Reflection — AI's first project isn't a tool, it's redesigning the roles

When an organization brings in AI, the deepest lever isn't the model: it's explicitly redefining the human roles around it. The operator becomes a supervisor of algorithms, the analyst becomes a narrator of what the data says, the front-line agent becomes an escalation specialist. The research converges: augmentation supersedes automation as the design principle, and role redesign is the micro-foundation that decides adoption.

The World Economic Forum measures it: three-quarters of companies plan to adopt AI, yet fewer than half have redesigned the roles around it. Leaving the role implicit manufactures resistance and shadow usage. For a small or mid-sized company, the first decision isn't which tool to buy: it is to rewrite, in black and white and before any rollout, the job description of every role AI will touch, then present it to leadership before the first hands-on use.

Source: World Economic Forum, "Future of Jobs Report," 2023; A. Quintin, integrative literature review on organizational capabilities for AI adoption in SMEs (282 papers, 2015-2025), EDBA École des Ponts Business School, 2025.