BCG: AI project failure is mainly a people problem

On 19 May 2026, Boston Consulting Group published the article Your AI Change Is Actually a People Change. Its main message is uncomfortable for companies hoping that a better tool will solve the problem. According to a BCG study from late 2025, only five percent of companies achieve substantial value from AI. BCG also says that 60 percent of companies report no material value. In successful AI transformations, according to the firm, 70 percent of value comes from people-related actions rather than technology. This is a sharp correction for organisations that confuse tool deployment with work change. Authors Julia Dhar, Kristy Ellmer, Philip Jameson, David Martin, Vladimir Lukic and Vikram Sivakumar base the article on behavioural science. They argue that AI transformations fail because of unclear goals, weak employee adoption, false leadership alignment and a poorly explained reason for change. BCG warns against false alignment. Managers may outwardly agree that AI is a priority, but each may mean something different. One wants lower costs, another wants growth, a third wants new capabilities and a fourth only does not want to fall behind. Without an explicit decision, companies create collections of projects that do not reinforce each other. The practical advice is to focus on a smaller number of essential workflows. According to BCG, successful companies prioritise three to four use cases on average, while less successful ones pursue six to seven. An AI agenda should therefore not be a workshop idea list, but an investment decision with an owner, a metric and an impact on how people work.

BCG: AI project failure is mainly a people problem

The full article is available to users registered at Hard Skills