As artificial intelligence (AI) continues to reshape how we work, many companies are focused on reselling teams for an AI-driven future. But new research from Harvard Business School’s Letian Zhang shows that the most important skills aren’t necessarily the newest ones, but the foundational ones. 

Harvard Business School’s Working Knowledge reports that in a study analyzing 70 million job transitions and nearly 1,000 occupations, Zhang and his coauthors found that soft skills like communication, reading comprehension, and critical thinking often form the base for more technical specialized abilities. These “nested structures” of skill-building mirror how we learn. 

The research found that nearly 80 percent of the wage premium associated with advanced, technical skills actually depends on the underlying general skills a person already has. 

For employers, it’s a reminder that teaching coding, for example, without supporting the ability to communicate or problem-solve won’t lead to long-term success. It also helps explain why many returning programs fall short, not because people can’t learn, but because we underestimate the scaffolding that learning requires.

Zhang’s findings carry weight across policy and education too, especially as the report highlights racial and gender disparities in access to high-value skill tracks. The solution isn’t teaching new tools but expanding access to the foundational skill-building early on.