Companies track employee AI use in detail to assess productivity
Updated
Updated · CNBC · May 5
Companies track employee AI use in detail to assess productivity
14 articles · Updated · CNBC · May 5
ModelOp said over two-thirds of enterprises still estimate AI returns, while Meta tests tracking keystrokes and mouse movements and Microsoft offers prompt, user and agent monitoring tools.
Companies can measure token costs closely, but executives and advisers say usage is a poor proxy for performance and proving AI caused productivity gains remains difficult.
Some firms are restructuring work around AI agents and pods, with Coinbase cutting 14% of staff, while Salesforce says workflow-level metrics better link AI activity to revenue, savings and customer outcomes.
Beyond tracking tokens, how are firms proving AI's real financial impact?
With AI doing the work, are 'token costs' the new corporate payroll?
Is measuring AI productivity creating a new era of workplace surveillance?
Balancing AI Monitoring, Employee Privacy, and Performance Metrics in the Modern Workplace
Overview
Companies are increasingly monitoring employee interactions with AI using advanced tools to improve productivity, security, and compliance. For example, Meta's software tracks keystrokes and screen activity, sparking employee privacy concerns and backlash. To address risks like unauthorized 'Shadow AI' use, firms deploy detection tools and implement safeguards. AI monitoring also influences performance reviews, with failure to meet AI benchmarks affecting promotions, while workforce shifts favor skilled technical roles over routine jobs. However, constant surveillance raises stress and fairness issues, leading to lawsuits and stricter regulations like the EU AI Act that demand transparency and human oversight. Effective AI governance, combining ethical principles and employee trust, is essential as AI becomes deeply embedded in workflows and workplace norms evolve.