible in the data is the emergence of two categories of employees: those who master these tools and become indispensable, and those who fall behind and are considered a cost. If that division forms on its own, a company ends up with one group who know how to use AI but do not deeply understand the trade, and another who know the trade but have been shut out. These are precisely the two halves that needed to work together. The sensible direction is the opposite, and it is already visible where adoption works: start with the person doing the work, use their knowledge to define what the system should do, and give them a tool that solves a problem they actually had. This is not organisational kindness; it is the only way to get a working system, because only they know the rules that make it useful. ## The question to ask The decision to adopt AI belongs to the people running the business, and it is right for it to come from the top. Nobody in accounts wakes up asking to automate their own work, and waiting for that means never starting. What cannot come from the top is the content. Which process, with which rules and which exceptions: these are things the people doing the work every day know, and there is no way to derive them from a management meeting or a supplier demo. So the question is not who asked for this tool. It is: **did the people who will use it help define it?** If the answer is no, the problem will not show at launch. It will show three months later, when usage figures reveal that four people use it and someone concludes that people are afraid of change. > **The fastest way to burn a budget is to reverse the order: choose the tool, then ask people to adapt. The fastest way to make it work is to decide from the top and define from the bottom.** ## Sources and further reading **Sources:** WRITER and Workplace Intelligence, “AI Adoption in the Enterprise”, survey of 2,400 executives and employees in the USA and Europe, December 2025–January 2026 (research commissioned by an AI platform provider) · Gallup, survey of 23,717 US workers, February 2026 · SHRM, “Navigating AI in the Workplace 2026” - [WRITER e Workplace Intelligence · AI Adoption in the Enterprise 2026](https://go.writer.com/ai-adoption-enterprise-2026) — Edition 2026. Accessed: 2026-09-18. - [Gallup · Rising AI Adoption Spurs Workforce Changes](https://www.gallup.com/workplace/704225/rising-adoption-spurs-workforce-changes.aspx) — Published 12 April 2026. Accessed: 2026-09-18. - [SHRM · Navigating AI in the Workplace: 2026](https://www.shrm.org/topics-tools/research/navigating-ai-in-the-workplace) — Published 17 June 2026. Accessed: 2026-09-18.