What's happened
Goldman Sachs has moved from evaluating individual AI tool use to tracking team velocity. CIO Marco Argenti says the focus is on how quickly teams move from idea to production and on cross‑team collaboration, rather than monitoring developers one by one. The bank is measuring feature delivery and AI token usage to guide budgeting and quality control.
What's behind the headline?
key takeaways
- Team over individual: Argenti argues that measuring an entire team’s velocity provides a clearer picture of AI-driven productivity than tracking individual developers.
- End-to-end focus: The emphasis is on moving from idea to production and improving delivery timelines rather than comparing personal usage statistics.
- Budget signals: Tracking AI token consumption informs budgeting decisions and resource allocation.
implications for firms
- Expect more cross‑functional dashboards that highlight flow and backlog burn-down rather than personal metrics.
- Performance metrics will likely shift toward feature lead time, quality controls, and production readiness rather than lines of code.
- This approach could reduce over‑scaling concerns tied to individual activity, encouraging collaboration across teams.
How we got here
The interview with Goldman Sachs’ CIO occurs against a backdrop of firms integrating AI into software development. While some companies track individuals, Goldman is prioritising team performance and project throughput to ensure AI initiatives deliver faster, more reliable outcomes.
Our analysis
Business Insider UK; Goldman Sachs leadership materials; industry commentary on developer productivity metrics.
Go deeper
- Will teams be rewarded for faster feature delivery or for higher quality outputs?
- How will Goldman ensure accountability if a project stalls despite high team velocity?
- What other banks are adopting similar team-centric AI metrics?