Efficiency metrics versus outcome-based goals

Volume-based success. Some organizations define AI progress through throughput and cost metrics. Incident.io increased its shipping frequency to 140 releases daily with 50 engineers, using AI to generate parts of its code in 43% of pull requests (Ivan Klaric from The Pragmatic Engineer). At the same time, Cognizant reports that over 50% of its code is now AI-assisted, while CareCloud reduced manual claim touches by 44% in its billing division (AI Weekly). These teams prioritize velocity, treating AI as a tool to uncouple platform scale from headcount growth.

Outcome-focused rigor. Other leaders argue that volume metrics mask poor quality and unclear thinking. BT Group tested a challenger AI that pushed employees to define specific goals, resulting in 3.5x as many objectives rated as excellent compared to standard, agreeable AI tools (Confluence). Gaurav Jain notes that AI often exposes existing management failures, such as poor delegation or review habits, rather than creating new ones (The Good Boss). Jason Rigby advocates for protecting the time saved by AI to focus on human-centric tasks, such as coaching or customer interaction, rather than simply cranking out more volume (Jason Rigby from The Self-Aware Leader).

These positions diverge on what constitutes a return on investment. One side views AI as a factory for output, while the other views it as a lever to improve the substance of work. Success depends on whether managers measure the volume of code shipped or the quality of the business outcomes achieved.