Is that AI agent worth it? Agentic economics and the modern operating model

By McKinsey & Company · July 22, 2026.

This McKinsey Artificial Intelligence alert email announces a new McKinsey Quarterly piece titled "Is that AI agent worth it? Agentic economics and the modern operating model." The mailing is signed by Dan Tinkoff and Lieven Van der Veken, global leaders of QuantumBlack, AI by McKinsey.

The body of the email is brief and works as a teaser: it does not include the full development of the article, only its central idea. According to the text itself, as companies focus on escalating AI costs, they must make sure not to lose sight of "the real prize" — that is, the business value these initiatives should generate beyond spending control. The email mentions that the article describes "6 patterns," though it does not specify what they are, so it is not possible to list them without inventing content.

Beyond the main article, the email points to two related pieces under the "Also Consider" section: "Frontiers of compute: The technologies to reduce AI inference costs" (on technologies for reducing AI inference costs) and "Cost versus value: Managing agentic AI system performance" (on how to manage the performance of agentic AI systems by balancing cost and value). The email does not include verifiable final URLs for any of these links, since the source HTML has stripped them of their addresses.

Broadly speaking, this type of McKinsey content usually fits into the industry debate over the return on investment (ROI) of agentic AI projects in enterprises, a recurring theme in 2025-2026 given rising spending on infrastructure and models against still-uneven business results. That said, this is general industry context and not an explicit data point from the email.