McKinsey’s QuantumBlack AI leader, Debasish Patnaik, says the firm controls costs by monitoring consultants’ AI usage, specifically token consumption. He describes a process where the company checks activity and “pings” staff to flag excessive use or potential overspending.
Patnaik’s comments focus on practical steps to manage spending as generative AI models charge based on tokens processed. The reporting frames the approach as part of McKinsey’s internal cost governance for AI tools, aimed at keeping usage aligned with budgets and preventing runaway expenses.
Across the coverage, the emphasis is on cost-management mechanics rather than on changes to products or client outcomes. One outlet reports the details of the internal pings, while the other provides no additional distinct information beyond the shared headline topic. Together, the sources indicate that McKinsey treats token usage as a trackable metric and uses internal alerts to steer staff toward more efficient consumption of AI resources.
