What’s the ROI of software engineers’ spending on AI?

Gergely Orosz of The Pragmatic Engineer asked 27 companies about AI spending and found costs per engineer have soared while gains in delivery range from none to threefold.

Gergely Orosz writes The Pragmatic Engineer, a newsletter for software engineers and their managers. A reader asked what other companies spend on AI per engineer and whether it raised output. Orosz gathered figures from 27 workplaces and from two AI code review vendors, CodeRabbit and Gitar. Spending runs from under $300 to $3,000 a developer a month, and the reported gains vary just as widely.

• CodeRabbit, which makes an AI code review tool, reports that its users open more than twice as many pull requests a week as a year ago.
• A two-person software agency in Europe spends $3,000 per developer a month and now opens ten or more pull requests a day, up from one or two.
• At a listed software company with about 600 engineers, more pull requests did not mean faster merges. Its human reviewers still catch unneeded or oversized changes that AI review misses.
• At another listed company pull requests did not rise in number but tripled in size. One team at a multinational closes three times as many user stories as before.

A team that tracks only pull request counts may not learn whether its AI spending delivers value. The team that compared closed user stories had a clearer answer. This is an inference and is not the author's claim. Only the open part of this paid article was read.

Derived by Working Surface from the article; more in the Playbook. Source line: However, when the team cares about the quality in production code and wants to keep it like before AI, then actual development speed had not increased much.

21 October 2026: as AI tools multiply code, documents and prototypes, the work that counts moves to judging, reviewing and keeping track of what they produce.