More AI-written code means tighter tests, not less risk
In brief
According to Spotify, wider use of AI in development saw monthly accepted code changes rise from about 8,100 to 17,000 in August versus a year earlier. The company then adjusted monitoring, testing, rollout and rollback after the extra volume outpaced some checks. The same report cites Google's 2025 DORA research, which linked AI adoption to higher throughput and lower delivery stability.
Source : InfoQ - « Spotify More Than Doubled Monthly Code Changes as AI Accelerated Development »
What it changes for an SME
You do not run systems on Spotify's scale. If your teams already use AI to write software, you may spend less time drafting code, but only if tests, review and rollback keep pace. According to Spotify, change volume outgrew some checks, and an automated upgrade failed in production after passing existing controls. For an SME of 10 to 250 people, the issue is operational risk; this report does not describe a new compliance duty.
What you can do
- Map where AI already writes or changes code in your firm, including automated upgrades.
- Check that you can roll back a release and that tests run before a change is accepted.
- Time AI-assisted or automated changes for hours when someone can respond.
Let's talk about the time you could save.
In 20 minutes, we find where AI could save you time or money. Free, no commitment.