AI-assisted delivery amplifies the system around it
What the DORA 2025 research means for engineering leaders adopting AI-assisted software development.
Treat AI as an amplifier
DORA's 2025 research describes AI as an amplifier of an organisation's existing strengths and weaknesses. That is a useful correction to programmes that focus on licences or generated-code volume without examining the delivery system.
If review queues, brittle tests, unclear ownership, or slow environments already constrain delivery, producing more change can increase work in progress and operational risk.
Strengthen the feedback system
Short-lived branches, reliable automated tests, fast environments, observability, and small releases make assisted changes easier to verify. Teams also need clear standards for provenance, review, security, and when generated output is inappropriate.
- Track lead time, change failure, recovery, and user outcomes.
- Review generated changes with the same accountability as other code.
- Invest in platform paths that make the safe action the easy action.
Run an adoption experiment
Choose a bounded workflow, establish a baseline, define the quality guardrails, and compare the full delivery outcome. Expand only when the evidence shows improved flow without hidden rework or degraded reliability.