The economic model of ESNs, or Digital Service Companies, is changing rapidly with the rise of generative AI. These companies have traditionally earned income by selling time—essentially billing clients based on how long it takes to install or maintain software. However, for the first time, ESNs are seeing a decline in the domestic market, with a reported -1.8% drop in 2025. This shift is partly due to the growing dominance of AI agents and the interfaces between them, known as MCP (Machine Communication Protocols), which are changing how software is integrated and used. Labor unions, such as Numeum, have pointed out that the way growth is measured is also changing. In the past, companies tracked progress using "man-days" or "month-days"—a way of quantifying human labor. However, AI does not require such metrics, as it can perform tasks without the need for human time. This has forced ESNs to rethink how they value their services and how they are paid for them. The increasing use of AI across industries has led to a surge in spending on AI tokens, which are digital credits used to power AI systems. For example, Uber used up its entire annual AI programming budget in just four months, while one Microsoft employee spent $28,000 on AI tokens in 28 days. This trend shows that businesses are investing heavily in AI, and ESNs, with their close relationships to customer IT systems, are well positioned to help manage and optimize this usage. The challenge for ESNs now is to shift from an economic model based on time spent to one based on the value delivered. Instead of billing by the hour or by the number of AI agents deployed, companies are increasingly using contracts tied to results and impact. McKinsey has reported that about a quarter of global fees in 2025 came from such impact-indexed contracts. To support this new model, ESNs are developing tools like AI fuel cards, which track and control AI token usage at the point of use. These tools help make AI costs more visible and encourage responsible use by setting budgets per use case rather than by volume. This transition requires ESNs to move from selling time to selling value, which means creating new contract structures, defining clear performance metrics, and building stronger partnerships with clients. While this shift presents challenges, it also offers opportunities for ESNs to evolve their services and skills to meet the demands of a rapidly changing AI-driven landscape.