Don't Want to Use Artificial Intelligence? Skeptical? Your Colleague Might Change Your Mind
Investment is not the biggest problem for companies when it comes to AI, but rather the concrete adoption of new generative tools by employees. Many workers show distrust towards generative AI, due to concerns about the reliability of results and the difficulty of changing established habits.
To change the status quo and make the best of their investments, according to a Wall Street Journal article, companies are introducing roles called AI Champions. These are regular employees who have shown more openness to this new opportunity. These employees receive early access to new platforms, dedicated training, and maintain a direct relationship with management.
Their task is not only to promote AI tools but, more importantly, to assist colleagues who are "struggling" during the initial stages of use, addressing doubts and demonstrating practical applications in daily activities.
This approach seems to produce results. According to a report published in June by the Boston Consulting Group, 74% of non-managerial employees currently use artificial intelligence tools regularly, meaning daily or multiple times a week. In the previous 2025 study, the percentage was 51%. The survey involved 11,749 workers across 14 markets.
One significant example cited by the WSJ comes from the law firm Ropes & Gray, where AI Champions have facilitated the spread of Harvey, a specialized platform for the legal sector. Two years ago, only 32 users were generating a few hundred prompts each month. Today, almost 2,200 employees use the service, producing over 282,000 monthly prompts. The firm has about 3,000 employees worldwide, of which around 1,500 are lawyers.
Howard Glazer, co-head of the global division dedicated to private equity operations and currently also the internal head of the AI practice, explains that resistance often stems from the natural difficulty of accepting change. In the legal field, the problem is amplified by widely reported cases in recent years of court documents containing non-existent citations generated by artificial intelligence, incidents that have fueled distrust towards these tools.
For this reason, Glazer believes that one-on-one interactions are more effective than traditional collective training sessions. In direct conversations, specific doubts can be addressed, and it can be demonstrated how AI can speed up work without replacing human oversight. Another frequently used argument is that even the work of younger professionals typically requires revisions before reaching the final version: the same principle can be applied to AI-generated content.
A similar strategy is adopted by Citigroup, where Josh Goldsmith, head of digital solutions and innovation for Internal Audit, plays the role of AI Champion. The goal is to teach colleagues to use artificial intelligence on real activities, such as generating draft audit reports through the proprietary platform Citi Stylus Workspaces.
The bank has stated its intention to integrate AI into numerous business processes, from client onboarding to financing practice analysis, as well as the development of virtual financial advisors, cybersecurity tools, and software programming support. To support this transformation, it has created a program involving over 4,000 people, divided into "Champions," generally senior figures, and "Accelerators," from various levels of the organization.
According to Citi, the adoption of Stylus Workspaces has increased from single-digit percentages, recorded at the end of 2024 during the initial rollout, to over 80% of the company population that can access the tool. The company attributes part of this growth to the Champion and Accelerator program.
The role of AI Champions does not require that ideal candidates are necessarily the most technically skilled. The most important characteristics are curiosity, communication skills, and the ability to bridge business needs and the potential offered by technology.
Companies hope that the acceleration of the cultural change needed to integrate AI into daily work processes will come not from a top-down solution, which is often resented by employees, but through a persuasive effort born of peer relationships.